Model training techniques for channel characteristic prediction procedures
AI/ML models at UEs predict uplink channel characteristics using SRS resource sets and measurement results, addressing inefficiencies in beam management, enhancing accuracy, latency, and resource use in wireless communications.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- QUALCOMM INC
- Filing Date
- 2024-12-23
- Publication Date
- 2026-07-02
Smart Images

Figure CN2024141350_02072026_PF_FP_ABST
Abstract
Description
MODEL TRAINING TECHNIQUES FOR CHANNEL CHARACTERISTIC PREDICTION PROCEDURESFIELD OF TECHNOLOGY
[0001] The following relates to wireless communications, including model training techniques for channel characteristic prediction procedures.BACKGROUND
[0002] Wireless communications systems are widely deployed to provide various types of communication content such as voice, video, packet data, messaging, broadcast, and so on. These systems may be capable of supporting communication with multiple users by sharing the available system resources (e.g., time, frequency, and power) . Examples of such multiple-access systems include fourth generation (4G) systems such as Long Term Evolution (LTE) systems, LTE-Advanced (LTE-A) systems, or LTE-A Pro systems, and fifth generation (5G) systems which may be referred to as New Radio (NR) systems. These systems may employ technologies such as code division multiple access (CDMA) , time division multiple access (TDMA) , frequency division multiple access (FDMA) , orthogonal FDMA (OFDMA) , or discrete Fourier transform spread orthogonal frequency division multiplexing (DFT-S-OFDM) . A wireless multiple-access communications system may include one or more base stations, each supporting wireless communication for communication devices, which may be known as user equipment (UE) .SUMMARY
[0003] The systems, methods, and devices of this disclosure each have several innovative aspects, no single one of which is solely responsible for the desirable attributes disclosed herein.
[0004] A method for wireless communications by a user equipment (UE) is described. The method may include receiving one or more messages including scheduling information for one or more sounding reference signal (SRS) resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results, and obtaining one or more predicted uplink channel characteristics associated uplink transmission based on training the one or more machine learning models.
[0005] A UE for wireless communications is described. The UE may include one or more memories storing processor executable code, and one or more processors coupled with the one or more memories. The one or more processors may individually or collectively be operable to execute the code to cause the UE to receive one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, receive, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, train one or more machine learning models based on the one or more SRS resource sets and the set of measurement results, and obtain one or more predicted uplink channel characteristics associated uplink transmission based on training the one or more machine learning models.
[0006] Another UE for wireless communications is described. The UE may include means for receiving one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, means for receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, means for training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results, and means for obtaining one or more predicted uplink channel characteristics associated uplink transmission based on training the one or more machine learning models.
[0007] A non-transitory computer-readable medium storing code for wireless communications is described. The code may include instructions executable by one or more processors to receive one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, receive, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, train one or more machine learning models based on the one or more SRS resource sets and the set of measurement results, and obtain one or more predicted uplink channel characteristics associated uplink transmission based on training the one or more machine learning models.
[0008] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, the set of measurement results correspond to an SRS resource set of the one or more SRS resources sets, and the set of measurement results include at least one of: one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest reference signal receive power (RSRP) or a highest selection probability, one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set. or one or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.
[0009] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, the one or more predicted uplink channel characteristics include one or more temporal predictions, and obtaining the one or more predicted uplink channel characteristics may include operations, features, means, or instructions for obtaining the one or more temporal predictions based on the training of the one or more machine learning models using one or more previously received SRS resource sets and corresponding sets of previously received measurement results, where the corresponding sets of previously received measurement results include one or more RSRP measurements associated with each SRS resource of the one or more previously received SRS resource sets.
[0010] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, the one or more predicted uplink channel characteristics include one or more spatial predictions, and obtaining the one or more predicted uplink channel characteristics may include operations, features, means, or instructions for obtaining the one or more spatial predictions using a set of candidate channel characteristics obtained during the training of the one or more machine learning models and corresponding one or more measurement results, where corresponding sets of previously received measurement results include one or more RSRP measurements associated with each SRS resource of one or more previously received SRS resource sets, and where the one or more spatial predictions indicate a quantity of channel characteristics that may be less than a quantity of the set of candidate channel characteristics.
[0011] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, the one or more SRS resource sets include at least a first SRS resource set and a second SRS resource set and the set of measurement results corresponding to the first SRS resource set includes a set of RSRP measurements for each SRS resource included in the first SRS resource set, and the set of measurement results corresponding to the second SRS resource set includes a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.
[0012] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for receiving signaling including an associated identifier (ID) that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink transmission reception point (TRP) , where the set of measurement results may be based on one or more SRS resource sets.
[0013] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, consistency of one or more network conditions between the training of the one or more machine learning models and an inference phase of the one or more machine learning models may be based on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.
[0014] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, the one or more network conditions include at least one of: a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase; a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics; a quantity of SRS resources or SRS resource sets; a periodicity of the one or more SRS resource sets; or one or more types of measurements included in the set of measurement results.
[0015] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, receiving the set of measurement results may include operations, features, means, or instructions for receiving an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, where a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink TRP, to an uplink channel characteristic, or both.
[0016] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for predicting, using the one or more machine learning models, a set of uplink TRP identifiers for future selection, a set of uplink channel characteristic identifiers for future selection, or both, based on the group of measurement results.
[0017] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for receiving one or more channel state information reference signal (CSI-RS) identifiers, one or more synchronization signal block (SSB) reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models, obtaining a set of reference signal measurements corresponding to the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof, where training the one or more machine learning models includes, and training the one or more machine learning models based on the set of reference signal measurements.
[0018] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, receiving the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof may include operations, features, means, or instructions for receiving an indication of a CSI-RS resource, an indication of an SSB reference signal resource, or any combination thereof, where the CSI-RS resource, the SSB reference signal resource, or any combination thereof may be associated with corresponding downlink channel characteristics.
[0019] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for receiving assistance information signaling including at least one of TRP location information including a location of a downlink TRP, a location of an uplink TRP, a location of the downlink TRP relative to the location of the uplink TRP, or spatial direction information for a quantity of downlink reference signals received by the UE.
[0020] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for transmitting one or more UE capability reports indicating content requested to be included in the set of measurement results.
[0021] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for receiving, for the training of the one or more machine learning models indicated by an associated ID, signaling indicative of: an indication of a quantity of candidate SRS resource sets of the one or more SRS resource sets, an indication of a quantity of candidate SRS resources corresponding to the quantity of candidate SRS resource sets, an indication of one or more types of measurements included in the set of measurement results, or any combination thereof.
[0022] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for transmitting one or more messages indicative of one or more preferences selected from the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof.
[0023] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for transmitting one or more UE capability reports indicating support for training and inference of the one or more machine learning models using the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, where the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, may be each indicated using the associated ID.
[0024] Some examples of the method, UEs, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for transmitting a request for a different quantity of candidate SRS resource sets, a different quantity of candidate SRS resources, or a different set of measurements included in the set of measurement results, or any combination thereof and receiving, via radio resource control (RRC) reconfiguration signaling, an indication of the different quantity of candidate SRS resource sets, the different quantity of candidate SRS resources, or the different set of measurements included in the set of measurement results, or any combination thereof, based on the request.
[0025] In some examples of the method, UEs, and non-transitory computer-readable medium described herein, the one or more parameters associated with the respective downlink reference signals include transmission configuration indicator (TCI) state information, spatial relation information, or both.
[0026] A method for wireless communications by a network entity is described. The method may include outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, and performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0027] A network entity for wireless communications is described. The network entity may include one or more memories storing processor executable code, and one or more processors coupled with the one or more memories. The one or more processors may individually or collectively be operable to execute the code to cause the network entity to output one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, output a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, and perform one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0028] Another network entity for wireless communications is described. The network entity may include means for outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, means for outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, and means for performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0029] A non-transitory computer-readable medium storing code for wireless communications is described. The code may include instructions executable by one or more processors to output one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals, output a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets, and perform one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0030] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, the set of measurement results correspond to an SRS resource set of the one or more SRS resource sets, and the set of measurement results include at least one of: one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest RSRP or a highest selection probability, one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set, or one or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.
[0031] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, the one or more SRS resource sets include at least a first SRS resource set and a second SRS resource set and the set of measurement results corresponding to the first SRS resource set includes a set of RSRP measurements for each SRS resource included in the first SRS resource set, and the set of measurement results corresponding to the second SRS resource set includes a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.
[0032] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for outputting signaling including an associated ID that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink TRP, where the set of measurement results may be based on one or more SRS resource sets.
[0033] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, consistency of one or more network conditions between training of the one or more machine learning models and an inference phase of the one or more machine learning models may be based on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.
[0034] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, the one or more network conditions include at least one of: a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase, a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics, a quantity of SRS resources or SRS resource sets, a periodicity of the one or more SRS resource sets, or one or more types of measurements included in the set of measurement results.
[0035] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, outputting the set of measurement results may include operations, features, means, or instructions for outputting an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, where a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink TRP, to an uplink channel characteristic, or both.
[0036] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for outputting one or more CSI-RS identifiers, one or more SSB reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models, where the one or more machine learning models may be based on a set of reference signal measurements corresponding to the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof.
[0037] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, outputting the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof may include operations, features, means, or instructions for outputting an indication of a CSI-RS resource, an indication of an SSB reference signal resource, or any combination thereof, where the CSI-RS resource, the SSB reference signal resource, or any combination thereof may be associated with corresponding downlink channel characteristic identifiers.
[0038] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for outputting assistance information signaling including at least one of TRP location information including a location of a downlink TRP, a location of an uplink TRP, a location of the downlink TRP relative to the location of the uplink TRP, or spatial direction information for a quantity of downlink reference signals received by the UE.
[0039] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for obtaining one or more UE capability reports indicating content requested to be included in the set of measurement results.
[0040] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for outputting, for training of the one or more machine learning models indicated by an associated ID, signaling indicative of an indication of a quantity of candidate SRS resource sets of the one or more SRS resource sets, an indication of a quantity of candidate SRS resources corresponding to the quantity of candidate SRS resource sets, an indication of one or more types of measurements included in the set of measurement results, or any combination thereof.
[0041] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for obtaining one or more messages indicative of one or more preferences selected from the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof.
[0042] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for obtaining one or more UE capability reports indicating support for training and inference of the one or more machine learning models using the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, where the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, may be each indicated using the associated ID.
[0043] Some examples of the method, network entities, and non-transitory computer-readable medium described herein may further include operations, features, means, or instructions for obtaining a request for a different quantity of candidate SRS resource sets, a different quantity of candidate SRS resources, or a different set of measurements included in the set of measurement results, or any combination thereof and outputting, via RRC reconfiguration signaling, an indication of the different quantity of candidate SRS resource sets, the different quantity of candidate SRS resources, or the different set of measurements included in the set of measurement results, or any combination thereof, based on the request.
[0044] In some examples of the method, network entities, and non-transitory computer-readable medium described herein, the one or more parameters associated with the respective downlink reference signals include TCI state information, spatial relation information, or both.
[0045] Details of one or more implementations of the subject matter described in this disclosure are set forth in the accompanying drawings and the description below. Other features, aspects, and advantages will become apparent from the description, the drawings, and the claims. Note that the relative dimensions of the following figures may not be drawn to scale.
[0046] The foregoing has outlined rather broadly the features and technical advantages of examples according to the disclosure in order that the detailed description that follows may be better understood. Additional features and advantages will be described hereinafter. The conception and specific examples disclosed may be readily utilized as a basis for modifying or designing other structures for carrying out the same purposes of the present disclosure. Such equivalent constructions do not depart from the scope of the appended claims. Characteristics of the concepts disclosed herein, both their organization and method of operation, together with associated advantages will be better understood from the following description when considered in connection with the accompanying figures. Each of the figures is provided for the purposes of illustration and description, and not as a definition of the limits of the claims.
[0047] While aspects and embodiments are described in this application by illustration to some examples, those skilled in the art will understand that additional implementations and use cases may come about in many different arrangements and scenarios. Innovations described herein may be implemented across many differing platform types, devices, systems, shapes, sizes, packaging arrangements. For example, embodiments and / or uses may come about via integrated chip embodiments and other non-module-component based devices (e.g., end-user devices, vehicles, communication devices, computing devices, industrial equipment, retail / purchasing devices, medical devices, artificial intelligence (AI) -enabled devices, etc. ) . While some examples may or may not be specifically directed to use cases or applications, a wide assortment of applicability of described innovations may occur. Implementations may range in spectrum from chip-level or modular components to non-modular, non-chip-level implementations and further to aggregate, distributed, or original equipment manufacturer (OEM) devices or systems incorporating one or more aspects of the described innovations. In some practical settings, devices incorporating described aspects and features may also necessarily include additional components and features for implementation and practice of claimed and described embodiments. For example, transmission and reception of wireless signals necessarily includes a number of components for analog and digital purposes (e.g., hardware components including antenna, radio frequency (RF) -chains, power amplifiers, modulators, buffer, processor (s) , interleaver, adders / summers, etc. ) . It is intended that innovations described herein may be practiced in a wide variety of devices, chip-level components, systems, distributed arrangements, end-user devices, etc. of varying sizes, shapes, and constitution.BRIEF DESCRIPTION OF THE DRAWINGS
[0048] FIG. 1 shows an example of a wireless communications system that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0049] FIG. 2 shows an example of a network architecture that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0050] FIG. 3 shows an example of a wireless communications system that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0051] FIG. 4 shows examples of artificial intelligence / machine learning (AI / ML) implementations that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0052] FIG. 5 shows an example of an AI / ML operation flow that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0053] FIGs. 6 and 7 show examples of AI / ML beam management procedures that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0054] FIG. 8 shows examples of AI / ML performance monitoring configurations that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0055] FIG. 9 shows an example of an AI / ML parameter consistency evaluation flow that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0056] FIGs. 10, 11, and 12 show examples of wireless communications systems that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0057] FIGs. 13 and 14 show block diagrams of devices that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0058] FIG. 15 shows a block diagram of a communications manager that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0059] FIG. 16 shows a diagram of a system including a device that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0060] FIGs. 17 and 18 show block diagrams of devices that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0061] FIG. 19 shows a block diagram of a communications manager that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0062] FIG. 20 shows a diagram of a system including a device that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.
[0063] FIGs. 21 and 22 show flowcharts illustrating methods that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure.DETAILED DESCRIPTION
[0064] Some wireless communications systems may experience a relatively dense deployment of wireless communications devices, including multiple user equipment (UEs) within a same location (e.g., within a same coverage area, cell, geographic location) . In such cases, the wireless communications system may experience asymmetric traffic loads for uplink and downlink transmissions. The wireless communications system may implement techniques for managing such traffic loads (e.g., for efficiently handling scenarios where uplink transmissions are relatively dense) . Such techniques may include the deployment of the one or more transmission reception points (TRPs) that are configured to support uplink transmissions from one or multiple UEs but not used for downlink transmissions to the UEs (e.g., uplink TRPs or otherwise “uplink only” TRPs) .
[0065] In some cases, a UE may encounter various challenges for beam management related to uplink TRPs, including power expenditure, overhead, coverage, latency, and interference challenges, for beam management and prediction. For example, to identify appropriate (e.g., best, most accurate, or highest performance) uplink transmission (Tx) and uplink receive (Rx) beams associated with the uplink TRPs, the UE may be scheduled with frequently transmitted sounding reference signals (SRSs) (e.g., SRS transmissions that have a periodicity that exceeds a threshold) so that the UE may sweep uplink Tx beams, while the uplink TRPs may use various uplink Rx beams to receive such SRSs.
[0066] In some cases, however, such beam sweeping techniques may be inefficient for the wireless system. For example, because no downlink reference signal may be transmitted by an uplink TRP for beam correspondence, the UE may implement relatively frequent transmission of SRSs, which may consume significant amounts of uplink transmission power for the UE and may consume correspondingly significant amount of uplink overhead for related cells or TRPs, which may perform frequent autonomous uplink Rx beam tracking.
[0067] In some wireless communications systems, a UE may use a prediction model such as an artificial intelligence (AI) / machine learning (ML) model to predict one or more parameters or characteristics associated with wireless communications. In such cases, the UE may measure respective beams that corresponds to one or more reference signals (e.g., synchronization signal blocks (SSBs) , channel state information reference signals (CSI-RSs) ) via a first set of resources, which may be referred to as “Set B” beams, during a first set of measurement occasions. Additionally, the UE may perform beam prediction for a set of beams associated with a second set of resources, which may be referred to as “Set A” beams, using an AI / ML model, and based on model training measurement results of the Set B beams and / or historical measurement results of the Set B beams.
[0068] Training of the prediction model may be performed using a set of training beam measurements that are associated with a model identification of the prediction model, which may correspond to one or more an associated identifiers (IDs) . Further, model inference for predicted measurements of beams and model performance monitoring may also be performed based on the associated ID (s) corresponding to the model. However, efficient signaling for performing model training, inference, and performance monitoring has not been defined. Further, it may be desirable in some cases to perform beam prediction at a UE for uplink-receive beam measurements at a network entity, which may be useful when beam reciprocity may not be present for one or more beams (e.g., when a relatively strong downlink received signal strength of a downlink beam has a relatively weak uplink received signal strength for a corresponding uplink beam) , or in a scenario where the UE is communicating with an uplink TRP (e.g., an “uplink only” TRP) .
[0069] In some cases, a UE may support various techniques to support efficient AI / ML model implementation, including training data collection, inference frameworks, performance monitoring, and ensuring consistency of network-side additional conditions (e.g., conditions for performing beam prediction) , for network-side receive beam prediction by a UE using AI / ML model (s) . The AI / ML models may be trained based on measurements (e.g., reference signal receive power (RSRP) measurements, SRS-resource indicator (SRI) measurements, signal to interference and noise ratio (SINR) measurements) from uplink reference signals (e.g., SRSs) transmitted by a UE.
[0070] In some aspects, for training data collection, a UE may be scheduled with a quantity of SRS resource sets (without associated TCI-State or SpatialRelationInfo) . The UE may acquire measured channel characteristics as measured at the network entity with respect to the uplink reference signals via control signaling, such as via downlink control information (DCI) , one or more medium access control (MAC) control elements (CE) , and / or radio resource control (RRC) signaling. AI / ML algorithm (s) at the UE with respect to the associated ID that output predicted channel characteristics for the uplink-receive beams may be trained, based on measurements of SRS resource sets and the network-indicated measurements, to predict spatial and / or temporal domain characteristics associated uplink Tx beams that the UE may use for transmission of uplink signals (e.g., physical uplink shared channel (PUSCH) , physical uplink control channel (PUCCH) towards an uplink TRP.
[0071] In some aspects, for inference, the UE may use an AL / ML model to predict channel characteristics with respect to uplink Tx beams or different uplink TRPs. For example, after verifying that the associated IDs for training data collection and for inference are the same, the UE may perform AI / ML inferencing to determine a set of applicable uplink Tx beams at the UE (and associated uplink Rx beams at one or more uplink TRPs) . In some examples, the UE may determine or predict beam switching, switching between uplink TRPs, or both, using AI / ML inferencing. For example, the UE may identify a current or future switch between uplink TRPs or between uplink Rx beams, and may report an indication of the switch to the network.
[0072] In some aspects, for performance monitoring, while the prediction associated with the inference step is on-going, the network entity may further schedule additional uplink reference signals (e.g., SRSs) for performance monitoring. In some cases, performance monitoring may be performed at the network, where the network entity may compare UE reported prediction results against measurements on the uplink reference signals to determine further actions. In other cases, performance monitoring may be performed at the UE, and for UE-based performance monitoring the UE may receive network-signaled measurement results for the uplink reference signals, representing channel characteristics on corresponding uplink-receive beams and the UE may determine or recommend further AI / ML action (e.g., activation, deactivation, or a switch of models) based on such signaled measurement results.
[0073] Aspects of the present disclosure may be implemented to realize one or more potential advantages. For example, AI / ML techniques including AI / ML training, AI / ML inference, and performance monitoring in uplink TRP deployments may support improved consistency between model training procedures, model inference procedures, and model performance monitoring procedures, which may enable reliable and efficient use of AI / ML models to predict channel characteristics, which in turn may provide for enhanced throughput (e.g., based on improved beam prediction accuracy) , reduced latency (e.g., based on improved beam prediction accuracy) , enhanced communications reliability, reduced power consumption, and efficient use of communications resources, including SRS resources and associated signaling.
[0074] Aspects of the disclosure are initially described in the context of wireless communications systems. Aspects of the disclosure are further illustrated by and described with reference to various AI / ML implementations, AI / ML operation flows, AI / ML beam management procedures, AI / ML performance monitoring procedures, AI / ML parameter consistency flows, apparatus diagrams, system diagrams, and flowcharts that relate to model training techniques for channel characteristic prediction procedures.
[0075] FIG. 1 shows an example of a wireless communications system 100 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The wireless communications system 100 may include one or more devices, such as one or more network devices (e.g., network entities 105) , one or more UEs 115, and a core network 130. In some examples, the wireless communications system 100 may be a Long Term Evolution (LTE) network, an LTE-Advanced (LTE-A) network, an LTE-A Pro network, a New Radio (NR) network, or a network operating in accordance with other systems and radio technologies, including future systems and radio technologies not explicitly mentioned herein.
[0076] The network entities 105 may be dispersed throughout a geographic area to form the wireless communications system 100 and may include devices in different forms or having different capabilities. In various examples, a network entity 105 may be referred to as a network element, a mobility element, a radio access network (RAN) node, or network equipment, among other nomenclature. In some examples, network entities 105 and UEs 115 may wirelessly communicate via communication link (s) 125 (e.g., a radio frequency (RF) access link) . For example, a network entity 105 may support a coverage area 110 (e.g., a geographic coverage area) over which the UEs 115 and the network entity 105 may establish the communication link (s) 125. The coverage area 110 may be an example of a geographic area over which a network entity 105 and a UE 115 may support the communication of signals according to one or more radio access technologies (RATs) .
[0077] The UEs 115 may be dispersed throughout a coverage area 110 of the wireless communications system 100, and each UE 115 may be stationary, or mobile, or both at different times. The UEs 115 may be devices in different forms or having different capabilities. Some example UEs 115 are illustrated in FIG. 1. The UEs 115 described herein may be capable of supporting communications with various types of devices in the wireless communications system 100 (e.g., other wireless communication devices, including UEs 115 or network entities 105) , as shown in FIG. 1.
[0078] As described herein, a node of the wireless communications system 100, which may be referred to as a network node, or a wireless node, may be a network entity 105 (e.g., any network entity described herein) , a UE 115 (e.g., any UE described herein) , a network controller, an apparatus, a device, a computing system, one or more components, or another suitable processing entity configured to perform any of the techniques described herein. For example, a node may be a UE 115. As another example, a node may be a network entity 105. As another example, a first node may be configured to communicate with a second node or a third node. In one aspect of this example, the first node may be a UE 115, the second node may be a network entity 105, and the third node may be a UE 115. In another aspect of this example, the first node may be a UE 115, the second node may be a network entity 105, and the third node may be a network entity 105. In yet other aspects of this example, the first, second, and third nodes may be different relative to these examples. Similarly, reference to a UE 115, network entity 105, apparatus, device, computing system, or the like may include disclosure of the UE 115, network entity 105, apparatus, device, computing system, or the like being a node. For example, disclosure that a UE 115 is configured to receive information from a network entity 105 also discloses that a first node is configured to receive information from a second node.
[0079] In some examples, network entities 105 may communicate with a core network 130, or with one another, or both. For example, network entities 105 may communicate with the core network 130 via backhaul communication link (s) 120 (e.g., in accordance with an S1, N2, N3, or other interface protocol) . In some examples, network entities 105 may communicate with one another via backhaul communication link (s) 120 (e.g., in accordance with an X2, Xn, or other interface protocol) either directly (e.g., directly between network entities 105) or indirectly (e.g., via the core network 130) . In some examples, network entities 105 may communicate with one another via a midhaul communication link 162 (e.g., in accordance with a midhaul interface protocol) or a fronthaul communication link 168 (e.g., in accordance with a fronthaul interface protocol) , or any combination thereof. The backhaul communication link (s) 120, midhaul communication links 162, or fronthaul communication links 168 may be or include one or more wired links (e.g., an electrical link, an optical fiber link) or one or more wireless links (e.g., a radio link, a wireless optical link) , among other examples or various combinations thereof. A UE 115 may communicate with the core network 130 via a communication link 155.
[0080] One or more of the network entities 105 or network equipment described herein may include or may be referred to as a base station 140 (e.g., a base transceiver station, a radio base station, an NR base station, an access point, a radio transceiver, a NodeB, an eNodeB (eNB) , a next-generation NodeB or giga-NodeB (either of which may be referred to as a gNB) , a 5G NB, a next-generation eNB (ng-eNB) , a Home NodeB, a Home eNodeB, or other suitable terminology) . In some examples, a network entity 105 (e.g., a base station 140) may be implemented in an aggregated (e.g., monolithic, standalone) base station architecture, which may be configured to utilize a protocol stack that is physically or logically integrated within one network entity (e.g., a network entity 105 or a single RAN node, such as a base station 140) .
[0081] In some examples, a network entity 105 may be implemented in a disaggregated architecture (e.g., a disaggregated base station architecture, a disaggregated RAN architecture) , which may be configured to utilize a protocol stack that is physically or logically distributed among multiple network entities (e.g., network entities 105) , such as an integrated access and backhaul (IAB) network, an open RAN (O-RAN) (e.g., a network configuration sponsored by the O-RAN Alliance) , or a virtualized RAN (vRAN) (e.g., a cloud RAN (C-RAN) ) . For example, a network entity 105 may include one or more of a central unit (CU) , such as a CU 160, a distributed unit (DU) , such as a DU 165, a radio unit (RU) , such as an RU 170, a RAN Intelligent Controller (RIC) , such as an RIC 175 (e.g., a Near-Real Time RIC (Near-RT RIC) , a Non-Real Time RIC (Non-RT RIC) ) , a Service Management and Orchestration (SMO) system, such as an SMO system 180, or any combination thereof. An RU 170 may also be referred to as a radio head, a smart radio head, a remote radio head (RRH) , a remote radio unit (RRU) , or a TRP. One or more components of the network entities 105 in a disaggregated RAN architecture may be co-located, or one or more components of the network entities 105 may be located in distributed locations (e.g., separate physical locations) . In some examples, one or more of the network entities 105 of a disaggregated RAN architecture may be implemented as virtual units (e.g., a virtual CU (VCU) , a virtual DU (VDU) , a virtual RU (VRU) ) .
[0082] The split of functionality between a CU 160, a DU 165, and an RU 170 is flexible and may support different functionalities depending on which functions (e.g., network layer functions, protocol layer functions, baseband functions, RF functions, or any combinations thereof) are performed at a CU 160, a DU 165, or an RU 170. For example, a functional split of a protocol stack may be employed between a CU 160 and a DU 165 such that the CU 160 may support one or more layers of the protocol stack and the DU 165 may support one or more different layers of the protocol stack. In some examples, the CU 160 may host upper protocol layer (e.g., layer 3 (L3) , layer 2 (L2) ) functionality and signaling (e.g., Radio Resource Control (RRC) , service data adaptation protocol (SDAP) , Packet Data Convergence Protocol (PDCP) ) . The CU 160 (e.g., one or more CUs) may be connected to a DU 165 (e.g., one or more DUs) or an RU 170 (e.g., one or more RUs) , or some combination thereof, and the DUs 165, RUs 170, or both may host lower protocol layers, such as layer 1 (L1) (e.g., physical (PHY) layer) or L2 (e.g., radio link control (RLC) layer, medium access control (MAC) layer) functionality and signaling, and may each be at least partially controlled by the CU 160. Additionally, or alternatively, a functional split of the protocol stack may be employed between a DU 165 and an RU 170 such that the DU 165 may support one or more layers of the protocol stack and the RU 170 may support one or more different layers of the protocol stack. The DU 165 may support one or multiple different cells (e.g., via one or multiple different RUs, such as an RU 170) . In some cases, a functional split between a CU 160 and a DU 165 or between a DU 165 and an RU 170 may be within a protocol layer (e.g., some functions for a protocol layer may be performed by one of a CU 160, a DU 165, or an RU 170, while other functions of the protocol layer are performed by a different one of the CU 160, the DU 165, or the RU 170) . A CU 160 may be functionally split further into CU control plane (CU-CP) and CU user plane (CU-UP) functions. A CU 160 may be connected to a DU 165 via a midhaul communication link 162 (e.g., F1, F1-c, F1-u) , and a DU 165 may be connected to an RU 170 via a fronthaul communication link 168 (e.g., open fronthaul (FH) interface) . In some examples, a midhaul communication link 162 or a fronthaul communication link 168 may be implemented in accordance with an interface (e.g., a channel) between layers of a protocol stack supported by respective network entities (e.g., one or more of the network entities 105) that are in communication via such communication links.
[0083] In some wireless communications systems (e.g., the wireless communications system 100) , infrastructure and spectral resources for radio access may support wireless backhaul link capabilities to supplement wired backhaul connections, providing an IAB network architecture (e.g., to a core network 130) . In some cases, in an IAB network, one or more of the network entities 105 (e.g., network entities 105 or IAB node (s) 104) may be partially controlled by each other. The IAB node (s) 104 may be referred to as a donor entity or an IAB donor. A DU 165 or an RU 170 may be partially controlled by a CU 160 associated with a network entity 105 or base station 140 (such as a donor network entity or a donor base station) . The one or more donor entities (e.g., IAB donors) may be in communication with one or more additional devices (e.g., IAB node (s) 104) via supported access and backhaul links (e.g., backhaul communication link (s) 120) . IAB node (s) 104 may include an IAB mobile termination (IAB-MT) controlled (e.g., scheduled) by one or more DUs (e.g., DUs 165) of a coupled IAB donor. An IAB-MT may be equipped with an independent set of antennas for relay of communications with UEs 115 or may share the same antennas (e.g., of an RU 170) of IAB node (s) 104 used for access via the DU 165 of the IAB node (s) 104 (e.g., referred to as virtual IAB-MT (vIAB-MT) ) . In some examples, the IAB node (s) 104 may include one or more DUs (e.g., DUs 165) that support communication links with additional entities (e.g., IAB node (s) 104, UEs 115) within the relay chain or configuration of the access network (e.g., downstream) . In such cases, one or more components of the disaggregated RAN architecture (e.g., the IAB node (s) 104 or components of the IAB node (s) 104) may be configured to operate according to the techniques described herein.
[0084] In the case of the techniques described herein applied in the context of a disaggregated RAN architecture, one or more components of the disaggregated RAN architecture may be configured to support model training techniques for channel characteristic prediction procedures as described herein. For example, some operations described as being performed by a UE 115 or a network entity 105 (e.g., a base station 140) may additionally, or alternatively, be performed by one or more components of the disaggregated RAN architecture (e.g., components such as an IAB node, a DU 165, a CU 160, an RU 170, an RIC 175, an SMO system 180) .
[0085] A UE 115 may include or may be referred to as a mobile device, a wireless device, a remote device, a handheld device, or a subscriber device, or some other suitable terminology, where the “device” may also be referred to as a unit, a station, a terminal, or a client, among other examples. A UE 115 may also include or may be referred to as a personal electronic device such as a cellular phone, a personal digital assistant (PDA) , a tablet computer, a laptop computer, or a personal computer. In some examples, a UE 115 may include or be referred to as a wireless local loop (WLL) station, an Internet of Things (IoT) device, an Internet of Everything (IoE) device, or a machine type communications (MTC) device, among other examples, which may be implemented in various objects such as appliances, vehicles, or meters, among other examples.
[0086] The UEs 115 described herein may be able to communicate with various types of devices, such as UEs 115 that may sometimes operate as relays, as well as the network entities 105 and the network equipment including macro eNBs or gNBs, small cell eNBs or gNBs, or relay base stations, among other examples, as shown in FIG. 1.
[0087] The UEs 115 and the network entities 105 may wirelessly communicate with one another via the communication link (s) 125 (e.g., one or more access links) using resources associated with one or more carriers. The term “carrier” may refer to a set of RF spectrum resources having a defined PHY layer structure for supporting the communication link (s) 125. For example, a carrier used for the communication link (s) 125 may include a portion of an RF spectrum band (e.g., a bandwidth part (BWP) ) that is operated according to one or more PHY layer channels for a given RAT (e.g., LTE, LTE-A, LTE-A Pro, NR) . Each PHY layer channel may carry acquisition signaling (e.g., synchronization signals, system information) , control signaling that coordinates operation for the carrier, user data, or other signaling. The wireless communications system 100 may support communication with a UE 115 using carrier aggregation or multi-carrier operation. A UE 115 may be configured with multiple downlink component carriers and one or more uplink component carriers according to a carrier aggregation configuration. Carrier aggregation may be used with both frequency division duplexing (FDD) and time division duplexing (TDD) component carriers. Communication between a network entity 105 and other devices may refer to communication between the devices and any portion (e.g., entity, sub-entity) of a network entity 105. For example, the terms “transmitting, ” “receiving, ” or “communicating, ” when referring to a network entity 105, may refer to any portion of a network entity 105 (e.g., a base station 140, a CU 160, a DU 165, a RU 170) of a RAN communicating with another device (e.g., directly or via one or more other network entities, such as one or more of the network entities 105) .
[0088] Signal waveforms transmitted via a carrier may be made up of multiple subcarriers (e.g., using multi-carrier modulation (MCM) techniques such as orthogonal frequency division multiplexing (OFDM) or discrete Fourier transform spread OFDM (DFT-S-OFDM) ) . In a system employing MCM techniques, a resource element may refer to resources of one symbol period (e.g., a duration of one modulation symbol) and one subcarrier, in which case the symbol period and subcarrier spacing may be inversely related. The quantity of bits carried by each resource element may depend on the modulation scheme (e.g., the order of the modulation scheme, the coding rate of the modulation scheme, or both) , such that a relatively higher quantity of resource elements (e.g., in a transmission duration) and a relatively higher order of a modulation scheme may correspond to a relatively higher rate of communication. A wireless communications resource may refer to a combination of an RF spectrum resource, a time resource, and a spatial resource (e.g., a spatial layer, a beam) , and the use of multiple spatial resources may increase the data rate or data integrity for communications with a UE 115.
[0089] The time intervals for the network entities 105 or the UEs 115 may be expressed in multiples of a basic time unit which may, for example, refer to a sampling period of Ts=1 / (Δfmax·Nf) seconds, for which Δfmax may represent a supported subcarrier spacing, and Nf may represent a supported discrete Fourier transform (DFT) size. Time intervals of a communications resource may be organized according to radio frames each having a specified duration (e.g., 10 milliseconds (ms) ) . Each radio frame may be identified by a system frame number (SFN) (e.g., ranging from 0 to 1023) .
[0090] Each frame may include multiple consecutively-numbered subframes or slots, and each subframe or slot may have the same duration. In some examples, a frame may be divided (e.g., in the time domain) into subframes, and each subframe may be further divided into a quantity of slots. Alternatively, each frame may include a variable quantity of slots, and the quantity of slots may depend on subcarrier spacing. Each slot may include a quantity of symbol periods (e.g., depending on the length of the cyclic prefix prepended to each symbol period) . In some wireless communications systems, such as the wireless communications system 100, a slot may further be divided into multiple mini-slots associated with one or more symbols. Excluding the cyclic prefix, each symbol period may be associated with one or more (e.g., Nf) sampling periods. The duration of a symbol period may depend on the subcarrier spacing or frequency band of operation.
[0091] A subframe, a slot, a mini-slot, or a symbol may be the smallest scheduling unit (e.g., in the time domain) of the wireless communications system 100 and may be referred to as a transmission time interval (TTI) . In some examples, the TTI duration (e.g., a quantity of symbol periods in a TTI) may be variable. Additionally, or alternatively, the smallest scheduling unit of the wireless communications system 100 may be dynamically selected (e.g., in bursts of shortened TTIs (sTTIs) ) .
[0092] Physical channels may be multiplexed for communication using a carrier according to various techniques. A physical control channel and a physical data channel may be multiplexed for signaling via a downlink carrier, for example, using one or more of time division multiplexing (TDM) techniques, frequency division multiplexing (FDM) techniques, or hybrid TDM-FDM techniques. A control region (e.g., a control resource set (CORESET) ) for a physical control channel may be defined by a set of symbol periods and may extend across the system bandwidth or a subset of the system bandwidth of the carrier. One or more control regions (e.g., CORESETs) may be configured for a set of the UEs 115. For example, one or more of the UEs 115 may monitor or search control regions for control information according to one or more search space sets, and each search space set may include one or multiple control channel candidates in one or more aggregation levels arranged in a cascaded manner. An aggregation level for a control channel candidate may refer to an amount of control channel resources (e.g., control channel elements (CCEs) ) associated with encoded information for a control information format having a given payload size. Search space sets may include common search space sets configured for sending control information to UEs 115 (e.g., one or more UEs) or may include UE-specific search space sets for sending control information to a UE 115 (e.g., a specific UE) .
[0093] A network entity 105 may provide communication coverage via one or more cells, for example, a macro cell, a small cell, a hot spot, or other types of cells, or any combination thereof. The term “cell” may refer to a logical communication entity used for communication with a network entity 105 (e.g., using a carrier) and may be associated with an identifier for distinguishing neighboring cells (e.g., a physical cell identifier (PCID) , a virtual cell identifier (VCID) ) . In some examples, a cell also may refer to a coverage area 110 or a portion of a coverage area 110 (e.g., a sector) over which the logical communication entity operates. Such cells may range from smaller areas (e.g., a structure, a subset of structure) to larger areas depending on various factors such as the capabilities of the network entity 105. For example, a cell may be or include a building, a subset of a building, or exterior spaces between or overlapping with coverage areas 110, among other examples.
[0094] In some examples, a network entity 105 (e.g., a base station 140, an RU 170) may be movable and therefore provide communication coverage for a moving coverage area, such as the coverage area 110. In some examples, coverage areas 110 (e.g., different coverage areas) associated with different technologies may overlap, but the coverage areas 110 (e.g., different coverage areas) may be supported by the same network entity (e.g., a network entity 105) . In some other examples, overlapping coverage areas, such as a coverage area 110, associated with different technologies may be supported by different network entities (e.g., the network entities 105) . The wireless communications system 100 may include, for example, a heterogeneous network in which different types of the network entities 105 support communications for coverage areas 110 (e.g., different coverage areas) using the same or different RATs.
[0095] The wireless communications system 100 may be configured to support ultra-reliable communications or low-latency communications, or various combinations thereof. For example, the wireless communications system 100 may be configured to support ultra-reliable low-latency communications (URLLC) . The UEs 115 may be designed to support ultra-reliable, low-latency, or critical functions. Ultra-reliable communications may include private communication or group communication and may be supported by one or more services such as push-to-talk, video, or data. Support for ultra-reliable, low-latency functions may include prioritization of services, and such services may be used for public safety or general commercial applications. The terms ultra-reliable, low-latency, and ultra-reliable low-latency may be used interchangeably herein.
[0096] In some examples, a UE 115 may be configured to support communicating directly with other UEs (e.g., one or more of the UEs 115) via a device-to-device (D2D) communication link, such as a D2D communication link 135 (e.g., in accordance with a peer-to-peer (P2P) , D2D, or sidelink protocol) . In some examples, one or more UEs 115 of a group that are performing D2D communications may be within the coverage area 110 of a network entity 105 (e.g., a base station 140, an RU 170) , which may support aspects of such D2D communications being configured by (e.g., scheduled by) the network entity 105. In some examples, one or more UEs 115 of such a group may be outside the coverage area 110 of a network entity 105 or may be otherwise unable to or not configured to receive transmissions from a network entity 105. In some examples, groups of the UEs 115 communicating via D2D communications may support a one-to-many (1: M) system in which each UE 115 transmits to one or more of the UEs 115 in the group. In some examples, a network entity 105 may facilitate the scheduling of resources for D2D communications. In some other examples, D2D communications may be carried out between the UEs 115 without an involvement of a network entity 105.
[0097] The core network 130 may provide user authentication, access authorization, tracking, Internet Protocol (IP) connectivity, and other access, routing, or mobility functions. The core network 130 may be an evolved packet core (EPC) or 5G core (5GC) , which may include at least one control plane entity that manages access and mobility (e.g., a mobility management entity (MME) , an access and mobility management function (AMF) ) and at least one user plane entity that routes packets or interconnects to external networks (e.g., a serving gateway (S-GW) , a Packet Data Network (PDN) gateway (P-GW) , or a user plane function (UPF) ) . The control plane entity may manage non-access stratum (NAS) functions such as mobility, authentication, and bearer management for the UEs 115 served by the network entities 105 (e.g., base stations 140) associated with the core network 130. User IP packets may be transferred through the user plane entity, which may provide IP address allocation as well as other functions. The user plane entity may be connected to IP services 150 for one or more network operators. The IP services 150 may include access to the Internet, Intranet (s) , an IP Multimedia Subsystem (IMS) , or a Packet-Switched Streaming Service.
[0098] The wireless communications system 100 may operate using one or more frequency bands, which may be in the range of 300 megahertz (MHz) to 300 gigahertz (GHz) . Generally, the region from 300 MHz to 3 GHz is known as the ultra-high frequency (UHF) region or decimeter band because the wavelengths range from approximately one decimeter to one meter in length. UHF waves may be blocked or redirected by buildings and environmental features, which may be referred to as clusters, but the waves may penetrate structures sufficiently for a macro cell to provide service to the UEs 115 located indoors. Communications using UHF waves may be associated with smaller antennas and shorter ranges (e.g., less than one hundred kilometers) compared to communications using the smaller frequencies and longer waves of the high frequency (HF) or very high frequency (VHF) portion of the spectrum below 300 MHz.
[0099] The wireless communications system 100 may utilize both licensed and unlicensed RF spectrum bands. For example, the wireless communications system 100 may employ License Assisted Access (LAA) , LTE-Unlicensed (LTE-U) RAT, or NR technology using an unlicensed band such as the 5 GHz industrial, scientific, and medical (ISM) band. While operating using unlicensed RF spectrum bands, devices such as the network entities 105 and the UEs 115 may employ carrier sensing for collision detection and avoidance. In some examples, operations using unlicensed bands may be based on a carrier aggregation configuration in conjunction with component carriers operating using a licensed band (e.g., LAA) . Operations using unlicensed spectrum may include downlink transmissions, uplink transmissions, P2P transmissions, or D2D transmissions, among other examples.
[0100] A network entity 105 (e.g., a base station 140, an RU 170) or a UE 115 may be equipped with multiple antennas, which may be used to employ techniques such as transmit diversity, receive diversity, multiple-input multiple-output (MIMO) communications, or beamforming. The antennas of a network entity 105 or a UE 115 may be located within one or more antenna arrays or antenna panels, which may support MIMO operations or transmit or receive beamforming. For example, one or more base station antennas or antenna arrays may be co-located at an antenna assembly, such as an antenna tower. In some examples, antennas or antenna arrays associated with a network entity 105 may be located at diverse geographic locations. A network entity 105 may include an antenna array with a set of rows and columns of antenna ports that the network entity 105 may use to support beamforming of communications with a UE 115. Likewise, a UE 115 may include one or more antenna arrays that may support various MIMO or beamforming operations. Additionally, or alternatively, an antenna panel may support RF beamforming for a signal transmitted via an antenna port.
[0101] Beamforming, which may also be referred to as spatial filtering, directional transmission, or directional reception, is a signal processing technique that may be used at a transmitting device or a receiving device (e.g., a network entity 105, a UE 115) to shape or steer an antenna beam (e.g., a transmit beam, a receive beam) along a spatial path between the transmitting device and the receiving device. Beamforming may be achieved by combining the signals communicated via antenna elements of an antenna array such that some signals propagating along particular orientations with respect to an antenna array experience constructive interference while others experience destructive interference. The adjustment of signals communicated via the antenna elements may include a transmitting device or a receiving device applying amplitude offsets, phase offsets, or both to signals carried via the antenna elements associated with the device. The adjustments associated with each of the antenna elements may be defined by a beamforming weight set associated with a particular orientation (e.g., with respect to the antenna array of the transmitting device or receiving device, or with respect to some other orientation) .
[0102] A network entity 105 or a UE 115 may use beam sweeping techniques as part of beamforming operations. For example, a network entity 105 (e.g., a base station 140, an RU 170) may use multiple antennas or antenna arrays (e.g., antenna panels) to conduct beamforming operations for directional communications with a UE 115. Some signals (e.g., synchronization signals, reference signals, beam selection signals, or other control signals) may be transmitted by a network entity 105 multiple times along different directions. For example, the network entity 105 may transmit a signal according to different beamforming weight sets associated with different directions of transmission. Transmissions along different beam directions may be used to identify (e.g., by a transmitting device, such as a network entity 105, or by a receiving device, such as a UE 115) a beam direction for later transmission or reception by the network entity 105.
[0103] Some signals, such as data signals associated with a particular receiving device, may be transmitted by a transmitting device (e.g., a network entity 105 or a UE 115) along a single beam direction (e.g., a direction associated with the receiving device, such as another network entity 105 or UE 115) . In some examples, the beam direction associated with transmissions along a single beam direction may be determined based on a signal that was transmitted along one or more beam directions. For example, a UE 115 may receive one or more of the signals transmitted by the network entity 105 along different directions and may report to the network entity 105 an indication of the signal that the UE 115 received with a highest signal quality or an otherwise acceptable signal quality.
[0104] In some examples, transmissions by a device (e.g., by a network entity 105 or a UE 115) may be performed using multiple beam directions, and the device may use a combination of digital precoding or beamforming to generate a combined beam for transmission (e.g., from a network entity 105 to a UE 115) . The UE 115 may report feedback that indicates precoding weights for one or more beam directions, and the feedback may correspond to a configured set of beams across a system bandwidth or one or more sub-bands. The network entity 105 may transmit a reference signal (e.g., a cell-specific reference signal (CRS) , a CSI-RS) , which may be precoded or unprecoded. The UE 115 may provide feedback for beam selection, which may be a precoding matrix indicator (PMI) or codebook-based feedback (e.g., a multi-panel type codebook, a linear combination type codebook, a port selection type codebook) . Although these techniques are described with reference to signals transmitted along one or more directions by a network entity 105 (e.g., a base station 140, an RU 170) , a UE 115 may employ similar techniques for transmitting signals multiple times along different directions (e.g., for identifying a beam direction for subsequent transmission or reception by the UE 115) or for transmitting a signal along a single direction (e.g., for transmitting data to a receiving device) .
[0105] A receiving device (e.g., a UE 115) may perform reception operations in accordance with multiple receive configurations (e.g., directional listening) when receiving various signals from a transmitting device (e.g., a network entity 105) , such as synchronization signals, reference signals, beam selection signals, or other control signals. For example, a receiving device may perform reception in accordance with multiple receive directions by receiving via different antenna subarrays, by processing received signals according to different antenna subarrays, by receiving according to different receive beamforming weight sets (e.g., different directional listening weight sets) applied to signals received at multiple antenna elements of an antenna array, or by processing received signals according to different receive beamforming weight sets applied to signals received at multiple antenna elements of an antenna array, any of which may be referred to as “listening” according to different receive configurations or receive directions. In some examples, a receiving device may use a single receive configuration to receive along a single beam direction (e.g., when receiving a data signal) . The single receive configuration may be aligned along a beam direction determined based on listening according to different receive configuration directions (e.g., a beam direction determined to have a highest signal strength, highest signal-to-noise ratio (SNR) , or otherwise acceptable signal quality based on listening according to multiple beam directions) .
[0106] In some examples, a UE 115 may support AI and / or ML models and / or functionalities, which the UE 115 may use to perform various wireless communications procedures (e.g., CSI prediction, beam selection, and / or beam prediction, among other examples) . In some examples, an AI and / or ML model may be referred to as a prediction model. In such cases, the UE 115 may generate inference data using one or more AI / ML models / functionalities. Additionally, or alternatively, the UE 115 may perform life cycle management (LCM) operations for a given prediction model (e.g., AI / ML model and / or functionality) , which may include, for example, model or functionality selection, activation, deactivation, switching, and fallback, among other examples, and may be based on one or more AI / ML models / functionalities. In some aspects, LCM may be model-based or functionality-based LCM procedures. As described herein, an AI functionality or AI model may be referred to as an ML functionality or ML model, or vice versa, and the term “prediction model” may be used to generally refer to models based on AI and / or ML functionality. That is, the terms “artificial intelligence” “machine learning, ” and “prediction model” may, in some examples, be used interchangeably to refer to similar technologies, models, functions, algorithms, or any combination thereof. In some cases, the terms “model” and “functionality” may be used interchangeably. In some examples, ML operations may be considered a subset of AI operations. In any case, aspects of the features described herein may be referred to as AI functionalities, AI functions, AI models, AI services, AI operations, or the like, and such features may be similarly applicable to ML functionalities, ML functions, ML models, ML services, ML operations, or any combination thereof. Thus, reference to “machine learning” or “artificial intelligence” may refer to ML, AI, or both, and the terms “artificial intelligence” or “machine learning” should not be considered limiting to the scope of the claims or the disclosure.
[0107] A quasi co-location (QCL) relationship between one or more transmissions or signals may refer to a relationship between the antenna ports (and the corresponding signaling beams) of the respective transmissions. For example, one or more antenna ports may be implemented by a network entity 105 for transmitting at least one or more reference signals (such as a downlink reference signal, a synchronization signal block (SSB) , or the like) and control information transmissions to a UE 115. However, the channel properties of signals sent via the different antenna ports may be interpreted (e.g., by a receiving device) to be the same (e.g., despite the signals being transmitted from different antenna ports) , and the antenna ports (and the respective beams) may be described as being quasi co-located (QCLed) . QCLed signals may enable the UE 115 to derive the properties of a first signal (e.g., delay spread, Doppler spread, frequency shift, average power) transmitted via a first antenna port from measurements made on a second signal transmitted via a second antenna port. Put another way, if two antenna ports are categorized as being QCLed in terms of, for example, delay spread then the UE 115 may determine the delay spread for one antenna port (e.g., based on a received reference signal, such as CSI-RS) and then apply the result to both antenna ports. Such techniques may avoid the UE 115 determining the delay spread separately for each antenna port. In some cases, two antenna ports may be said to be spatially QCLed, and the properties of a signal sent over a directional beam may be derived from the properties of a different signal over another, different directional beam. That is, QCL relationships may relate to beam information for respective directional beams used for communications of various signals.
[0108] Different types of QCL relationships may describe the relationship between two different signals or antenna ports. For instance, QCL-TypeA may refer to a QCL relationship between signals including Doppler shift, Doppler spread, average delay, and delay spread. QCL-TypeB may refer to a QCL relationship including Doppler shift and Doppler spread, whereas QCL-TypeC may refer to a QCL relationship including Doppler shift and average delay. A QCL-TypeD may refer to a QCL relationship of spatial parameters, which may indicate a relationship between two or more directional beams used to communicate signals. Here, the spatial parameters may indicate that a first beam used to transmit a first signal may be similar (or the same) as another beam used to transmit a second, different, signal, or, that the same receive beam may be used to receive both the first and the second signal. Thus, the beam information for various beams may be derived through receiving signals from a transmitting device, where, in some cases, the QCL information or spatial information may help a receiving device efficient identify communications beams (e.g., without having to sweep through a large quantity of beams to identify a beam (e.g., the beam having a highest signal quality) ) . In addition, QCL relationships may exist for both uplink and downlink transmissions and, in some cases, a QCL relationship may also be referred to as spatial relationship information.
[0109] In some examples, TCI states may include one or more parameters associated with a QCL relationship between transmitted signals. For example, each TCI state includes parameters for configuring a QCL relationship between one or two downlink reference signals and the DMRS ports of PDSCH, the DMRS port of PDCCH or the CSI-RS port (s) of a CSI-RS resource. The QCL relationship is configured by a first higher layer parameter for the first downlink reference signal, and by a second higher layer parameter for the second downlink reference signal (if configured) . That is, a network entity 105 may configure one or more QCL relationships that provides a mapping between a reference signal and antenna ports of another signal, and a corresponding TCI state may be indicated to the UE 115 by the network entity 105. In some cases, a set of TCI states (e.g., a list of TCI states) may be indicated to a UE 115 via RRC signaling, where some quantity of TCI states may be configured via RRC and one or more TCI states may be indicated (e.g., activated) via a medium access control (MAC) -control element (MAC-CE) , and further indicated via DCI (e.g., within a CORESET) . The QCL relationship associated with the TCI state (and further established through higher-layer parameters) may provide the UE 115 with the QCL relationship for respective antenna ports and reference signals transmitted by the network entity 105.
[0110] In some cases, a unified TCI state framework may be utilized in the wireless communications system 100, where a joint TCI state may be indicated (e.g., via a codepoint in DCI) for multiple channels and / or signals, and a joint TCI (e.g., joint downlink / uplink TCI state) may be associated with a single TCI state identifier for a codepoint. Here, one or more joint TCI states may be configured via RRC signaling, and the TCI state may be jointly applied for multiple uplink and / or downlink channels. The use of joint TCI states may enable a network to efficiently use TCI states that may be intended for use with both uplink and downlink transmissions (e.g., corresponding to respective beams for uplink and downlink) .
[0111] An SRS may be transmitted by a UE 115 using a predetermined sequence (e.g., a Zadoff-Chu sequence) so that a network entity 105 may estimate the uplink channel quality. An SRS transmission may not be associated with transmission of data on another channel, and may be transmitted periodically on a wide bandwidth (e.g., a bandwidth including more subcarriers than are allocated for uplink data transmission) . In some examples, an SRS may be scheduled on multiple antenna ports and still considered to be a single SRS transmission. An SRS transmission may be categorized as a Type 0 (periodically transmitted at equally spaced intervals) SRS or as a Type 1 (aperiodic) SRS. In either case, the network entity 105 may control the timing of SRS transmissions by notifying the UE 115 of which TTIs (e.g., subframes) may support the transmission of the SRS. Additionally, a sounding period (e.g., 2 to 230 subframes) and an offset within the sounding period may be configured for the UE 115. As a result, the UE 115 may transmit the SRS when a subframe that supports SRS transmissions coincides with the configured sounding period. In some cases, the SRS may be transmitted during a temporally last OFDM symbol of the subframe or, in some cases, may be sent during an uplink portion of a special subframe. Data gathered by a network entity 105 from an SRS may be used to inform the scheduling of uplink transmissions by the UE 115, such as frequency dependent transmissions. A network entity 105 may also utilize an SRS to check timing alignment status and send time alignment commands to the UE 115.
[0112] In some examples, one or more SRS resource sets may be configured via higher-layer signaling (e.g., RRC signaling comprising an SRS-ResourceSet information element) , and an applicability of an SRS resource set may be configured using a parameter (e.g., a usage parameter) included in the configuration. As an example, the usage associated with an SRS resource set configuration may be configured as beam management, codebook, non-codebook, or antenna switching. Each SRS resource set may be configured with one or more (e.g., up to 16) SRS resources, and each SRS resource set may be associated with aperiodic, semi-persistent, or periodic signaling. The periodic SRS resources may be configured via RRC signaling, whereas semi-persistent SRS may be configured via RRC and activated via MAC-CE signaling that includes an indication of an SRS resource set ID.
[0113] Some wireless communications systems may experience a relatively dense deployment of wireless communications devices, including multiple UEs 115 within a same location (e.g., within a same coverage area, cell, geographic location, or the like) . In one example, dense deployments may correspond to a relatively large event, such as a sporting event, a concert, or other types of events and situations in which a relatively large quantity of UEs 115 may be in a same area, region, and / or location. In such cases, the wireless communications system may experience asymmetric traffic loads for uplink and downlink transmissions. The wireless communications system may implement techniques for managing such traffic loads (e.g., for efficiently handling scenarios where uplink transmissions are relatively dense) . Such techniques may include the deployment of the one or more TRPs that are configured to support uplink transmissions from one or multiple UEs 115, but not used for downlink transmissions to the UEs 115. The one or more TRPs configured for uplink transmissions (but not downlink transmissions) may be referred to as uplink-only TRPs, or some similar terminology.
[0114] Thus, a UE 115 may receive downlink transmissions from a network entity 105 via a downlink communication link. The UE 115 may also communicate in the uplink with the network entity 105. However, in relatively dense deployment scenarios (e.g., where a relatively high quantity of UEs 115 are communicating in the wireless communications system) , the UE 115 may communicate in the uplink with the one or more TRPs, for example, via an uplink communication link. Accordingly, the network entity 105 may receive or otherwise obtain uplink signaling from one or more UEs 115 (including the UE 115) via one or more TRPs (via the uplink-only TRPs) . That is, the network entity 105 may support both uplink and downlink communication via collocated (or approximately collocated) antenna panels and may use the one or more TRPs to supplement the uplink coverage or capacity provided by the network entity 105 (e.g., to receive signals from the UE 115 via an uplink-only TRP) . Here, the network entity 105 may be collocated with one or more downlink TRPs and may additionally include uplink reception capabilities via one or more antenna panels, and the network entity 105 may control or otherwise be associated with the (e.g., non-collocated) TRPs (e.g., uplink-only TRPs) to supplement uplink coverage and capacity. The TRPs may be communicatively coupled with the network entity 105 via a backhaul link, which may be one or more wired or wireless links.
[0115] A wireless communications system may be configured to improve coverage and / or capacity for uplink communications and may be associated with an asymmetric downlink / uplink densification. By providing and using the one or more uplink-only TRPs, the network entity 105 may reduce uplink pathloss, which may be beneficial in scenarios in which uplink coverage is a bottleneck for uplink communications, as well as in terms of deployment cost and / or complexity (e.g., because the TRPs may not transmit any downlink signaling) . Instead, an uplink-only TRP may receive an uplink signal and / or uplink channel transmissions from a UE 115 and send (e.g., forward, relay, or transmit) the signal or channel (or information parsed or decoded from the signal or channel) to the network entity 105. A TRP may send the signal or channel (or information parsed or decoded therefrom) to the network entity 105 with complete or partial processing or without any processing.
[0116] A wireless communication system implementing such techniques may be an example of a heterogeneous network that supports deployments for improving uplink throughput. In such cases, the network entity 105 (e.g., a macro network entity) and one or more uplink-only TRPs (e.g., one or more micro nodes) may have different power ratings, and a UE 115 may receive downlink transmissions from the network entity and transmit uplink messages to either the network entity or non-co-located micro nodes (e.g., TRPs) to maximize uplink throughput. As a technique to further reduce energy consumption, the one or more TRPs may, for instance, reduce the use of or turn off various components and / or circuitry associated with downlink transmissions. To support such deployments, enhancements for uplink power control may be used. In some examples, when one or more reference signals (e.g., pathloss reference signals) are transmitted from the network entity 105 and the UE 115 transmits in the uplink to the TRPs, pathloss measured from the reference signals sent from the network entity may not accurate. Therefore, the UE 115 may be configured with a pathloss offset to facilitate accurate calculation of the pathloss associated with the micro nodes. Further, an additional SRS closed-loop power control for downlink CSI acquisition with the network entity 105 (e.g., for downlink transmissions) , separate from that for the SRS to the TRPs (for uplink multi-TRP reception) may be used. Therefore, the wireless communication system may support multiple closed-loop power control adjustment states for SRS, which may be separate from PUSCH. As an example, there may be two closed-loop power control adjustment states for SRS that are both separate from PUSCH, and there may be a configuration of one or more pathloss offset configurations for pathloss calculation to one or more uplink TRPs (e.g., uplink-only TRPs) , such as when pathloss reference signals are received from the network entity 105 (and / or one or more downlink TRPs) .
[0117] In some wireless communications systems, a UE 115, may perform beam management in which the UE 115 may establish a beam pair associated with threshold connectivity (e.g., a quality that satisfies a threshold) and may update the beam pair to maintain the threshold connectivity. Additionally, the UE 115 may perform beam prediction to support updating and / or reselection, of the beam pair to maintain the threshold connectivity. For example, the UE 115 may perform beam prediction in a time domain, a spatial domain, or both, to reduce overhead and latency, and to improve beam selection accuracy (e.g., as compared to not performing beam prediction) .
[0118] In some cases, the UE 115 may perform the beam prediction using one or more beam prediction (e.g., machine learning, artificial intelligence) models. For example, the UE 115 may perform spatial domain beam prediction (e.g., spatial domain downlink beam prediction) using a model (e.g., an AI / ML model / functionality, a prediction model) for a first beam set (e.g., Set A beams) associated with downlink beam prediction based on measurement results of a second beam set (e.g., Set B beams) associated with downlink beam measurement. Such beam prediction may, in some examples, be referred to as a first case of beam management (e.g., beam management (BM) -Case 1) . The second beam set may be a subgroup, or subset, of the first beam set. Alternatively, the first beam set may be different than the second beam set. For example, the first beam set may include or be associated with relatively narrow beams and the second beam set may include or be associated with relatively wide beams. In some implementations, the model may generate predictions for uplink beam reception characteristics at the network entity 105. For example, the model may generate RSRP or SINR predictions for receptions of uplink communications at the network entity 105.
[0119] Additionally, or alternatively, the UE 115 may perform temporal beam prediction (e.g., temporal or time-domain downlink beam prediction) using the model for the first beam set (e.g., Set A beams) based on measurement results (e.g., historic measurement results) of the second beam set (e.g., Set B beams) . Such beam predication may, in some examples, be referred to as a second case of beam management (e.g., BM-Case 2) . In any case, one or more signaling mechanisms may be used to facilitate LCM operation related to beam management cases, and there may be one or more techniques implemented to ensure consistency between training and inference regarding network-side additional conditions (e.g., when identified) for inference procedures by the UE 115. In some examples, a common framework may be implemented for both spatial-domain and temporal beam prediction.
[0120] Thus, one or more UEs 115 may use one or more AI / ML models / functionalities to perform beam prediction. In such cases, the UE 115 may measure respective beams that corresponds to one or more reference signals (e.g., SSBs, CSI-RSs) via a first set of resources, which may be referred to as Set B beams, during a first set of measurement occasions. Additionally, the UE 115 may perform beam prediction (e.g., inference, assumption) for a set of beams associated with a second set of resources, which may be referred to as Set A beams, using the AI / ML model / functionality, which may be based on historical measurement results of the Set B beams. That is, the UE 115 may use various measurements of Set B beams to predict one or more Set A beams. In such cases, the UE 115 may perform the beam prediction (e.g., temporal beam prediction, spatial beam prediction) using the AI / ML model / functionality in accordance with a set of parameters, including a Set B beam measurement window length, a Set B beam measurement periodicity, and a prediction duration for Set A beams, among other examples. The use of the AI / ML models / functionalities may be associated with a training of the AI / ML model / functionality (e.g., based on collected or known data) and one or more inference processes by which the AI / ML model / functionality uses training to analyze additional data and make one or more predictions.
[0121] In some cases, wireless communications systems may support functionality-based LCM operations or model-based (e.g., model ID-based) LCM operations for ML and / or AI-enabled processes and functions. LCM may refer to the use of AI and / or ML for operations that maintain one or more wireless communication links, such as CSI reporting (e.g., CSI prediction) , beam management operations (e.g., spatial beam prediction and / or temporal beam prediction) , positioning (e.g., AI and / or ML-assisted positioning) , among other examples. Functionality-based LCM operations may be associated with AI and / or ML-enabled features (e.g., AI / ML models / functionalities) enabled by configurations that are supported by a UE 115. Model-based LCM operations may be associated with specific configurations or conditions of an AI and / or ML model supported by the UE 115. In some aspects, the UE 115 and / or network entity 105 may support one or more enhancements related to AI / ML inference procedures. For example, the UE 115 and / or network entity 105 may support reporting enhancements for carrying spatial and / or temporal beam prediction results. Additionally, or alternatively, the UE 115 and / or network entity 105 may support Set A / Set B configurations for reporting of inference results. The UE 115 and / or network entity 105 may support one or more enhancements related to AI / ML performance monitoring. For example, the UE 115 and / or network entity 105 may support network-side performance monitoring and / or UE-assisted performance monitoring.
[0122] In some examples, there may need to be some consistency with additional conditions (e.g., network-side additional conditions) across both training and inference for AI / ML models / functionalities used by one or more UEs 115. The network-side additional conditions may include aspects related to the network that are transparent to one or more UEs 115 and may impact generalization capabilities of the UEs 115. An example of network-side additional conditions may include a network entity codebook, as some UE-side AI / ML models / functionalities for beam prediction may not generalize well across different network entity codebooks. Further examples of network-side additional conditions may include, a numbering of Set A / Set B beams, an ordering of Set A / Set B beams, an indexing of Set A / Set B beams, pointing directions (e.g., absolute pointing directions, relative pointing directions) for beamforming (e.g., with respect to a boresight direction relative to the center of an antenna panel) , one or more beam shapes (e.g., angular specific beamforming gains) , QCL relationships across / within Set A / Set B beams, temporal parameters (e.g., periodicity of Set A / Set B beams, target future occasions for temporal prediction) , among other examples. Consistency with network-side conditions may be needed to enable efficient communications between the UE 115 and the network entity.
[0123] In some examples, various techniques may be used to support the consistency of network-side additional condition across training and inference for UE-sided models for beam management BM-Case 1 and BM-Case 2, where the network-side additional conditions may at least impact UE assumptions on beams of Set A and / or Set B. As an example, there may be some parameter consistency that is guaranteed by the network for a same model ID that is used for data collection and inference (e.g., related to the Set A and Set B beams) . As an example, there may be a similar or same quantity of reference signals and / or resources (e.g., SSBs, CSI-RSs, other reference signals, resources) configured for the Set A and / or Set B beams across training and inference. In another example, a relative pointing direction and beamwidth between directional beams (e.g., with respect to different resources) may be the same (e.g., across SSB resource sets for training and inference, across a CSI-RS resources set for training and a prediction CSI-RS resource set for inference) . In some further examples, there may be consistency of QCL parameters, consistency of down-sampling patterns (e.g., for narrow beam-to-narrow beam predictions) , and / or consistency of temporal network-side additional conditions across training and inference, among other examples that enable consistency of additional conditions.
[0124] Additionally, or alternatively, the techniques for consistency with additional conditions (e.g., network-side additional conditions) may be based on an associated ID, where some information may be assumed by UE 115 with the same associated ID across training and inference. An “associated ID” may refer to some identifier that is indicative of one or more beam parameters (e.g., beam shapes, beam pointing angles, or other aspects associated with beamforming) . The associated ID may, in some examples, be referred to as a data set ID, a data configuration ID, or some similar terminology. In some cases, the associated ID may be interpreted as a dataset, a configuration, a scenario, a codebook, a functionality, a model identifier, or any combination thereof, where the associated ID indicates network-side additional conditions related to UE assumptions associated with LCM (e.g., AI / ML LCM, which may include data collection, training, deployment, inference, performance monitoring, activation, deactivation, switching, or any combination thereof) .
[0125] In some cases, different infrastructure vendors may use different associated IDs for different beam parameters, which may present challenges for a UE to determine the beam parameters across different vendors. In other examples, the associated IDs may be the same across some vendors. In any case, the UE 115 may benefit from receiving additional information (e.g., supplemental information) corresponding to an associated ID to help determine information about relevant beams used, for example, for beam prediction. For instance, the UE 115 may have one or more AI / ML models / functionalities that are trained using a set of associated IDs (e.g., for beam prediction or other operations) . As such, for inference operations, it may be important for the UE 115 to know whether an indicated associated ID corresponds to one of the associated IDs used to train an AI / ML model. In particular, as long as a same associated ID is identified across training and inference, corresponding network-side additional conditions may be assumed to be the same by across training and inference. In some examples, an associated ID may be indicate or be equivalent to a model ID. In some aspects, the UE 115 may determine that it may use a particular AI / ML model that was trained using a same associated ID that was indicated to the UE 115 for beam prediction operations.
[0126] In some examples, one or more associated IDs may be used, for example, within a CSI framework or outside of the CSI framework, among other examples. That is, for a UE-sided model in beam management, the associated ID may be supported, and the associated ID may at least be configured within a CSI framework. In some aspects, various techniques may be used for configuring / indicating the associated ID via one or more signal (s) and / or in other procedure (s) / framework (s) , which may correspond to whether / how the associated ID is configured / indicated using such techniques. In some examples, the UE 115 may assume similar properties of a downlink transmission beam or beam set / list associated with the same associated ID, where the similar properties of the downlink transmission beam or beam set / list may be defined in some way.
[0127] In some implementations, a UE 115 may provide a capability message that indicates one or more parameters or feature groups related to a capability of the UE 115 to perform beam prediction for uplink-receive beams at a network entity 105. Additionally, or alternatively, the capability message may indicate a UE 115 capability to perform UE-based or UE-assisted performance monitoring of AI / ML models. For example, the UE 115 may receive a capability enquiry message from the network entity 105 (e.g., a message including UECapabilityEnquiry) requesting one or more AI / ML-related capabilities of the UE 115. The network entity 105 may request information regarding AI / ML capabilities of the UE 115 (e.g., whether the UE 115 is capable of using AI / ML, one or more functions that the UE 115 is capable of performing using AI / ML) . The UE 115 may transmit capability information (e.g., a message including UECapabilityInformation) to the network entity 105. For example, the UE 115 may provide UE capability parameters associated with one or more feature groups for AI / ML operations / functions. In some cases, the UE 115 may indicate a set of feature groups and / or a set of parameters within each feature group (e.g., within each functionality) . The UE 115 may transmit the capability information in response to the capability enquiry message.
[0128] In some cases, the network entity 105 may transmit a parameter configuration to the UE 115, which may indicate one or more parameters associated with AI / ML model LCM, such as one or more associated IDs of one or more models, one or more beams associate with each model, or any combination thereof. In some aspects, the network entity 105 may transmit, to the UE 115, a control message including (or indicating) one or more sets of inference information (e.g., combinations of associated IDs and configurations for inference operations and / or parameters associated with the inference operations) corresponding to current network-side additional conditions of the network entity 105, future additional conditions of the network entity 105, and / or additional conditions associated with one or more other network entities 105. In some examples, the control message may be an RRC reconfiguration message that includes (or indicates) the one or more sets of inference information.
[0129] The network entity 105 may indicate one or more configurations (e.g., CSI-ReportConfig) that may be used for the purpose of enabling inference operation for beam prediction and / or the one or more parameters. In some examples, the one or more configurations and / or the one or more parameters may include the associated ID. In some cases, the network entity 105 may provide one or more other configurations to the UE 115. The one or more other configurations may include, for example, whether the UE 115 is enabled (e.g., allowed) to perform UAI reporting via OtherConfig, whether the network entity 105 may provide network-side additional conditions (e.g., conditions which may be signaled via RRC signaling, and may be mandatory or optional) , and / or one or more configurations (e.g., inference configurations) of supported functionalities.
[0130] In some examples, a UE 115 may determine (e.g., identify, select) applicable functionalities based on the network-side additional conditions (e.g., if provided) , one or more UE-side additional conditions (e.g., internally known by the UE 115) , and / or model availability in the device. In some examples one or more other configurations may be considered by the UE 115 (e.g., inference configuration) , and the UE 115 may, in some cases, be capable of determining the applicable functionality when network-side additional condition are not provided.
[0131] To support AI / ML functionality and / or model-based LCM operations, applicability information associated with AI / ML functionalities and models may be provided to a network entity 105. Thus, the UE 115 may report applicable functionality (e.g., applicability information) to the network entity 105. In some examples, the applicable functionality may be reported after configuration of the UE 115 to provide applicable functionality, after a change of applicable functionality via UAI, and / or as a response to network-side additional condition requesting applicable functionality reporting. The applicability information may indicate whether AI / ML functionalities or AI / ML models, or a combination thereof, are applicable to the UE 115 (e.g., supported by the UE 115, usable by the UE 115) . In some cases, the network entity 105 may provide configurations to the UE 115 for reporting functionality and model applicability information. For example, the network entity 105 may transmit or output a message (e.g., a control message) indicating a configuration for reporting applicability information associated with AI / ML functionalities and models for maintaining AI / ML-based operations. The applicability information may indicate an applicability of one or more ML functionalities (e.g., AI functionalities) or one or more ML models (e.g., AI models) supported. That is, the applicability information may indicate whether the UE 115 supports and / or uses the one or more AI / ML functionalities or the one or more AI / ML models for one or more cells, in a RAN notification area, in a target area, or the like. In some examples, the UE 115 may report the applicability information using UAI, using a measurement report, or some other signaling.
[0132] As described herein, supported functionalities may refer to one or more functionalities (e.g., AI / ML / prediction functionalities and / or models) that the UE 115 can indicate via UE capability information (via RRC / LPP signaling) , and applicable functionalities may refer to one or more functionalities (e.g., AI / ML / prediction functionalities and / or models) that the UE 115 is ready to apply for inference (where an applicable functionality may be included in the supported functionalities) . Further, activated functionalities (e.g., AI / ML / prediction functionalities and / or models) may refer to functionalities already enabled for performing inference. A “functionality” may refer to one or more feature groups that correspond to capabilities (e.g., UE capabilities) . As an example, for a first functionality or sub-functionality (e.g., corresponding to a first use case) , there may be a first features group that corresponds to, for example, temporal beam prediction (e.g., predicting a set of beams from prior measurements of a set of beams) . Further, for a second functionality or sub-functionality (e.g., corresponding to a second use case) , there may be a second feature group that correspond to, for example, spatial beam prediction (e.g., predicting relatively narrow beams from relatively wide beams) . In some examples, the first feature group and the second feature group may be different.
[0133] In some examples, the UE 115 and / or the network entity 105 may activate or deactivate one or more AI / ML models / functionalities, and / or the UE 115 and / or network entity 105 may perform inference using the AI / ML models / functionalities. Additionally, or alternatively, the UE 115 and / or the network entity 105 may perform monitoring procedures (e.g., performance monitoring) . In some cases, the UE 115 may provide a report (e.g., a CSI report) , which may indicate one or more predicted beam parameters based on a selected AI / ML model.
[0134] AI / ML, and / or prediction models for beam management may be used in deployments supporting uplink-only TRPs and may be used to address various issues and inefficiencies associated with conventional beam management techniques. As an example, because one or more uplink-only TRPs may not transmit in the downlink (e.g., may not support downlink transmissions, may power down components used for downlink transmissions, or the like) , there may be an absence of downlink reference signals from those TRPs that may otherwise indicate some beam correspondence (e.g., between uplink beams and downlink beams at a UE 115) . This may result in relatively frequent sweeping of SRS transmission beams by the UE 115, and may further require tracking of uplink receive beams (e.g., autonomous uplink receive beam tracking) by the TRPs. But such techniques may be inefficient, consume excess resources, consume additional power, and introduce latency, among other issues. Further, sweeping of SRSs by a UE 115 (e.g., in multiple directions) may result in interference (e.g., inter-cell interference, intra-cell interference) to other devices, which may be caused by the absence of downlink reference signals (e.g., corresponding to uplink-only TRPs) that may otherwise “guide” beam searching / sweeping by the UE 115. In cases where one or more UEs 115 are mobile, rotating, and / or experiencing different levels of signal blockage (e.g., when moving with respect to a network entity 105 and / or one or more TRPs in an environment) , the UEs 115 may have to rely on relatively wider uplink transmission beams when communicating with uplink-only TRPs, which may lead to degraded coverage and / or throughput (e.g., because relatively narrower beams may not be practically used by the UE 115) .
[0135] AI / ML-based beam prediction in deployments with one or more uplink-only TRPs, however, may enable more efficient and robust techniques for beam management procedures. As an example, AI / ML-based beam prediction may implement one or more beam prediction cycles (e.g., via temporal beam prediction techniques) , which may enable efficient coordination between the UE 115 and TRPs for identifying uplink transmission and reception beams. Moreover, spatial uplink beam prediction may enable a UE 115 predict relatively narrower transmission beams, which may improve coverage and throughput for the UE 115, enabling transmission via one or more uplink channels using relatively narrow beams. Additionally, or alternatively, AI / ML-based beam prediction may utilize some specific information (e.g., location information of downlink and / or uplink TRPs, pointing direction information associated with downlink reference signals, measurement results of downlink reference signals, or any combination thereof) to predict (e.g., derive) uplink transmission beams for the transmission of SRS. Such information may further be used as AI / ML inputs for temporal and / or spatial beam prediction, and may also be used to modify one or more uplink beams with respect to SRSs. In some cases, AI / ML-based beam prediction using such techniques and information may reduce an SRS transmission density, thereby reducing or eliminating interference (inter-cell interference, intra-cell interference) .
[0136] As an example of one or more inference procedures for beam prediction (e.g., uplink transmission beam prediction) , a UE 115 may measure downlink reference signals from a network entity 105 (or one or more downlink TRPs) . In some examples, the UE 115 may measure and / or determine a respective angle of arrival (AoA) and reference signal metrics (e.g., RSRP) of respective downlink reference signals. The UE 115 may identify one or more receive beams that correspond to the downlink reference signals during the measurement procedures. After measuring the downlink reference signals, the UE 115 may acquire respective locations of one or more TRPs and the UE 115. As an example, the locations of downlink TRPs and / or locations of uplink-only TRPs may be signaled to the UE 115. In some cases, a location of the UE 115 may be determined (e.g., autonomously retrieved) by the UE 115 itself, for example, using various techniques and positioning schemes.
[0137] In some aspects, the UE 115 may optionally acquire relative pointing directions of downlink transmission beams (e.g., associated with the network entity 105, associated with one or more downlink TRPs) and uplink receive beams (e.g., associated with one or more uplink-only TRPs) . In some aspects, the UE 115 may receive the relative pointing directions via signaling from the network entity. As an example, one or more positioning reference signals (PRSs) may be used for the downlink reference signals received from the network entity 105 (or the one or more downlink TRPs) .
[0138] After acquiring the locations of the TRPs and the UE 115 (and optionally the pointing directions of one or more beams) , the UE 115 may, using one or more prediction models (e.g., AI / ML models / functionalities) , to derive (e.g., predict) one or more (initial) uplink transmission beams for SRS transmissions (e.g., for SRS sweeping) . The inputs to the prediction model (s) may include, for example, the measurements and related information obtained by the UE 115, the position (s) of the UE 115 and TRPs, the pointing direction (s) , relative pointing direction relationships between downlink receive beams and candidate uplink transmissions beams, or any combination thereof. In some aspects, uplink transmission beams used to transmit SRS (and / or uplink transmission beams for uplink beam prediction targets) may be modified through the use of beam prediction results, the derived uplink transmission beams, or both.
[0139] In some aspects, AI / ML-based beam prediction may be used for predictions associated with switching between TRPs (e.g., uplink-only TRPs) , which may include the use of one or more cell-specific prediction models (e.g., a model specific to a particular cell) . As an example, during the collection of training data, a UE 115 may receive information associated with respective uplink-only TRPs, and the information may be used during one or more inference procedures. For instance, when transmitting SRSs via multiple uplink transmission beams (e.g., beam sweeping via SRS) , the UE 115 may receive signaling that indicates, for respective uplink-only TRPs, a set of SRS resource indicators (SRIs) , a set of SRS-RSRPs, or both. Such information may be used for training one or more prediction models (e.g., one or more AI / ML algorithms) , and the UE 115 may use such prediction models for switching (e.g., jointly) one or more uplink transmission beams and uplink-only TRPs. In such cases, predictions of an uplink-only TRP switch may be reported to the network by the UE 115, for example, during one or more beam prediction cycles (e.g., in accordance with temporal beam prediction techniques) . In some examples, the set of SRIs and / or SRS-RSRPs may be signaled during one or more SRS transmission cycles, whereas an indication of an uplink-only TRP preference and / or predicted uplink-only TRP switch may be reported by the UE 115 during one or more beam prediction cycles.
[0140] In some examples, AI / ML-based beam prediction may be used for uplink reception beam switching prediction, which may include the use of one or more prediction models supported in various cells (e.g., prediction models that may be supported or implemented in two or more cells) . As an example, during the collection of training data, a UE 115 may receive information associated with respective uplink reception beams (e.g., beam IDs) , and the information may be used during one or more inference procedures. For instance, when transmitting SRSs during an SRS occasion and via a same uplink transmission beam (e.g., beam sweeping via SRS) , the UE 115 may receive signaling that indicates, for respective uplink reception beams (of one or more uplink-only TRPs) , SRIs, SRS-RSRPs, or both. Such information may be used for the training of one or more prediction models (e.g., one or more AI / ML algorithms) , and the UE 115 may use such prediction models for switching one or more uplink transmission beams, which may correspond to a joint switch of an uplink reception beam. In such cases, the SRI (s) and / or RSRP (s) corresponding to SRS cycles may be used for the prediction of uplink receive beam switching reported by the UE 115, which may be reported during one or more beam prediction cycles. In such cases, the uplink reception beams of one or more TRPs (e.g., uplink-only TRPs) may be identifiable by the UE 115. In some aspects, uplink reception beams may be referred to as uplink receive beams, virtual downlink transmission beams, or some other terminology.
[0141] Thus, as described herein, various techniques may be used to support LCM for UE-side beam prediction in deployments including one or more uplink-only TRPs. For example, in accordance with associated ID-based consistency for AI / ML / prediction model input parameters across training and inference, the configuration of certain quantities of SRS resource sets, and / or quantities of SRS resources for each SRS resources set, may be considered for inference procedures and data collection (e.g., for model training) . Additionally, types of network-indicated measurement results of SRSs transmitted by the UE 115 (e.g., whether SRS RSRPs are available during inference) may be used for inference procedures and training data collection.
[0142] In the example of data collection for model training, a UE 115 may use one or multiple SRS resource sets together with network-indicated SRIs / SRS-RSRPs. For instance, the UE 115 may use one SRS resource set for temporal uplink transmission beam prediction, use multiple resource sets for different beam sets (e.g., resource set {#A, #B} as {Set A, Set B} beams) for spatial beam prediction, and / or use multiple resource sets for uplink receive beam switching prediction and / or uplink-only TRP switching prediction. In some aspects, an indication of one or more associated IDs corresponding to uplink-only TRPs may enable the UE to determine spatial filters (e.g., spatial reception filters) with respect to uplink beams of an uplink-only TRP, which may be assumed to be consistent across training and inference (e.g., as long as the same associated ID is identified) . In some examples, such as for cell-specific AI / ML / prediction models, the network and UE 115 may support signaling of downlink reference signal IDs, and the measurements of such reference signals may be considered as “signatures” for optional AI / ML inputs. In such cases, there may be signaling from the network to the UE 115 indicating location information for one or more TRPs (e.g., uplink TRPs, downlink TRPs) and / or beam pointing directions associated with downlink reference signals. Here, as long as the UE 115 identifies a same cell ID, the downlink reference signals indicated by the network may correspond to (e.g., include) the same downlink transmission beam spatial filters across training and inference.
[0143] In the example of AI / ML inference procedures, the UE 115 may predict one or more beams based on network-indicated SRIs / SRS-RSRPs for Set B SRS resource sets (which may have relatively extended periodicities) . In such cases, the UE 115 may predict and manage candidate uplink transmission beams considered for transmitting one or more uplink channels (e.g., PUCCH, PUSCH) . In some examples, such as for spatial beam prediction techniques, the UE 115 may use Set A SRS resource sets together with network-indicated SRIs / SRS-RSRPs as AI / ML inputs (e.g., supplemental inputs) , which may improve prediction accuracy. Here, a greater quantity of SRSs may be scheduled in Set A SRS resource sets, as compared to the SRSs scheduled in Set B resource sets. In some aspects, the described techniques may support predictive UE feedback schemes for uplink reception beams (e.g., for AI / ML models supported in two or more cells) and / or uplink-only TRP switching (e.g., for cell-specific AI / ML models) . Such schemes may be further based on the network indicating one or more associated uplink receive beam-IDs and / or one or more uplink-only TRP IDs for corresponding SRIs / SRS-RSRPs. In some cases, such as for cell-specific AI / ML models, one or more downlink reference signal IDs may be used as “signatures” for optional AI / ML inputs, and there may be some signaling, from the network, indicating location information for one or more TRPs (e.g., downlink TRPs, uplink TRPs) and / or beam pointing directions for the one or more downlink reference signals. Here, as long as the UE 115 identifies a same cell ID, the downlink reference signals indicated by the network may correspond to (e.g., include) the same downlink transmission beam spatial filters across training and inference.
[0144] The described techniques may further enable efficient performance monitoring procedures. As an example, the UE 115 may utilize SRSs configured for performance monitoring, where such SRSs may exclude an association with downlink reference signals (and corresponding TCI-states, spatial filters) , and the SRS may be scheduled and transmitted based on UE-predicted beams. In such cases, the network entity 105 may signal SRIs / SRS-RSRPs for the performance monitoring SRSs transmitted by the UE 115. In some examples, such as for UE-based monitoring, the UE 115 may determine or recommend AI / ML model activation or deactivation by comparing predictive and measured sets of uplink beams. Additionally, for network-based and / or UE-assisted performance monitoring, the UE 115 may report raw and / or statistical differences between predictive and measured sets of uplink beams.
[0145] In some aspects, performance monitoring SRSs may be utilized for uplink reception beam and / or uplink-only TRP switching prediction accuracy monitoring. In such cases, the network entity 105 may separately schedule SRS resource sets associated with different uplink reception beams and / or uplink-only TRPs. The UE 115 may indicate identified / predicted uplink transmission beams via such SRSs corresponding to different uplink reception beam (s) and / or uplink-only TRPs. Such techniques may be used for UE-based performance monitoring, network-based performance monitoring, and / or UE-assisted performance monitoring schemes.
[0146] FIG. 2 shows an example of a network architecture 200 (e.g., a disaggregated base station architecture, a disaggregated RAN architecture) that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The network architecture 200 may illustrate an example for implementing one or more aspects of the wireless communications system 100. The network architecture 200 may include one or more CUs 160-a that may communicate directly with a core network 130-a via a backhaul communication link 120-a, or indirectly with the core network 130-a through one or more disaggregated network entities 105 (e.g., a Near-RT RIC 175-b via an E2 link, or a Non-RT RIC 175-a associated with an SMO 180-a (e.g., an SMO Framework) , or both) . A CU 160-a may communicate with one or more DUs 165-a via respective midhaul communication links 162-a (e.g., an F1 interface) . The DUs 165-a may communicate with one or more RUs 170-a via respective fronthaul communication links 168-a. The RUs 170-a may be associated with respective coverage areas 110-a and may communicate with UEs 115-a via one or more communication links 125-a. In some implementations, a UE 115-a may be simultaneously served by multiple RUs 170-a.
[0147] Each of the network entities 105 of the network architecture 200 (e.g., CUs 160-a, DUs 165-a, RUs 170-a, Non-RT RICs 175-a, Near-RT RICs 175-b, SMOs 180-a, Open Clouds (O-Clouds) 205, Open eNBs (O-eNBs) 210) may include one or more interfaces or may be coupled with one or more interfaces configured to receive or transmit signals (e.g., data, information) via a wired or wireless transmission medium. Each network entity 105, or an associated processor (e.g., controller) providing instructions to an interface of the network entity 105, may be configured to communicate with one or more of the other network entities 105 via the transmission medium. For example, the network entities 105 may include a wired interface configured to receive or transmit signals over a wired transmission medium to one or more of the other network entities 105. Additionally, or alternatively, the network entities 105 may include a wireless interface, which may include a receiver, a transmitter, or transceiver (e.g., an RF transceiver) configured to receive or transmit signals, or both, over a wireless transmission medium to one or more of the other network entities 105.
[0148] In some examples, a CU 160-a may host one or more higher layer control functions. Such control functions may include RRC, PDCP, SDAP, or the like. Each control function may be implemented with an interface configured to communicate signals with other control functions hosted by the CU 160-a. A CU 160-a may be configured to handle user plane functionality (e.g., CU-UP) , control plane functionality (e.g., CU-CP) , or a combination thereof. In some examples, a CU 160-a may be logically split into one or more CU-UP units and one or more CU-CP units. A CU-UP unit may communicate bidirectionally with the CU-CP unit via an interface, such as an E1 interface when implemented in an O-RAN configuration. A CU 160-a may be implemented to communicate with a DU 165-a, as necessary, for network control and signaling.
[0149] A DU 165-a may correspond to a logical unit that includes one or more functions (e.g., base station functions, RAN functions) to control the operation of one or more RUs 170-a. In some examples, a DU 165-a may host, at least partially, one or more of an RLC layer, a MAC layer, and one or more aspects of a PHY layer (e.g., a high PHY layer, such as modules for FEC encoding and decoding, scrambling, modulation and demodulation, or the like) depending, at least in part, on a functional split, such as those defined by the 3rd Generation Partnership Project (3GPP) . In some examples, a DU 165-a may further host one or more low PHY layers. Each layer may be implemented with an interface configured to communicate signals with other layers hosted by the DU 165-a, or with control functions hosted by a CU 160-a.
[0150] In some examples, lower-layer functionality may be implemented by one or more RUs 170-a. For example, an RU 170-a, controlled by a DU 165-a, may correspond to a logical node that hosts RF processing functions, or low-PHY layer functions (e.g., performing fast Fourier transform (FFT) , inverse FFT (iFFT) , digital beamforming, physical random access channel (PRACH) extraction and filtering, or the like) , or both, based at least in part on the functional split, such as a lower-layer functional split. In such an architecture, an RU 170-a may be implemented to handle over the air (OTA) communication with one or more UEs 115-a. In some implementations, real-time and non-real-time aspects of control and user plane communication with the RU (s) 170-a may be controlled by the corresponding DU 165-a. In some examples, such a configuration may enable a DU 165-a and a CU 160-a to be implemented in a cloud-based RAN architecture, such as a vRAN architecture.
[0151] The SMO 180-a may be configured to support RAN deployment and provisioning of non-virtualized and virtualized network entities 105. For non-virtualized network entities 105, the SMO 180-a may be configured to support the deployment of dedicated physical resources for RAN coverage requirements which may be managed via an operations and maintenance interface (e.g., an O1 interface) . For virtualized network entities 105, the SMO 180-a may be configured to interact with a cloud computing platform (e.g., an O-Cloud 205) to perform network entity life cycle management (e.g., to instantiate virtualized network entities 105) via a cloud computing platform interface (e.g., an O2 interface) . Such virtualized network entities 105 can include, but are not limited to, CUs 160-a, DUs 165-a, RUs 170-a, and Near-RT RICs 175-b. In some implementations, the SMO 180-a may communicate with components configured in accordance with a 4G RAN (e.g., via an O1 interface) . Additionally, or alternatively, in some implementations, the SMO 180-a may communicate directly with one or more RUs 170-a via an O1 interface. The SMO 180-a also may include a Non-RT RIC 175-a configured to support functionality of the SMO 180-a.
[0152] The Non-RT RIC 175-a may be configured to include a logical function that enables non-real-time control and optimization of RAN elements and resources, Artificial Intelligence (AI) or Machine Learning (ML) workflows including model training and updates, or policy-based guidance of applications / features in the Near-RT RIC 175-b. The Non-RT RIC 175-a may be coupled to or communicate with (e.g., via an A1 interface) the Near-RT RIC 175-b. The Near-RT RIC 175-b may be configured to include a logical function that enables near-real-time control and optimization of RAN elements and resources via data collection and actions over an interface (e.g., via an E2 interface) connecting one or more CUs 160-a, one or more DUs 165-a, or both, as well as an O-eNB 210, with the Near-RT RIC 175-b.
[0153] In some examples, to generate AI / ML models to be deployed in the Near-RT RIC 175-b, the Non-RT RIC 175-a may receive parameters or external enrichment information from external servers. Such information may be utilized by the Near-RT RIC 175-b and may be received at the SMO 180-a or the Non-RT RIC 175-a from non-network data sources or from network functions. In some examples, the Non-RT RIC 175-a or the Near-RT RIC 175-b may be configured to tune RAN behavior or performance. For example, the Non-RT RIC 175-a may monitor long-term trends and patterns for performance and employ AI or ML models to perform corrective actions through the SMO 180-a (e.g., reconfiguration via O1) or via generation of RAN management policies (e.g., A1 policies) .
[0154] As described herein, a UE 115-a may support various techniques to support efficient AI / ML model implementation, including training data collection, inference frameworks, performance monitoring, and ensuring consistency of network-side additional conditions (e.g., conditions for performing beam prediction) , for network-side receive beam prediction by a UE using AI / ML model (s) . The AI / ML models may be trained based on measurements (e.g., RSRP measurements, SRI measurements, SINR measurements) from uplink reference signals (e.g., SRSs) transmitted by the UE 115-a.
[0155] In some aspects, for training data collection, a UE 115-a may be scheduled with a quantity of SRS resource sets (without associated TCI-State or SpatialRelationInfo) . The UE 115-a may acquire measured channel characteristics as measured at the network entity with respect to the uplink reference signals via control signaling, such as via DCI, a MAC-CE, or RRC signaling. AI / ML algorithm (s) at the UE 115-a with respect to the associated ID that output predicted channel characteristics for the uplink-receive beams may be trained, based on measurements of SRS resource sets and the network-indicated measurements, to predict spatial and / or temporal domain characteristics associated uplink Tx beams that the UE 115-a may use for transmission of uplink signals (e.g., PUSCH, PUCCH towards an uplink TRP.
[0156] In some aspects, for inference, the UE 115-a may use an AL / ML model to predict channel characteristics with respect to uplink Tx beams or different uplink TRPs. For example, after verifying that the associated IDs for training data collection and for inference are the same, the UE 115-a may perform AI / ML inference procedures to determine a set of applicable uplink Tx beams at the UE 115-a (and associated uplink Rx beams at one or more uplink TRPs) . In some examples, the UE 115-a may determine or predict beam switching, switching between uplink TRPs, or both, using AI / ML inferencing. For example, the UE 115-a may identify a current or future switch between uplink TRPs or between uplink Rx beams, and may report an indication of the switch to the network.
[0157] In some aspects, for performance monitoring, while the prediction associated with the inference step is on-going, the network entity may further schedule additional uplink reference signals (e.g., SRSs) for performance monitoring. In some cases, performance monitoring may be performed at the network, where the network entity may compare UE-reported prediction results against measurements on the uplink reference signals to determine further actions. In other cases, performance monitoring may be performed at the UE 115-a, and for UE-based performance monitoring the UE 115-a may receive network-signaled measurement results for the uplink reference signals, representing channel characteristics on corresponding uplink-receive beams and the UE 115-a may determine or recommend further AI / ML action (e.g., activation, deactivation, or a switch of models) based on such signaled measurement results.
[0158] FIG. 3 shows an example of a wireless communications system 300 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. For example, the wireless communications system 300 may be an example of a wireless communications system supporting uplink TRP deployment. In some aspects, the wireless communications system may support communications between a UE 115, a network entity 105, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0159] In some wireless communications systems, a UE 115 may use a prediction model to predict one or more parameters or characteristics associated with uplink or downlink transmissions. In such cases, the UE 115 may measure respective beams that corresponds to one or more reference signals (e.g., SSBs or CSI-RSs) via a first set of resources, which may be referred to as Set B beams, during a first set of measurement occasions. Additionally, the UE 115 may perform beam prediction (e.g., prediction of one or more parameters and / or characteristics associated with uplink and / or downlink transmissions) for a set of beams associated with a second set of resources, which may be referred to as Set A beams, using a prediction model (e.g., an AI / ML model, a prediction model) and based on model training measurement results for resources associated with the Set B beams and / or historical measurement results for resources associated with the Set B beams. In such cases, the UE 115 may perform the beam prediction using the prediction model in accordance with a set of prediction parameters that may include, for example, a Set B beam measurement window length, a Set B beam measurement periodicity, and a furthest temporal prediction duration for Set A beams. Further, in some cases, performance monitoring for the prediction model may be performed to identify how closely the predictions of the prediction model match actual measurements, and model updates may be performed based on the performance monitoring.
[0160] Training of the prediction model may be performed using a set of training beam measurements that are associated with a model identification of the prediction model, which may be referred to as an associated ID. Further, model inference for predicted measurements of beams, and model performance monitoring, may also be performed based on the associated ID of the model. However, efficient signaling for performing model training, inference, and performance monitoring has not been defined. Further, in some cases it may be desirable to perform beam prediction at a UE 115 for uplink-receive beam measurements at a network entity, which may be useful when beam reciprocity may not be present for one or more beams (e.g., when a relatively strong downlink received signal strength of a downlink beam has a relatively weak uplink received signal strength for a corresponding uplink beam) .
[0161] In some cases, a UE 115 may support various techniques to support efficient AI / ML model implementation, including training data collection, inference frameworks, performance monitoring, and ensuring consistency of network-side additional conditions (e.g., conditions for performing beam prediction) , for network-side receive beam prediction (e.g., prediction of one or more parameters or characteristics associated with uplink or downlink transmission) by a UE 115 using AI / ML model (s) . The AI / ML models may be trained based on measurements (e.g., RSRP or SINR measurements) from uplink reference signals (e.g., SRSs) transmitted by a UE.
[0162] In some aspects, for training data collection, a UE 115 may be scheduled with a quantity of downlink reference signals (e.g., SSB or CSI-RSs) , and the same quantity of uplink reference signals (e.g., SRSs) together with an associated ID, where the downlink reference signals are associated with the same quantity of network entity downlink-transmit / uplink-receive beams, and spatial relation information (e.g., indicated by transmission configuration indicator (TCI) states or srs-SpatialRelationInfo) with respect to the uplink reference signals are respectively associated with the downlink reference signals. The UE 115 may acquire measured channel characteristics as measured at the network entity with respect to the uplink reference signals via control signaling, such as via downlink control information (DCI) a medium access control (MAC) control element (CE) , or RRC signaling. AI / ML algorithm (s) at the UE 115 with respect to the associated ID that output predicted channel characteristics for the uplink-receive beams may be trained, where the inputs include at least channel characteristics of the corresponding downlink-transmit beams based on measurements of the downlink reference signals, and the outputs may be trained based at least on the channel characteristics measured from the uplink reference signals.
[0163] In some aspects, for inference, an AL / ML model may predict channel characteristics with respect to uplink-receive beams using downlink-transmit beam measurements. In some cases, the UE 115 may be instructed by the network entity to predict and report channel characteristics on a quantity of uplink-receive beams (e.g., predicted RSRPs / SINRs that would be measured at the network entity for an uplink signal) , which are respectively associated with the same quantity of downlink-receive beams transmitted via the downlink reference signals, and the prediction is based at least in part on measurements of the downlink reference signals and the indicated associated ID for the AI / ML model. In some cases, the network entity may also signal the associated ID for the instructed UE-side prediction.
[0164] In some aspects, for performance monitoring, while the prediction associated with the inference step is on-going, the network entity may further schedule additional uplink reference signals (e.g., SRSs) with spatial relation information (e.g., TCI-states and / or srs-SpatialRelationInfos) associated with one or more of the downlink reference signals considered in the inference step. In some cases, performance monitoring may be performed at the network, where the network entity may compare UE reported prediction results against measurements on the uplink reference signals to determine further actions. In other cases, performance monitoring may be performed at the UE 115, and for UE-based performance monitoring the UE 115 may receive network-signaled measurement results for the uplink reference signals, representing channel characteristics on corresponding uplink-receive beams and the UE 115 may determine or recommend further AI / ML action (e.g., activation, deactivation, or a switch of models) based on such signaled measurement results. In further cases, UE-assisted performance monitoring may be used, where the UE 115 may receive network-signaled measurement results for the uplink reference signals, representing actually measured channel characteristics on corresponding uplink-receive beams; the UE may further calculate and report statistical differences between its predicted channel characteristics for the uplink-receive beams and the actually measured channel characteristics on the same uplink-receive beams in their overlapping occasions; and the network entity may determine further actions according to such UE reports.
[0165] The wireless communications system 300 may experience a relatively dense deployment of wireless communications devices, including multiple UEs 115 within a same location (e.g., within a same coverage area, cell, geographic location, or the like) . In such cases, the wireless communications system may experience asymmetric traffic loads for uplink and downlink transmissions. The wireless communications system may implement techniques for managing such traffic loads (e.g., for efficiently handling scenarios where uplink transmissions are relatively dense) . Such techniques may include the deployment of the one or more TRPs that are configured to support uplink transmissions from one or multiple UEs 115, but not used for downlink transmissions to the UEs 115 (e.g., uplink TRPs or otherwise “uplink only” TRPs) . The one or more TRPs configured for uplink transmissions (but not downlink transmissions) may be referred to as uplink-only TRPs, or some similar terminology.
[0166] In some cases, a UE 115 may encounter various challenges for beam management related to uplink TRPs, including power expenditure, overhead, coverage, latency, and interference challenges, for beam management and prediction. For example, to identify appropriate (e.g., best, most accurate, or highest performance) uplink transmission (Tx) and uplink receive (Rx) beams associated with the uplink TRPs, the UE 115 may be scheduled with frequently transmitted SRSs (e.g., SRS transmissions that have a periodicity that exceeds a threshold) so that the UE 115 may sweep uplink Tx beams, while the uplink TRPs may use various uplink Rx beams to receive such SRSs. In such cases, the network entity may signal SRIs and / or SRS-RSRPs for scheduling PUxCH (e.g., PUSCH and / or PUCCH) or for “guiding” the UE 115 towards alternative uplink Tx beams) .
[0167] In some aspects, however, such beam sweeping techniques may be inefficient for the wireless system. For example, because no downlink reference signal is transmitted by the TRP 102-b (e.g., an uplink-only TRP) for beam correspondence, the UE 115 may use frequent transmission of SRSs, which may consume significant amounts of uplink Tx power for the UE 115, and may consume correspondingly significant amount of uplink overhead for related cells or TRPs, which may perform frequent autonomous uplink Rx beam tracking. Additionally, or alternatively, due to lack of downlink RSs as unified TCI states and / or uplink TCI-states to “guide” SRS transmission of the UE 115 (while the UE 115 may be randomly or constantly moving and / or rotating, or when the UE 115 experiences blockage) , wider and weaker (rather than narrower and sharper) uplink Tx beams may be used for such SRS transmissions (and for subsequent PUSCH transmissions) . Such use of wider and weaker uplink Tx beams may reduce the potential uplink coverage and / or throughput benefits considered by heterogeneous networks, because UE 115 may be capable of using the narrow and / or shaper (e.g., more targeted, higher performance) uplink Tx beams towards the downlink TRP, but may instead use wider and / or weaker beams. Additionally, or alternatively, lack of reciprocity between uplink TRPs and the UE 115 may be that autonomous (initial) SRS sweeping by the UE 115 may create intra-cell interference and / or inter-cell interference. For example, because no downlink reference signals may be provided by an uplink TRP, the initial SRS sweeping and beam search performed by the UE 115 may cause interference (e.g., because one or more beam directions of the UE 115 may not be guided in an initial direction using a downlink reference signal) .
[0168] In some other aspects, some such beam sweeping techniques may reduce the efficiency of uplink Tx beam alignment and / or uplink Rx beam alignment. That is, due to movement of the UE 115 (translational movements, rotations) , latency of the uplink Tx beam alignment and / or Rx beam alignment may be increased. Such increased latencies may also be due to the uplink TRPs being implemented with multiple uplink Rx beams. In some such aspects, to facilitate low-cost uplink TRPs, analog beamforming may be preferred such that at each beam monitoring instance, the network entity 105 may use a single uplink Rx beam to receive a certain SRS transmitted by the UE 115, which may lead to increased latency to track and maintain uplink Tx / Rx beam pairings. Additionally, or alternatively, the UE 115 may maintain separate and simultaneous uplink beam management procedures with different uplink multi-TRPs (mTRPs) (for example, if the UE 115 is moving from a first uplink TRP towards a second uplink TRP) . In such cases, challenges such as increased latency, increased power, and overhead may be linearly increased or affected as the quantity of mTRPs increases.
[0169] To increase the efficiency of beam management for communications between a UE 115 and one or more uplink TRPs (e.g., TRP 102-a and TRP 102-b) , the UE 115 may implement beam prediction (e.g., prediction of one or more parameters or characteristics associated with uplink or downlink transmission) including temporal beam prediction 305 and spatial beam prediction 310 using downlink Tx beams for downlink-uplink collocated TRPs, to uplink TRPs, where beam prediction using the downlink Tx beams is performed using one or more AI / ML models. For example, the UE 115 may use temporal beam prediction, which may reduce SRS transmission frequency, which may reduce power expenditure, beam misalignments, and signaling overhead. In such temporal beam prediction, the UE 115 may predicted (e.g., using AI / ML) uplink Tx and / or uplink Rx beams that are to be used between adjacent SRS transmission instances. The UE 115 may also implement spatial beam prediction to increase coverage and throughput. For example, the UE 115 may utilize AI / ML to predict narrower or sharper uplink Tx beams for transmitting PUSCH, even if wider and weaker uplink Tx beams are exclusively (or more frequently) transmitted via SRSs. In some other examples, the UE 115 may receive assistance information which may include location and / or pointing information which the UE 115 may use to derive and / or predict uplink Tx beams associated with SRSs, which may reduce interference, SRS transmission density, or both. For example, the assistance information may include location information associated with uplink and / or downlink TRPs, and / or pointing direction information for downlink reference signals, and / or measurement results associated with the downlink reference signals. The UE 115 may use the assistance information to predict initial uplink Tx beams associated with SRSs (e.g., the UE 115 may use the assistance information as inputs to one or more AI / ML models for temporal or spatial prediction) , which may reduce SRS transmission density.
[0170] FIG. 4 shows examples of an AI / ML implementation 401, an AI / ML implementation 402, and an AI / ML implementation 403 that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. In some examples, the AI / ML implementation 401, AI / ML implementation 402, and the AI / ML implementation 403 may support AI / ML beam-related prediction for uplink TRP deployments, and may support communications between a UE 115, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0171] A UE 115 may implement various techniques to support AI / ML training and inference for communications in a wireless network, including communication with uplink TRPs. In some aspects, the UE 115 may determine SRS uplink Tx beams autonomously (e.g., without use of a TCI state provided by a guiding downlink reference signal) using AI / ML predictions. In some cases, the UE 115 may utilize SRS and network indications of SRIs and / or SRS RSRPs as inputs to an AI / ML model to predict SRS uplink Tx beams. In some aspects, the UE 115 may implement AI / ML implementation 401 for AI / ML and spatial prediction supported in one or more cells. For example, to facilitate wide-to-narrow UE-side uplink Tx spatial beam prediction (e.g., via wide beams 405 and narrow beams 410) , the UE 115 may be scheduled with (at least) two sets of SRSs (e.g., SRS resource set #1 and SRS resource set #2) , where the first SRS resource set (e.g., SRS resource set #1) is scheduled for wider uplink Tx beams and the second SRS resource set (e.g., SRS resource set #2) is scheduled for narrower uplink Tx beams. In such examples, a network entity may separately indicate SRIs and / or SRS-RSRPs corresponding to each SRS resource set, so that the UE 115 may use respective SRIs and / or SRS-RSRPs as AI / ML inputs and output labels for model training. In addition, in some examples, a quantity of SRSs in the first SRS resource set (e.g., corresponding to wide beams 405) may be smaller than the quantity of SRSs in the second SRS resource set (e.g., corresponding to narrow beams 410) , but the quantity of SRS resources addressed by the network-indicated SRIs and / or RSRPs regarding the first SRS resource set may be greater than the SRIs and / or RSRPs indicated in the second SRS resource set (e.g., the first SRS resource set may include wider beam inputs) . In such examples, the network may indicate a SRIs and / or RSRPs corresponding to each wide beam input of the first SRS resource set, and may indicate one or more “top” SRIs and / or RSRPs (e.g., SRIs and / or RSRPs having a greatest value of the total quantity of SRIs and / or RSRPs) corresponding to the narrow beam inputs of the second SRS resource set.
[0172] In some aspects, the UE 115 may utilize AI / ML implementation 401 that may be used in various different cells (e.g., the AI / ML implementation 401 may be used across multiple cells, the AI / ML implementation 401 may not be limited to a particular cell, the AI / ML implementation 401 may be common across cells) for implementation-based uplink Tx beam prediction (e.g., prediction of one or more parameters or characteristics associated with uplink transmission) . For example, temporal and / or spatial uplink Tx beam prediction results may be used directly by the UE 115 to derive predicted uplink Tx beams used for transmitting PUxCH (e.g., PUSCH or PUCCH) with or without informing the network. For temporal predictions, the UE 115 may be scheduled with SRSs that have periodicities that are greater than SRSs utilized during training data collection, and the uplink Tx beams used for uplink transmissions (e.g., PUxCH transmissions) between SRS occasions may be predicted based on network-signaled SRIs and / or SRS-RSRPs. For spatial predictions, the UE may use wider uplink Tx beams to transmit SRSs scheduled in a single SRS resource set, and relatively narrower uplink Tx beams that are to be used for uplink transmissions (e.g., PUxCH transmissions, which may include PUCCH transmission, PUSCH transmissions, among other examples) may be predicted based on network-signaled SRIs and / or SRS-RSRPs. In such examples, the SRSs used for beam predictions may lack associated downlink reference signals (e.g., from an uplink TRP)
[0173] In some implementations, the UE 115 may utilize AI / ML implementation 402 across various cells, and may be applicable to various TRPs in various cells where the input and / or output of the AI / ML model may be, in some cases, associated with a single TRP during a training and / or inference session. The UE 115 may receive signaling that includes associated-ID (s) for one or more uplink TRPs (e.g., TRP 102-a and TRP 102-b) , and / or uplink Rx beams of the one or more TRPs associated with the indicated SRIs and / or SRS-RSRPs transmitted by the network. Such signaling of the associated-ID (s) , the uplink Rx beams may allow for the UE 115 to determine consistency of related network-side additional conditions (e.g., spatial Rx filters) between training and inference of the one or more AI / ML models. In some aspects, the UE 115 may switch between TRPs having the same associated ID (e.g., TRP switching from TRP 102-a to TRP 102-b) , where SRIs and / or SRS-RSRPs provided to the UE 115 may correspond to the same associated ID (e.g., associated ID “A” ) . Additionally, or alternatively, the UE 115 may switch between TRPs having different associated IDs (e.g., TRP switching from TRP 102-a having associated ID “A” to TRP 102-c having associated ID “B” ) , where SRIs and / or SRS-RSRPs provided to the UE 115 may correspond to the different associated IDs (e.g., associated ID “A” prior to TRP switching and associated ID “B” after switching) . In some cases, different UL TRPs may be associated with different antenna panels and different uplink Rx beam codebooks, and AI / ML models trained on a first type of uplink TRP may perform either better or worse on the same or different types of uplink TRPs. In such cases, use of associated IDs may allow for the UE 115 to support network-side additional condition consistency for uplink TRPs (and uplink Rx beams associated with the uplink TRPs) . That is, if the UE 115 determines consistency of associated IDs for beams used between training and inference, the UE 115 may be able to determine consistency of network-side additional conditions.
[0174] In some aspects, for some AI / ML implementations, the UE 115 may in some cases transmit relatively narrower uplink Tx beams to improve beam prediction accuracy. For example, in wide-to-narrow spatial prediction, additional SRS resource set (s) may be scheduled during an inference stage of AI / ML, so that the UE 115 may sparsely transmit the candidate narrower uplink Tx beams using the scheduled SRSs (e.g., where sparse transmission may be transmission of narrow beams relatively less frequently than wider beams) . In such cases, the UE 115 may utilize network-indicated SRIs and / or SRS-RSRPs associated with such SRSs as additional AI / ML inputs to improve prediction accuracy. In some examples, the UE 115 may receive separate groups of SRIs and / or SRS-RSRPs for the SRS resource set scheduled for wider uplink Tx beams and for the SRS resource set (s) scheduled for narrower uplink Tx beams.
[0175] In some aspects, the UE 115 may utilize AI / ML implementation 403 for cell-specific AI / ML. In some aspects, the UE 115 may use downlink reference signals transmitted from the TRP 102-d (e.g., an uplink and downlink TRP) to use as “signatures” for AI / ML inputs (e.g., for cell-specific models due to various downlink and uplink TRP versus uplink-only mTRPs configurations across different cells) . In order to support AI / ML inputs based on downlink reference signaling, a network entity may instruct the UE 115 regarding which downlink reference signals (e.g., certain SSBs and / or CSI-RSs) may be considered as “signatures, ” which the UE 115 may consider as being consistent (e.g., consistent spatial Tx filters) across training and inference (at least within the considered cell) . In some examples, the UE 115 may determine how or when to utilize such “signatures” as input into the one or more AI / ML models to derive beam prediction results.
[0176] Additionally, or alternatively, the UE 115 may support uplink Rx beam switching (for an AI / ML implementation used in various cells) and TRP switching (for cell-specific AI / ML prediction) . For example, based on AI / ML design, training, and inference, the UE 115 may predict and report which uplink TRP (s) and / or which uplink Rx beam (s) may be preferred for receiving uplink transmissions from the UE 115 (e.g., for receiving PUxCH) , based on network-signaled associated-IDs regarding the uplink TRP (s) and / or uplink Rx beams of the uplink TRPs. In some examples, the UE 115 may support cell-specific AI / ML based on network-indication of downlink reference signals as “signatures” for inference. For example, prior to inference, the UE 115 may receive and measure downlink reference signals which are cell-specific “signatures” to obtain additional AI / ML inputs. In some examples, the downlink reference signals may be indicated as cell specific signatures based on additional network signaling.
[0177] FIG. 5 shows an example of an AI / ML operation flow 500 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. In some examples, the AI / ML operation flow 500 may support AI / ML beam-related prediction for uplink TRP deployments, and may support communications between a UE 115, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0178] AI / ML operation flow 500 may support beam prediction and beam management using temporal uplink Tx beam prediction (using AI / ML) and spatial uplink Tx beam prediction (using AI / ML) . In a first step 505, the UE 115 may measure downlink reference signals from the TRP 102-b (which may be a downlink TRP) , including measuring RSRPs of the respective downlink reference signals. In addition, the UE 115 may determine downlink receive beams which correspond with or “match” the downlink reference signals during measurement of the downlink reference signals. In a second step 510, the UE 115 may acquire locations of TRPs (e.g., TRP 102-a and TRP 102-b) and the location of the UE 115. For example, the UE 115 may determine or acquire a set of location coordinates for the TRP 102-a (e.g., an uplink TRP) as (xUL, yUL, zUL) , a set of location coordinates for the TRP 102-b (e.g., a downlink TRP) as (xDL, yDL, zDL) , and a set of location coordinates for the UE 115 as (xUE, yUE, zUE) . In some examples, the locations of the TRP 102-a and the location of the TRP 102-b may be signaled by the network, and the location of the UE 115 may be autonomously retrieved by the UE 115.
[0179] For example, the UE 115 may be equipped with one or more sensors that may identify orientations, locations, rotations, or any combination thereof, of the UE 115. Based on additional network-side assistance information regarding the locations of the TRP 102-a and the TRP 102-b, as well as downlink Tx beam pointing directions, the UE 115 may be able to determine the directions where uplink Tx beams should be directed toward an uplink TRP (such as TRP 102-b) . The UE 115 may transmit candidates of such uplink Tx beams via SRS transmissions. In such cases, the UE 115 may use SRIs and / or SRS-RSRPs associated with the transmit candidates as AI / ML inputs for temporal and / or spatial uplink Tx beam prediction. Such uplink Tx beam predictions may reduce SRS overhead, and uplink Tx power. In some examples, the UE 115 may derive an estimate of the uplink Tx beams using AI / ML, or using other prediction processes. For example, using information such as the absolute locations of the TRP 102-a and the TRP 102-b, pointing directions of downlink reference signals (e.g., positioning reference signals) together with location information of the UE 115, orientation and / or rotation information autonomously identified by the UE 115, RSRPs and / or angle of arrival estimated from the downlink reference signals, the UE 115 may identify uplink Tx beams without using AI / ML, or may be used in combination with AI / ML. In some examples, advanced AI / ML design may also utilize such location and directional information as inputs for AI / ML-based temporal and spatial beam predictions (e.g., prediction of one or more parameters or characteristics associated with uplink or downlink transmission) .
[0180] In some implementations, periodic and / or aperiodic SRSs may be utilized in addition to such NW-side assistance information to facilitate the AI / ML predictions. For example, utilization of SRS may reduce errors or latency associated with UE-side orientation and / or rotation estimations, may reduce impacts related to uplink Tx beam selection due to reflection / multi-path and blockage around the UE (e.g., by body blockage, hand blockage, or nearby vehicles) , and may also be used in cases of uplink TRP uplink Tx beam estimation, where beam tracking, beam refinement, and beam prediction of uplink Rx beams at uplink TRPs may utilize SRSs. In at least some cases, however, signaling of network-side assistance information may reduce SRS transmission densities, which may increase power savings and reduce device overhead.
[0181] In some aspects, the UE 115 may receive assistance information during training data collection, or during inferencing. In some examples, the assistance information may include a location of a downlink TRP with respect to one or more uplink TRPs, beam pointing directions of the respective downlink Tx spatial filters associated with downlink reference signals, beam pointing directions of uplink Rx beams for one or more uplink TRPs, or any combination thereof. In some examples, the UE 115 may determine or decide whether to use such assistance information as AI / ML inputs, or whether to use the assistance information to derive AI / ML inputs for the one or more beam prediction models (e.g., a model for prediction of one or more parameters or characteristics associated with uplink or downlink transmission) . In some examples, the UE 115 may transmit one or more requests to acquire different types of assistance information during training data collection, or during inference.
[0182] In an optional third step 515, the UE 115 may acquire pointing directions of the downlink Tx beams (associated with the TRP 102-b) and the uplink Rx beams (associated with the TRP 102-a) . In some examples, the pointing directions may be signaled by the network. In some examples, the downlink reference signals may be positioning reference signals. In a fourth step 520, the UE 115 may derive and / or predict uplink Tx beams (e.g., initial uplink Tx beams) for SRS transmission and / or SRS beam sweeping. In some implementations, the UE 115 may utilize one or more AI / ML models to predict the uplink Tx beams from a set of candidate uplink Tx beams using different inputs. For example, the inputs may include information acquired from first step 505, second step 510, optional third step 515, or any combination thereof, together with relative pointing direction relationships between downlink Rx beams and candidate uplink Tx beams.
[0183] In some examples, the UE 115 may perform temporal prediction 525 using the one or more AI / ML models to determine the predicted uplink Tx beams. Additionally, or alternatively, the UE 115 may use spatial prediction 530 using the one or more AI / ML models to determine the predicted uplink Tx beams.
[0184] FIG. 6 shows an example of an AI / ML beam management procedure 600 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. In some examples, the AI / ML beam management procedure 600 may support AI / ML beam-related prediction for uplink TRP deployments, and may support communications between a UE 115, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0185] The AI / ML beam management procedure 600 may support cell common or cell specific AI / ML based on TRP-ID and / or uplink Rx beam ID (e.g., an uplink signal ID) indication obtained during training data collection. For example, during training data collection, the UE 115 may receive multiple groups of measurement results for a certain SRS resource set in a given SRS transmission instance. In such examples, each group of measurement results may be associated with an additional identifier of a certain uplink TRP and / or a certain uplink Rx beam.
[0186] The network may provide (e.g., signal) a set of measurements (e.g., SRIs, SRS-RSRPs, or both) for a first TRP (e.g., uplink TRP 1, TRP 102-a) and may provide a separate set of measurements (e.g., SRIs, SRS-RSRPs, or both) for a second TRP (e.g., uplink TRP 2, TRP 102-b) . In some examples, the UE 115 may preform uplink beam sweeping via SRSs, and may receive the measurements (which may include uplink TRP IDs signaled together with the SRIs / SRS-RSRPs) for different swept SRSs. In such examples, the uplink receive beams of the TRP 102-a and the TRP 102-b may be transparent to the UE 115, and the UE 115 may be capable of receiving SRSs from multiple uplink TRPs.
[0187] In some other examples, the network may provide (e.g., signal) a set of measurements (e.g., SRIs, SRS-RSRPs, or both) for the first TRP (e.g., uplink TRP 1, TRP 102-a) and may provide a separate set of measurements (e.g., SRIs, SRS-RSRPs, or both) for the second TRP (e.g., uplink TRP 2, TRP 102-b) . In some examples, the UE 115 may perform uplink beam sweeping via SRSs, and may receive the measurements (which may include uplink Rx beam IDs signaled together with the SRIs / SRS-RSRPs) for different swept SRSs. In such examples, the uplink receive beams of the TRP 102-aand the TRP 102-b may be identifiable to the UE 115. After training data collection, the UE 115 may train one or more AI / ML algorithms or AI / ML models (e.g., perform offline training) using the data.
[0188] In some implementations, the UE 115 may perform AI / ML inferencing using the one or more trained AI / ML algorithms or trained AI / ML models, and may feedback TRP-ID and / or uplink Rx Beam ID preference during inference. For example, during inference, the UE 115 may use various type of control signaling (e.g., RRC signaling, MAC-CE signaling, UCI signaling, or any combination thereof) to report a current switch or a predicted future switch towards an indicated Rx beam-ID associated with a respective uplink TRP ID, or towards another uplink TRP ID. In some cases, AI / ML inferencing-based beam switching or TRP switching may be further based on various network-provided measurement results for the SRS resource sets during inference (which may include the corresponding uplink TRP ID (s) and / or the associated uplink Rx beam ID (s) ) .
[0189] In some aspects, the UE 115 may support cell-specific AI / ML for joint uplink Tx beam and uplink TRP beam switch predictions. For example, the UE 115 may perform temporal uplink Tx beam prediction to determine an uplink TRP switch during beam prediction cycles. In such cases, the UE 115 may additionally receive SRIs and / or SRS-RSRPs associated with different uplink TRPs during SRS transmission cycles, which the UE 115 may additionally use as inputs to the AI / ML model. The UE 115 may determine a preferred uplink TRP (or may determine a switch from one uplink TRP to another TRP) during beam prediction cycles, and may transmit an indication of the preferred uplink TRP or an indication of the switch to the network.
[0190] FIG. 7 shows an example of an AI / ML beam management procedure 700 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. In some examples, the AI / ML beam management procedure 700 may support AI / ML beam-related prediction for uplink TRP deployments, and may support communications between a UE 115 and a TRP 102-a, each of which may be examples of corresponding devices described herein.
[0191] The AI / ML beam management procedure 700 may support an AI / ML implementation used across various cells or an AI / ML implementation used in a single cell based on TRP-ID and / or uplink Rx beam ID (e.g., uplink Rx signal ID) indication obtained during training data collection. For example, during training data collection, the UE 115 may receive multiple groups of measurement results for a certain SRS resource set in a given SRS transmission instance. In such examples, each group of measurement results may be associated with an additional identifier of a certain uplink TRP and / or a certain uplink Rx beam.
[0192] The network may provide (e.g., signal) a set of measurements (e.g., SRIs, SRS-RSRPs, or both) for a first TRP. In some examples, the UE 115 may preform uplink beam sweeping via SRSs, and may receive the measurements (which may include uplink TRP IDs signaled together with the SRIs / SRS-RSRPs) for different swept SRSs. In some other examples, the network may provide (e.g., signal) a set of measurements (e.g., SRIs, SRS-RSRPs, or both) for the TRP 102-a. In some examples, the UE 115 may perform uplink beam sweeping via SRSs, and may receive the measurements (which may include uplink Rx beam IDs signaled together with the SRIs / SRS-RSRPs) for different swept SRSs. After training data collection, the UE 115 may train one or more AI / ML algorithms or AI / ML models (e.g., perform offline training) using the data.
[0193] In some implementations, the UE 115 may perform AI / ML inferencing using the one or more trained AI / ML algorithms or trained AI / ML models, and may feedback TRP-ID and / or uplink Rx Beam ID (e.g., uplink signal ID) preference during inference. For example, during inference, the UE 115 may use various type of control signaling (e.g., RRC signaling, MAC-CE signaling, UCI signaling, or any combination thereof) to report a current switch or a predicted future switch towards an indicated Rx beam-ID associated with a respective uplink TRP ID, or towards another uplink TRP ID. In some cases, AI / ML inferencing-based beam switching or TRP switching may be further based on various network-provided measurement results for the SRS resource sets during inference (which may include the corresponding uplink TRP ID (s) and / or the associated uplink Rx beam ID (s) ) .
[0194] In some aspects, the UE 115 may support an AI / ML model that is used in various cells for joint uplink Tx beam and uplink TRP beam switch predictions. For example, the UE 115 may perform uplink Tx beam sweeping via SRSs to determine an uplink TRP switch during beam prediction cycles. The UE 115 may identify a predicted or preferred beam ID (or a beam ID switch) and may indicate the predicted or preferred beam to the network.
[0195] FIG. 8 shows an example of an AI / ML performance monitoring configuration 801 and an AI / ML performance monitoring configuration 802 that support model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. In some examples, the AI / ML performance monitoring configurations 801 and 802 may support AI / ML beam-related prediction for uplink TRP deployments, and may support communications between a UE 115 and a TRP 102-a (and one or more other TRPs) each of which may be examples of corresponding devices described herein.
[0196] In some cases, the UE 115 may support an AI / ML performance monitoring procedure for the one or more AI / ML models to identify how closely the predictions of the one or more AI / ML models match actual measurements, and model updates may be performed based on the performance monitoring. In some examples, the UE 115 may support AI / ML performance monitoring across one or more cells using performance monitoring SRSs for uplink Tx beam prediction monitoring. In some such examples, performance monitoring SRSs (without any downlink reference signals to determine TCI-states or SRS-SpatialRelationInfo) may be scheduled and transmitted based on candidate uplink Tx beams that the UE 115 predicts, and the network may further indicate corresponding SRIs and / or SRS-RSRPs to the UE 115 for verifying the corresponding uplink Tx beam prediction accuracy. In some cases, for a given instance where the performance monitoring SRSs are scheduled, the UE 115 may perform uplink Tx beam prediction without referring to SRIs and / or SRS-RSRPs and the corresponding monitoring SRSs.
[0197] In some implementations, the UE 115 may support UE-based performance monitoring, where the UE may receive network-signaled SRIs and / or SRS-RSRPs on the SRSs, and the UE 115 may compare such network indications with previously determined uplink Tx beam prediction results. Based on the comparison, the UE 115 may determine and / or recommend AI / ML activation, deactivation, or switch actions. Additionally, or alternatively, the UE 115 may support network-based or UE-assisted performance monitoring. For example, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs associated with the SRSs, and the UE 115 may calculate and report raw differences (e.g., for network-based performance monitoring) or statistical differences (e.g., for UE-assisted performance monitoring) between uplink Tx beam prediction results and uplink Tx beam measurement results determined based on the SRIs and / or RSRPs. In such examples, a network entity (such as a gNB) may determine further actions according to such UE performance monitoring reports.
[0198] In performance monitoring configuration 801, the UE 115 may receive performance monitoring SRSs for uplink Rx beam switching (e.g., for performance monitoring for an AI / ML model used in various cells) . In some examples, performance monitoring SRSs may be associated with uplink Tx beams that are respectively identified and / or predicted for different candidate uplink Rx beams associated with an uplink TRP (e.g., TRP 102-a) . In such examples, the performance monitoring SRSs may be scheduled by a network entity so that the UE 115 may verify related UE-side uplink Rx beam prediction accuracy.
[0199] For UE-based performance monitoring, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs associated with SRSs (e.g., SRS 1, SRS 2, SRS 3, SRS 4, and SRS 5) , and the UE 115 may use the received SRIs and / or SRS-RSRPs to derive measurement-based uplink Rx beams, and compare the measurement-based uplink Rx beams with predicted uplink Rx beams (e.g., corresponding to uplink Rx beam 1, uplink Rx beam 2, uplink Rx beam 3, uplink Rx beam 4, and uplink Rx beam 5) , and then determine and / or recommend further AI / ML activation, deactivation, or switch actions based on the comparisons. In some examples, the UE 115 may select a candidate uplink beam (e.g., a preferred uplink Rx beam) using a beam ID for communication with the TRP 102-a, from a set of candidate uplink beams. In some implementations, each candidate uplink Tx beam may be transmitted via a unique SRS resource, and may be received by the TRP 102-a using different uplink Rx beams, respectively.
[0200] For network-based (or UE-assisted) performance monitoring, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs associated with the SRSs, and the UE 115 may use the received SRIs and / or SRS-RSRPs to derive measurement-based uplink Rx beams, and may calculate and report raw differences (for network-based performance monitoring) or statistical differences (for UE-assisted performance monitoring) between the measurement-based uplink Rx beams and the predicted uplink Rx beams (e.g., corresponding uplink Rx beam 1, uplink Rx beam 2, uplink Rx beam 3, uplink Rx beam 4, and uplink Rx beam 5) . The UE 115 may report the determined preferences to the network, and based on the UE reporting, a network entity (such as a gNB) may determine further actions or instructions for the UE 115.
[0201] In performance monitoring configuration 802, the UE 115 may receive performance monitoring SRSs for uplink TRP switching prediction monitoring (e.g., cell-specific AI / ML) . In some examples, performance monitoring SRSs may be associated with uplink Tx beams respectively identified and / or predicted for different candidate uplink TRPs (e.g., uplink TRP 1, uplink TRP 2, uplink TRP 3, or any other quantity of uplink TRPs) . In some such examples, the performance monitoring SRSs may be scheduled by a network entity so that the UE 115 may verify related UE-side uplink TRP prediction accuracies.
[0202] For UE-based performance monitoring, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs associated with SRSs (e.g., SRS 1, SRS 2, and SRS 3) , and the UE 115 may use the received SRIs and / or SRS-RSRPs to derive measurement-based uplink TRP preferences, and compare the measurement-based uplink TRP preferences with predicted uplink TRP preferences (e.g., corresponding to uplink TRP 1, uplink TRP 2, and uplink TRP 3) , and then determine and / or recommend further AI / ML activation, deactivation, or switch actions based on the comparisons. In some examples, the UE 115 may select a candidate uplink TRP (e.g., a preferred uplink TRP) using an TRP ID for communication, from a set of candidate uplink TRPs. In some implementations, each candidate uplink TRP may be associated with an uplink Tx beam transmitted via a unique SRS resource, and may be received by the different candidate TRPs using different uplink Rx beams, respectively.
[0203] For network-based (or UE-assisted) performance monitoring, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs associated with the SRSs, and the UE 115 may use the received SRIs and / or SRS-RSRPs to derive measurement-based uplink TRP preferences, and may calculate and report raw differences (for network-based performance monitoring) or statistical differences (for UE-assisted performance monitoring) between the measurement-based TRP preferences and the predicted TRP preferences (e.g., corresponding uplink TRP 1, uplink TRP 2, and uplink TRP 3) . The UE 115 may report the determined preferences to the network, and based on the UE reporting, a network entity (such as a gNB) may determine further actions or instructions for the UE 115.
[0204] FIG. 9 shows an example of an AI / ML parameter consistency evaluation flow 900 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. In some examples, the AI / ML parameter consistency evaluation flow 900 may support AI / ML beam-related prediction for uplink TRP deployments, and may be implemented at or by a UE 115, which may be an example of a UE 115 as described herein.
[0205] In some implementations, the UE 115 may support various procedures that allow for the determination of consistency for AI / ML input parameters (e.g., SRS resource or resource Set numbers, together with network-indicated measurement types for AI / ML inputs) across training and inference. To enable more efficient UE-side beam prediction deployment, the network and the UE 115 may “pre-negotiate” various parameters such as how many SRS resources or SRS resource sets may be supported, which SRS transmission periodicities may be supported, what kind of SRS measurement results are supported for AI / ML inference (e.g., whether SRIs are supported, whether SRS-RSRPs are supported, or both) , or any combination thereof. Such pre-negotiated parameters may be consistent across training and inference, for a given associated-ID.
[0206] During training data collection, and for a given applicable associated-ID, the network may signal one or more candidate SRS resource set numbers, the SRS resource numbers for the respective SRS resource numbers, the types of network indicated measurement results, or any combination thereof, to be considered by the UE 115 for AI / ML inference. In some examples, candidate types of network-indicated measurement results (as AI / ML inputs) may include a threshold quantity (e.g., a minimum number) of SRS resource IDs addressed per SRS resource set by network indication, whether SRS-RSRPs on such SRS resource IDs are to be used for inferencing, among other measurement results, or a combination of different measurement results. In some aspects, the SRS scheduled and the network-indicated measurement results, may indicate support of such network-signaled candidates or UE preferred candidates (e.g., if network supported candidate SRS resource number per resource set is at least 4, the actual number of SRS resources may be greater than or equal to 4) .
[0207] In some aspects, the UE 115 may assume consistency across AI / ML training and inference. For example, the UE 115 may be expected to support or report UE capabilities for AI / ML inference regarding the same associated-ID that the UE 115 used for trained an AI / ML algorithm, for the network indicated candidates. That is, if the network schedules the UE 115 with 1 SRS resource set during training data collection, and the 1 SRS resource set was indicated as the only supported candidate, the UE 115 may not expect to be scheduled with 2 separate SRS resource sets for wide-to-narrow beam predictions. Additionally, or alternatively, if an SRS-RSRP based measurement results indication was considered as the only supported candidate for AI / ML inputs, the UE 115 may not expect to be (only) signaled with SRIs during inference on the transmitted SRS resource sets for AI / ML inputs.
[0208] FIG. 10 shows an example of various training procedures for a wireless communications system 1000 and which also supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. For example, the wireless communications system 1000 may be an example of a wireless communications system supporting uplink TRP deployment. In some aspects, the wireless communications system may support communications between a UE 115, a network entity 105, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0209] In some aspects, the UE 115 may support AI / ML model training using SRSs transmitted towards one or more UL TRPs (e.g., TRP 102-a and TRP 102-b) . For example, the UE 115 may be scheduled with one or more SRS resource sets without any downlink reference signals signaled for obtaining TCI-state’s or SpatialRelationInfo’s (e.g., based on the one or more UL TRPs lacking downlink reference signals) , and the UE 115 may receive network-indicated measurement results on the respective SRS resource sets. In such cases, the UE 115 may be expected to train one or more AI / ML algorithms based on the SRS resource sets and the network-indicated measurement results, and predict uplink Tx beams used for transmitting PUxCH (towards UL-only TRP (s) ) in spatial and / or temporal domains.
[0210] In some examples, the network-indicated measurement results on a particular SRS resource set may include the “Top K” SRS resource ID (s) in the SRS resource set (e.g., the SRS resource IDs having the strongest relative SRS-RSRP strengths or with the highest relative probability of being the “Top 1” or “Top K” SRS resources) , explicit SRS-RSRP (s) associated with the indicated “Top K” SRS resource ID (s) , explicit probabilities being the “Top1” or “Top K” SRS resource (s) associated with the indicated “Top K” SRS resource ID (s) , or any combination thereof.
[0211] In some examples, the UE 115 may perform temporal prediction using one or more AI / ML models. For example, a single SRS resource set may be transmitted based on the same candidate uplink Tx beams across training &inference. In such examples, the UE 115 may use the one or more AI / ML algorithms to predict future uplink Tx beams out of the candidate uplink Tx beams based on historically transmitted SRSs and the network-indicated measurement results associated with the SRSs. In some examples, the UE 115 may receive the network-indicated measurement results during training data collection, including SRS-RSRPs associated with every transmitted SRS resources for a given SRS resource set transmission instance.
[0212] In some examples, the UE 115 may perform wide-to-narrow spatial prediction using the one or more AI / ML models. For example, a single SRS resource set may be transmitted based on a set of candidate uplink Tx beams during training data collection. In such examples, the UE 115 may use the AI / ML algorithm to predict an uplink Tx beam from the set of candidate beams, based on another SRS resource set transmitted via a relatively smaller (e.g., much smaller) quantity of uplink Tx beams (from the candidate beams) together with NW-indicated measurement results associated with the uplink Tx beams. In some cases, the UE 115 may receive the network indicated measurement results during training data collection, and the network indicated measurement results may include SRS-RSRPs associated with every transmitted SRS resource for a given SRS resource set transmission instance.
[0213] In some examples, the UE 115 may use two SRS resource sets for wide-to-narrow beam prediction. For example, the quantity of SRS resource sets scheduled for the UE 115 may be at least two, where a {S1, S2} quantity of SRS resources may include the {1st, 2nd} SRS resource set such that S1<<S2, and the values of K with respect to the “Top K” SRS resources associated with the {1st, 2nd} SRS resource set may be {K1, K2} such that K1>K2. For example, if K1=S1 and K2=1, the network-indicated measurement results on the first SRS resource set includes at least SRS-RSRPs associated with every transmitted SRS resource for a given SRS resource set transmission instance, and the network indicated measurement results on the second SRS resource set may only include a single SRS resource ID associated with the strongest SRS-RSRP (e.g., highest SRS-RSRP value) for a given SRS resource set transmission instance. For example, training data collection for a classifier predicting the “Top 1” narrow uplink Tx beam among S2>>S1 candidate beams, based on measurements of S1 wider uplink Tx beams.
[0214] In some implementations, the UE 115 may receive one or more associated IDs (e.g., an associated ID for each uplink TRP and / or uplink Rx beams associated with an uplink TRP. The UE 115 may also receive one or more associated-IDs regarding the uplink TRPs and / or regarding the one or more uplink Rx beams associated with the uplink TRPs. During training data collection, the UE 115 may receive a network-indicated associated ID for one or more signaled measurement results on one or more transmitted SRS resource sets, where the network indicated associated ID may correspond to an uplink TRP and / or the uplink Rx beams associated with the uplink TRP. In such case, the network indicated measurement results on the certain UE transmitted SRS resource set (s) may be derived based on receiving the one or more SRS resource sets via the uplink TRP and / or the uplink Rx beams associated with the uplink TRP.
[0215] Additionally, or alternatively, during AI / ML inference, the UE 115 may receive a network indicated associated-ID for network signaled measurement results on one or more UE-transmitted SRS resource set (s) , where the network indicated associated ID may correspond to an uplink TRP and / or the uplink Rx beams associated with the uplink TRP. In such cases, the network indicated measurement results associated with the UE-transmitted SRS resource set (s) may be derived based on receiving the SRS resource set (s) via the uplink TRP and / or the uplink Rx beams associated with the uplink TRP.
[0216] In some examples, the UE 115 may identify (e.g., ensure) consistency of network side additional conditions associated with uplink TRPs using associated IDs. For example, if the UE 115 identifies a same associated-ID during inference as the associated-ID used during training data collection, the UE 115 may assume certain network side additional conditions are the same as that used during model training data collection. In some examples, the network-side additional conditions may include spatial Rx filters. For example, the use of a same associated-ID during training and inference may indicate consistent spatial Rx filters used across training and inference. In some examples, the spatial Rx filters may be defined by beam pointing directions and beam widths, referring to the center of the antenna panel associated with the uplink TRPs. Additionally, or alternatively, the network additional conditions may include a quantity and / or ordering of the uplink Rx beam-IDs within the same applicable uplink TRP, a quantity of uplink Rx beams per uplink TRP, SRS resource or resource set quantities, SRS periodicities, network indicated SRS measurement result types, or any combination thereof.
[0217] In some implementations, the UE 115 may support training data collection for TRP switch predictions, uplink Rx-beam switch predictions, or both. For example, during training data collection, the UE 115 may receive one or more groups of network-indicated measurement results for an SRS resource set in a given SRS transmission instance, where each group of network indicated measurement results may be associated with an additional identifier of a certain uplink TRP and / or a certain uplink Rx beam.
[0218] To perform uplink Rx beam switch predictions using AI / ML supported in various cells, the AI / ML algorithm trained by the UE 115 may predict one or more (current and / or future) preferred uplink Rx beam identifiers from the total set of candidate uplink Rx beams, based on network indicated measurement results associated with a restricted quantity of uplink Rx beam identifiers from the total set of candidate uplink Rx beams. Such AI / ML techniques may be based on the same associated ID (for uplink Rx beams) being identified across training and inference.
[0219] To perform uplink Rx beam switch predictions using cell-specific AI / ML, the AI / ML algorithm trained by the UE 115 may predict one or more (current and / or future) preferred uplink TRP identifiers out of a total set of candidate uplink TRP identifiers, based on network indicated measurement results associated with a restricted quantity of uplink TRP identifiers from the total set of candidate uplink TRP identifiers. Such cell-specific AI / ML techniques may be based on the same associated ID (for uplink Rx beams) being identified across training and inference.
[0220] In some implementations, the UE 115 may use downlink reference signals as “signatures” for training data collection. For example, during training data collection, the UE 115 may receive network information regarding a quantity of SSB resource identifiers and / or CSI-RS resource identifiers, such that measurement results on such SSB resource identifiers and / or CSI-RS resources may be used as additional AI / ML inputs for the AI / ML algorithms trained by the UE 115. In some examples, downlink Tx beam identifiers associated with the SSB resources and / or CSI-RS resources may be associated with a quantity of downlink Tx beam IDs and respective Tx spatial filters, which may be mapped to the respective SSB resources and / or CSI-RS resources that the network may explicitly indicate or implicitly indicate. For explicit indication, the network may utilize an explicit mapping such that each network indicate SSB resource and / or CSI-RS resource may be signaled with a downlink Tx beam ID. For implicit indication, the network may utilize an implicit mapping, such that the total quantity of downlink Tx beams may be the same as the total quantity of network indicated SSB resources and / or CSI-RS resources. In such cases, the SSB resources and / or the CSI-RS resources may be indicated via one or more SSB resource sets and / or CSI-RS resource sets, and mappings from the downlink Tx beam IDs to the SSB resources and / or the CSI-RS resources may be based on ordering of the SSB resource set IDs and / or the CSI-RS resource set IDs, and the resource entry-IDs within the respective resource sets (e.g., first via set IDs second via resource entry IDs) .
[0221] In some cases, implicit and explicit mappings may be based on the AI / ML algorithm used for inference being for the same cell-ID that is used during training data collection. For example, the UE 115 may expect to be indicated with implicit and or explicit information during the AI / ML algorithm inference procedures, such that mapping orders from the downlink Tx beams to the SSB resource sets and / or the CSI-RS resource sets may also be explicitly or implicitly identified. Additionally, or alternatively, the UE 115 may determine that, for the same cell-ID, the same Tx spatial filter may be used for SSB and / or CSI-RS resources associated with the same downlink Tx beam ID, across training and inference.
[0222] In some implementations, the network may provide various location information and / or pointing-direction assistance information during training data collection. For example, during training data collection, the UE 115 may receive network-signaled assistance information, which may indicate a location of a downlink TRP, one or more locations associated with one or more involved uplink TRPs, one or more beam pointing directions of the respective downlink Tx spatial filters associated with downlink reference signaling provided by the network, or any combination thereof. In some aspects, the UE 115 may determine how to use such assistance information, for example, the UE 115 may use the assistance information as inputs to the one or more AI / ML models, or to derive AI / ML inputs for one or more AI / ML models. In some aspects, the UE 115 may transmit one or more messages which may include requests regarding a type of information that the network is to include in the assistance information, and the network may transmit the requested assistance information during training data collection. Additionally, or alternatively, the UE 115 may transmit one or more UE capability reports which indicate different information that the UE 115 may support for AI / ML training, including information that the UE 115 may use as AI / ML inputs.
[0223] As described herein, the UE 115 may identify or support consistency of AI / ML input parameters (e.g., SRS resources and SRS resource set quantities, together with network-indicated measurement types for AI / ML inputs) across both training and inference stages of the one or more AI / ML modes. In some examples, during training data collection, the network entity 105 may indicate candidate SRS resources and / or SRS resource set quantities, together with candidate types of network-indicated measurement results, to be considered during inference. In such examples, during training data collection and for a given applicable associated-ID the network entity 105 may signal the candidate SRS resource set quantities, and / or the SRS resource quantities for the respective SRS resource set quantities, and / or the types of network-indicated measurement results, to be considered by the UE 115 for AI / ML inference. In some aspects, the indication of the same associated ID for training and inference may indicate the same network additional conditions across training and inference.
[0224] In some implementations, “candidate” information (e.g., network-indicated measurement results as AI / ML inputs including a threshold quantity (e.g., a minimum number) of SRS resource IDs addressed per SRS resource set, whether SRS-RSRPs on such SRS resource IDs are needed, among other examples) may be signaled to the UE 115 before over the air training data collection is actually triggered, and the UE 115 may further report down-selected preferences out of such candidate information (e.g., to lower signaling overhead, resource overhead, power expenditure, or any combination thereof, for such data collection) . In some aspects, the scheduled SRS and the network-indicated measurement results, may indicate or guarantee such network-signaled candidates or UE-preferred candidates. For example, if a network-supported candidate SRS resource quantity per SRS resource set is at least 4, the actual number of SRS resources may be greater than or equal to 4.
[0225] In some implementations, the UE 115 may be expected to support or report UE capabilities for AI / ML inference for the same associated-ID that the UE 115 uses to collect training data and train the AI / ML algorithm (e.g., the UE 115 identifies consistency of associated IDs used across training and inference) . For example, if only one SRS resource set was scheduled during training data collection (and also indicated by the network as the supported candidate) , the UE 115 may not expect to be scheduled with two separate SRS resource sets for wide-to-narrow beam prediction. Additionally, or alternatively, if an SRS-RSRP based measurement results indication was considered as the only supported candidate for AI / ML inputs, the UE 115 may not expect to be (only) signaled with SRIs during inference on the transmitted SRS resource sets for AI / ML inputs.
[0226] FIG. 11 shows an example of various inferencing procedures supported by a wireless communications system 1100 and which also supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. For example, the wireless communications system 1100 may be an example of a wireless communications system supporting uplink TRP deployment. In some aspects, the wireless communications system may support communications between a UE 115, a network entity 105, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0227] In some implementations, the UE 115 may support AI / ML model inferencing for beam prediction (e.g., prediction of one or more parameters or characteristics associated with uplink or downlink transmission) for communication with uplink mTRPs. In some aspects, the UE 115 may utilize identification of associated IDs (associated with different uplink TRPs such as TRP 102-a and TRP 102-b, or associated with different uplink Rx beams associated with the different uplink TRPs) and related UE capabilities and / or recommendations during inferencing. For example, the UE 115 may identify one or more associated IDs via network signaling (e.g., from remaining minimum system information (RMSI) signaling or from ServCell-specific RRC configurations) , for beam prediction on uplink mTRPs located in a cell or a cell-group. For one or more such associated-IDs, the UE 115 may report related UE capabilities and / or recommendations to the network entity 105.
[0228] In some aspects, if the UE 115 identifies the same associated-ID for AI / ML inferencing as the associated-ID identified or used during training data collection, the UE 115 may consider related network side additional conditions considered for AI / ML inference to be the same as network side additional conditions considered during training data collection. Some such examples of network side additional conditions may include spatial Rx filters associated with the uplink TRPs and / or uplink Rx beam (s) associated with the uplink TRPs, a quantity of uplink Rx beams per uplink TRP, a quantity of SRS resources or a quantity of SRS resource sets, one or more SRS periodicities or network indicated SRS measurement result types, or any combination thereof. Some examples of UE capabilities that the UE 115 may report may include an indication of a threshold (e.g., minimum, maximum) quantity of candidate SRS resource sets or SRS resources per SRS resource set supported by the UE 115, a minimum quantity of candidate SRS resource sets and / or a supported quantity of SRS resources within the respective SRS resource sets supported during inference, or any combination thereof.
[0229] For some implementations such as for spatial beam prediction inference, the UE 115 may support two SRS resource sets (e.g., one for transmitting wider uplink Tx beams relatively more frequently with fewer SRSs, and one for transmitting relatively narrower uplink Tx beams relatively less frequently with a relatively greater quantity of SRSs) . Additionally, or alternatively, for temporal beam prediction inference, the UE 115 may support scheduling of a single SRS resource set with a restricted number of SRSs.
[0230] In some examples, the threshold SRS resource transmission periodicities (e.g., the minimum or maximum candidate SRS resource transmission periodicities) may indicate a threshold periodicity of the SRS resource sets that the UE 115 may support during inference, and such periodicity capabilities may be reported as UE capabilities. Additionally, or alternatively, when the UE 115 transmits UE capability reporting, the UE 115 may also indicate various measurement result types, including a minimum supported quantity of SRS resource IDs addressed per SRS resource set by network indication, whether SRS-RSRPs on such SRS resource IDs are requested by the UE 115, among other examples.
[0231] As long as UE capabilities are not violated, the supported SRS resource quantities, the supported SRS resource set quantities, the supported SRS periodicities, or any combination thereof that are scheduled by the network, may be variable and subject to network-side implementation choices. In some examples, the UE 115 may dynamically update indicate changes to the one or more UE capabilities using RRC signaling, MAC-CE signaling, UCI signaling (e.g., depending on UEs real-time movement, rotation, or observed speed, among other changes) , as a recommendation to the network, and the network may or may not follow such dynamic updates based on whether the initial UE capabilities are not violated.
[0232] In some implementations, signaling of UE capabilities and / or recommendations may be negotiated between the network entity 105 and the UE 115, as applicable inference functionalities or non-applicable inference functionalities, via RRC signaling (e.g., RRCReconfiguration signaling) . For example, the network entity 105 may output one or more candidate parameters (or sets of parameters) to UE 115 that comply with the indicated UE capabilities and / or recommendations, and the UE 115 may report the applicability of the different candidate parameters or sets of parameters, after UE capability reporting. (e.g., the UE 115 may directly report applicable parameters (or sets of parameters) associated with relevant capabilities and / or recommendations after UE capability reporting) .
[0233] In some implementations, the UE 115 may support two SRS resource sets for wide-to-narrow beam prediction during inference. For example, during inference, the UE 115 may receive a scheduling that includes at least two SRS resource sets, where a {S1, S2} quantity of SRS resources and a periodicity of {P1, P2} includes the {1st, 2nd} SRS resource set, respectively, such that S1<<S2 and P1<<P2. The UE 115 may also receive, from the network, SRIs and / or SRS-RSRPs separately that are associated with the at least two SRS resource sets. In some examples, the UE 115 may use the {1st, 2nd} SRS resource sets to transmit wider or narrower uplink Tx beams, and the UE 115 may use measurement results indicated by network on the second SRS resource set as additional AI / ML inputs to improve wide-to-narrow prediction accuracy.
[0234] In some implementations, the UE 115 may support transmission of an uplink Rx beam switch report and / or an uplink TRP switch prediction report. For example, during inference, the UE 115 may use RRC signaling, MAC-CE signaling, UCI signaling, or any combination thereof, to report predicted current or future switches towards a respective uplink Rx beam-ID associated with a respective uplink TRP ID (or towards another uplink TRP ID) . Such switching between uplink TRPs and / or beams may also be based on network indicated measurement results associated with the one or more SRS resource sets, which may also include the corresponding uplink TRP ID (s) and / or the associated uplink Rx beam ID (s) . For example, the network entity 105 may indicate multiple groups of measurement results for a single SRS resource set, with each group being associated with a unique uplink TRP or a unique UL-Rx beam-ID, or both. In such cases, the UE 115 may interpret the groups of measurement results as being derived based on the associated uplink TRP and / or the uplink beam-ID. In some such examples, the network entity 105 may indicate which uplink Rx beam ID and / or UL-only TRP ID is selected as the “serving” uplink Rx beam or serving uplink TRP for the UE 115. In some aspects, the UE 115 may assume that quantity and ordering of the uplink Rx beam IDs are the same, as long as it the associated-ID identified during inference as the same as the associated ID identified during training data collection.
[0235] In some implementations, the UE 115 may support use of downlink reference signals as “signatures” for optional AI / ML inputs. For example, during inference, the UE 115 may receive one or more network indications regarding a quantity of SSB resource identifiers, CSI-RS resource identifiers, or both, such that measurement results on such SSB resources, CSI-RS resources, or both, may be used as additional AI / ML inputs for inference. In some aspects, the use of the pretrained AI / ML algorithm may be for the same cell-ID as that considered during training data collection and for inference, and mapping orders from the downlink Tx beams to the downlink reference signals may be explicitly or implicitly identified across training &inference, based on the implicit mappings and explicit mappings described herein. In some aspects, the UE 115 may assume the same Tx spatial filter is associated with the SSB resources and / or the CSI-RS resources associated with the same downlink Tx beam ID across training & inference within the same cell-ID.
[0236] In some implementations, the UE 115 may receive location information and / or beam pointing-direction assistance information from the network. For example, during inference, the UE 115 may receive network-signaled assistance information, which may include a location of one or more downlink TRPs, one or more locations of the involved uplink TRP (s) , beam pointing directions of the respective downlink Tx spatial filters associated with the one or more downlink reference signals, or any combination thereof. In some cases, the UE 115 may determine or decide whether or how to use the assistance information as AI / ML inputs for the one or more AI / ML models for beam prediction. Additionally, or alternatively, the UE 115 may use the assistance information to derive AI / ML inputs. In some cases, the UE 115 may transmit one or more requests for assistance information from the network during inference.
[0237] FIG. 12 shows an example of various AI / ML performance management procedures supported by a wireless communications system 1200 and which also supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. For example, the wireless communications system 1100 may be an example of a wireless communications system supporting uplink TRP deployment. In some aspects, the wireless communications system may support communications between a UE 115, a network entity 105, a TRP 102-a, and a TRP 102-b, each of which may be examples of corresponding devices described herein.
[0238] In some implementations, the UE 115 may support performance monitoring procedures to monitor ongoing performance of prediction accuracy for the one or more AI / ML models. For an AI / ML model that is supported for use in various cells, the UE 115 may receive one or more performance monitoring SRSs for uplink Tx beam prediction monitoring. For example, the UE 115 may receive a scheduling which includes an indication of one or more performance monitoring SRS resource set (s) without any TCI-states or SRS-SpatialRelationInfos, and the network may also indicate corresponding SRIs and / or SRS-RSRPs to the UE 115, for verifying the corresponding uplink Tx beam prediction accuracy.
[0239] When the performance monitoring SRSs are scheduled, the UE 115 may be expected to perform uplink Tx beam prediction (without referring to SRIs and / or SRS-RSRPs associated with the corresponding SRSs) using one or more AI / ML models using UE-based performance monitoring processes. The UE 115 may receive network-signaled SRIs and / or SRS-RSRPs on the SRSs, and the UE 115 may use the network indicated SRIs and / or SRS-RSRPs to compare with the uplink Tx beam prediction results. Based on the prediction, the UE 115 may determine and recommend further AI / ML activation, deactivation. or switching actions.
[0240] In some examples, the UE 115 may support network-based or UE-assisted performance monitoring. For example, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs on the SRSs, and the UE 115 may calculate and report raw differences (for network-based performance monitoring) or statistical differences (for UE-assisted performance monitoring) between the Tx beam prediction results and uplink Tx beam measurement results determined based on the SRIs and / or SRS-RSRPs. After the UE 115 determines the differences between the predicted and actual beam measurement results, the network entity 105 may determine further actions, and may instruct the UE 115 to perform AI / ML activation, deactivation. or switching actions.
[0241] In some aspects, the UE 115 may identify one or more “virtual” TCI-states or SRS-SpatialRelationInfos via uplink Rx beam-IDs or uplink TRP IDs. Such virtual” TCI-states or SRS-SpatialRelationInfos may be based on the SRS resource set (s) which may be respectively indicated with certain uplink Rx beam-ID (s) and / or an associated UL-only TRP ID, such that the UE 115 may be expected to use the predicted uplink Tx beams for the corresponding uplink Rx beam-ID (s) and / or the uplink TRP ID to transmit the corresponding SRS resources. In some aspects, the virtual” TCI-states or “virtual” SRS-SpatialRelationInfos may be additionally, or alternatively, referred to as virtual SSBs, virtual CSI-RSs (associated with uplink TRPs) , virtual downlink Tx spatial filters (associated with uplink TRPs) . Additionally, or alternatively, “uplink-Rx beam IDs” may be referred to as or may be equivalent to “virtual downlink Tx beam IDs, ” “virtual SSB IDs, ” “virtual CSI-RS IDs, ” or “virtual downlink Tx spatial filter IDs, ” when referring to beam prediction for uplink TRPs.
[0242] In some implementations, the UE 115 may support performance monitoring SRSs for uplink Rx beam switching (e.g., for an AI / ML model supported for use across various cells) and uplink TRP switching (e.g., for cell-specific AI / ML) prediction monitoring. For example, the UE 115 may be scheduled with one or more performance monitoring SRS resource sets with multiple “virtual” TCI-states or SRS-SpatialRelationInfos associated with uplink Rx beam-IDs and / or uplink TRP IDs, and the network may further indicate corresponding SRIs and / or SRS-RSRPs to the UE 115 for verifying the corresponding UE-side uplink Rx beam or uplink TRP switch prediction accuracy. In order to use the “virtual” TCI-states or virtual SRS-SpatialRelationInfos, the UE 115 may be expected to use the predicted uplink Tx beams for the corresponding uplink Rx beam-ID (s) and / or the UL-only TRP ID to transmit the corresponding SRS resources.
[0243] In some examples, the UE 115 may support UE-based performance monitoring. For example, the UE 115 may receive one or more network-signaled SRIs and / or SRS-RSRPs associated with the SRSs, and the UE 115 may use the received SRIs and / or SRS-RSRPs to derive a measurement-based uplink Rx beam and / or uplink TRP preferences. The UE 115 may then compare the measurement-based uplink Rx beam and / or the uplink TRP preferences with predicted uplink Rx beams and / or predicted uplink TRP preferences, and may determine and / or recommend additional AI / ML activation, deactivation, or switching actions based on the comparisons.
[0244] In some examples, the UE 115 may support network-based performance monitoring or UE-assisted performance monitoring. For example, the UE 115 may receive network-signaled SRIs and / or SRS-RSRPs associated with the SRSs, and the UE 115 may use the received SRIs and / or SRS-RSRPs to derive measurement-based uplink Rx beams and / or measurement-based uplink TRP preferences. The UE 115 may calculate and report raw differences (for network-based performance monitoring) or statistical differences (e.g., for UE-assisted performance monitoring) between such measurements based on comparing the measurement-based uplink Rx beam and / or measurement-based uplink TRP preferences with the predicted uplink Rx beam and / or predicted uplink TRP preferences. Based on the comparisons and the reported differences, the network entity 105 may determine further recommended actions (e.g., additional AI / ML activation, deactivation, or switching actions) and may report the recommended actions to the UE 115.
[0245] FIG. 13 shows a block diagram 1300 of a device 1305 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The device 1305 may be an example of aspects of a UE 115 as described herein. The device 1305 may include a receiver 1310, a transmitter 1315, and a communications manager 1320. The device 1305, or one or more components of the device 1305 (e.g., the receiver 1310, the transmitter 1315, the communications manager 1320) , may include at least one processor, which may be coupled with at least one memory, to, individually or collectively, support or enable the described techniques. Each of these components may be in communication with one another (e.g., via one or more buses) .
[0246] The receiver 1310 may provide a means for receiving information such as packets, user data, control information, or any combination thereof associated with various information channels (e.g., control channels, data channels, information channels related to model training techniques for channel characteristic prediction procedures) . Information may be passed on to other components of the device 1305. The receiver 1310 may utilize a single antenna or a set of multiple antennas.
[0247] The transmitter 1315 may provide a means for transmitting signals generated by other components of the device 1305. For example, the transmitter 1315 may transmit information such as packets, user data, control information, or any combination thereof associated with various information channels (e.g., control channels, data channels, information channels related to model training techniques for channel characteristic prediction procedures) . In some examples, the transmitter 1315 may be co-located with a receiver 1310 in a transceiver module. The transmitter 1315 may utilize a single antenna or a set of multiple antennas.
[0248] The communications manager 1320, the receiver 1310, the transmitter 1315, or various combinations or components thereof may be examples of means for performing various aspects of model training techniques for channel characteristic prediction procedures as described herein. For example, the communications manager 1320, the receiver 1310, the transmitter 1315, or various combinations or components thereof may be capable of performing one or more of the functions described herein.
[0249] In some examples, the communications manager 1320, the receiver 1310, the transmitter 1315, or various combinations or components thereof may be implemented in hardware (e.g., in communications management circuitry) . The hardware may include at least one of a processor, a digital signal processor (DSP) , a central processing unit (CPU) , an application-specific integrated circuit (ASIC) , a field-programmable gate array (FPGA) or other programmable logic device, a microcontroller, discrete gate or transistor logic, discrete hardware components, or any combination thereof configured as or otherwise supporting, individually or collectively, a means for performing the functions described in the present disclosure. In some examples, at least one processor and at least one memory coupled with the at least one processor may be configured to perform one or more of the functions described herein (e.g., by one or more processors, individually or collectively, executing instructions stored in the at least one memory) .
[0250] Additionally, or alternatively, the communications manager 1320, the receiver 1310, the transmitter 1315, or various combinations or components thereof may be implemented in code (e.g., as communications management software or firmware) executed by at least one processor (e.g., referred to as a processor-executable code) . If implemented in code executed by at least one processor, the functions of the communications manager 1320, the receiver 1310, the transmitter 1315, or various combinations or components thereof may be performed by a general-purpose processor, a DSP, a CPU, an ASIC, an FPGA, a microcontroller, or any combination of these or other programmable logic devices (e.g., configured as or otherwise supporting, individually or collectively, a means for performing the functions described in the present disclosure) .
[0251] In some examples, the communications manager 1320 may be configured to perform various operations (e.g., receiving, obtaining, monitoring, outputting, transmitting) using or otherwise in cooperation with the receiver 1310, the transmitter 1315, or both. For example, the communications manager 1320 may receive information from the receiver 1310, send information to the transmitter 1315, or be integrated in combination with the receiver 1310, the transmitter 1315, or both to obtain information, output information, or perform various other operations as described herein.
[0252] The communications manager 1320 may support wireless communications in accordance with examples as disclosed herein. For example, the communications manager 1320 is capable of, configured to, or operable to support a means for receiving one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The communications manager 1320 is capable of, configured to, or operable to support a means for receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The communications manager 1320 is capable of, configured to, or operable to support a means for training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results. The communications manager 1320 is capable of, configured to, or operable to support a means for obtaining one or more predicted uplink channel characteristics associated with uplink transmission based on training the one or more machine learning models.
[0253] By including or configuring the communications manager 1320 in accordance with examples as described herein, the device 1305 (e.g., at least one processor controlling or otherwise coupled with the receiver 1310, the transmitter 1315, the communications manager 1320, or a combination thereof) may support techniques for reduced processing, reduced latency, reduced power consumption, more efficient utilization of communication resources, including for SRS and associated resources.
[0254] FIG. 14 shows a block diagram 1400 of a device 1405 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The device 1405 may be an example of aspects of a device 1305 or a UE 115 as described herein. The device 1405 may include a receiver 1410, a transmitter 1415, and a communications manager 1420. The device 1405, or one or more components of the device 1405 (e.g., the receiver 1410, the transmitter 1415, the communications manager 1420) , may include at least one processor, which may be coupled with at least one memory, to support the described techniques. Each of these components may be in communication with one another (e.g., via one or more buses) .
[0255] The receiver 1410 may provide a means for receiving information such as packets, user data, control information, or any combination thereof associated with various information channels (e.g., control channels, data channels, information channels related to model training techniques for channel characteristic prediction procedures) . Information may be passed on to other components of the device 1405. The receiver 1410 may utilize a single antenna or a set of multiple antennas.
[0256] The transmitter 1415 may provide a means for transmitting signals generated by other components of the device 1405. For example, the transmitter 1415 may transmit information such as packets, user data, control information, or any combination thereof associated with various information channels (e.g., control channels, data channels, information channels related to model training techniques for channel characteristic prediction procedures) . In some examples, the transmitter 1415 may be co-located with a receiver 1410 in a transceiver module. The transmitter 1415 may utilize a single antenna or a set of multiple antennas.
[0257] The device 1405, or various components thereof, may be an example of means for performing various aspects of model training techniques for channel characteristic prediction procedures as described herein. For example, the communications manager 1420 may include an SRS scheduling component 1425, a measurement evaluation component 1430, an AI / ML training component 1435, an AI / ML inferencing component 1440, or any combination thereof. The communications manager 1420 may be an example of aspects of a communications manager 1320 as described herein. In some examples, the communications manager 1420, or various components thereof, may be configured to perform various operations (e.g., receiving, obtaining, monitoring, outputting, transmitting) using or otherwise in cooperation with the receiver 1410, the transmitter 1415, or both. For example, the communications manager 1420 may receive information from the receiver 1410, send information to the transmitter 1415, or be integrated in combination with the receiver 1410, the transmitter 1415, or both to obtain information, output information, or perform various other operations as described herein.
[0258] The communications manager 1420 may support wireless communications in accordance with examples as disclosed herein. The SRS scheduling component 1425 is capable of, configured to, or operable to support a means for receiving one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The measurement evaluation component 1430 is capable of, configured to, or operable to support a means for receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The AI / ML training component 1435 is capable of, configured to, or operable to support a means for training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results. The AI / ML inferencing component 1440 is capable of, configured to, or operable to support a means for obtaining one or more predicted uplink channel characteristics associated with uplink transmission based on training the one or more machine learning models.
[0259] FIG. 15 shows a block diagram 1500 of a communications manager 1520 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The communications manager 1520 may be an example of aspects of a communications manager 1320, a communications manager 1420, or both, as described herein. The communications manager 1520, or various components thereof, may be an example of means for performing various aspects of model training techniques for channel characteristic prediction procedures as described herein. For example, the communications manager 1520 may include an SRS scheduling component 1525, a measurement evaluation component 1530, an AI / ML training component 1535, an AI / ML inferencing component 1540, or any combination thereof. Each of these components, or components or subcomponents thereof (e.g., one or more processors, one or more memories) , may communicate, directly or indirectly, with one another (e.g., via one or more buses) .
[0260] The communications manager 1520 may support wireless communications in accordance with examples as disclosed herein. The SRS scheduling component 1525 is capable of, configured to, or operable to support a means for receiving one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The measurement evaluation component 1530 is capable of, configured to, or operable to support a means for receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The AI / ML training component 1535 is capable of, configured to, or operable to support a means for training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results. The AI / ML inferencing component 1540 is capable of, configured to, or operable to support a means for obtaining one or more predicted uplink channel characteristics associated with uplink transmission based on training the one or more machine learning models.
[0261] In some examples, one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest RSRP or a highest selection probability, one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set. or one or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.
[0262] In some examples, to support obtaining the one or more predicted uplink channel characteristics, the AI / ML inferencing component 1540 is capable of, configured to, or operable to support a means for obtaining the one or more temporal predictions based on the training of the one or more machine learning models using one or more previously received SRS resource sets and corresponding sets of previously received measurement results, where the corresponding sets of previously received measurement results include one or more RSRP measurements associated with each SRS resource of the one or more previously received SRS resource sets.
[0263] In some examples, to support obtaining the one or more predicted uplink channel characteristics, the AI / ML inferencing component 1540 is capable of, configured to, or operable to support a means for obtaining the one or more spatial predictions using a set of candidate channel characteristics obtained during the training of the one or more machine learning models and corresponding one or more measurement results, where corresponding sets of previously received measurement results include one or more RSRP measurements associated with each SRS resource of one or more previously received SRS resource sets, and where the one or more spatial predictions indicate a quantity of channel characteristics that is less than a quantity of the set of candidate channel characteristics.
[0264] In some examples, the one or more SRS resource sets include at least a first SRS resource set and a second SRS resource set. In some examples, the set of measurement results corresponding to the first SRS resource set includes a set of RSRP measurements for each SRS resource included in the first SRS resource set, and the set of measurement results corresponding to the second SRS resource set includes a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.
[0265] In some examples, the AI / ML training component 1535 is capable of, configured to, or operable to support a means for receiving signaling including an associated identifier (ID) that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink TRP, where the set of measurement results are based on one or more SRS resource sets.
[0266] In some examples, consistency of one or more network conditions between the training of the one or more machine learning models and an inference phase of the one or more machine learning models is based on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase. In some examples, a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase; a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics; a quantity of SRS resources or SRS resource sets; a periodicity of the one or more SRS resource sets; or one or more types of measurements included in the set of measurement results.
[0267] In some examples, to support receiving the set of measurement results, the measurement evaluation component 1530 is capable of, configured to, or operable to support a means for receiving an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, where a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink TRP, to an uplink channel characteristic, or both.
[0268] In some examples, the AI / ML inferencing component 1540 is capable of, configured to, or operable to support a means for predicting, using the one or more machine learning models, a set of uplink TRP identifiers for future selection, a set of uplink channel characteristic identifiers for future selection, or both, based on the group of measurement results.
[0269] In some examples, the AI / ML training component 1535 is capable of, configured to, or operable to support a means for receiving one or more CSI reference signal identifiers, one or more SSB reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models. In some examples, the measurement evaluation component 1530 is capable of, configured to, or operable to support a means for obtaining a set of reference signal measurements corresponding to the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof. In some examples, to train the one or more machine learning models, the AI / ML training component 1535 is capable of, configured to, or operable to support a means for training the one or more machine learning models based on the set of reference signal measurements.
[0270] In some examples, to support receiving the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof, the measurement evaluation component 1530 is capable of, configured to, or operable to support a means for receiving an indication of a CSI reference signal resource, an indication of an SSB reference signal resource, or any combination thereof, where the CSI reference signal resource, the SSB reference signal resource, or any combination thereof are associated with corresponding downlink channel characteristics.
[0271] In some examples, the measurement evaluation component 1530 is capable of, configured to, or operable to support a means for receiving assistance information signaling including at least one of TRP location information including a location of a downlink TRP, a location of an uplink TRP, a location of the downlink TRP relative to the location of the uplink TRP, or spatial direction information for a quantity of downlink reference signals received by the UE.
[0272] In some examples, the measurement evaluation component 1530 is capable of, configured to, or operable to support a means for transmitting one or more UE capability reports indicating content requested to be included in the set of measurement results. In some examples, the AI / ML training component 1535 is capable of, configured to, or operable to support a means for receiving, for the training of the one or more machine learning models indicated by an associated identifier (ID) , signaling indicative of: an indication of a quantity of candidate SRS resource sets of the one or more SRS resource sets, an indication of a quantity of candidate SRS resources corresponding to the quantity of candidate SRS resource sets, an indication of one or more types of measurements included in the set of measurement results, or any combination thereof. In some examples, the measurement evaluation component 1530 is capable of, configured to, or operable to support a means for transmitting one or more messages indicative of one or more preferences selected from the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof.
[0273] In some examples, the AI / ML training component 1535 is capable of, configured to, or operable to support a means for transmitting one or more UE capability reports indicating support for training and inference of the one or more machine learning models using the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, where the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, are each indicated using the associated ID.
[0274] In some examples, the SRS scheduling component 1525 is capable of, configured to, or operable to support a means for transmitting a request for a different quantity of candidate SRS resource sets, a different quantity of candidate SRS resources, or a different set of measurements included in the set of measurement results, or any combination thereof. In some examples, the SRS scheduling component 1525 is capable of, configured to, or operable to support a means for receiving, via RRC reconfiguration signaling, an indication of the different quantity of candidate SRS resource sets, the different quantity of candidate SRS resources, or the different set of measurements included in the set of measurement results, or any combination thereof, based on the request. In some examples, the one or more parameters associated with the respective downlink reference signals include TCI state information, spatial relation information, or both.
[0275] FIG. 16 shows a diagram of a system 1600 including a device 1605 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The device 1605 may be an example of or include components of a device 1305, a device 1405, or a UE 115 as described herein. The device 1605 may communicate (e.g., wirelessly) with one or more other devices (e.g., network entities 105, UEs 115, or a combination thereof) . The device 1605 may include components for bi-directional voice and data communications including components for transmitting and receiving communications, such as a communications manager 1620, an input / output (I / O) controller, such as an I / O controller 1610, a transceiver 1615, one or more antennas 1625, at least one memory 1630, code 1635, and at least one processor 1640. These components may be in electronic communication or otherwise coupled (e.g., operatively, communicatively, functionally, electronically, electrically) via one or more buses (e.g., a bus 1645) .
[0276] The I / O controller 1610 may manage input and output signals for the device 1605. The I / O controller 1610 may also manage peripherals not integrated into the device 1605. In some cases, the I / O controller 1610 may represent a physical connection or port to an external peripheral. In some cases, the I / O controller 1610 may utilize an operating system such as or another known operating system. Additionally, or alternatively, the I / O controller 1610 may represent or interact with a modem, a keyboard, a mouse, a touchscreen, or a similar device. In some cases, the I / O controller 1610 may be implemented as part of one or more processors, such as the at least one processor 1640. In some cases, a user may interact with the device 1605 via the I / O controller 1610 or via hardware components controlled by the I / O controller 1610.
[0277] In some cases, the device 1605 may include a single antenna. However, in some other cases, the device 1605 may have more than one antenna, which may be capable of concurrently transmitting or receiving multiple wireless transmissions. The transceiver 1615 may communicate bi-directionally via the one or more antennas 1625 using wired or wireless links as described herein. For example, the transceiver 1615 may represent a wireless transceiver and may communicate bi-directionally with another wireless transceiver. The transceiver 1615 may also include a modem to modulate the packets, to provide the modulated packets to one or more antennas 1625 for transmission, and to demodulate packets received from the one or more antennas 1625. The transceiver 1615, or the transceiver 1615 and one or more antennas 1625, may be an example of a transmitter 1315, a transmitter 1415, a receiver 1310, a receiver 1410, or any combination thereof or component thereof, as described herein.
[0278] The at least one memory 1630 may include random access memory (RAM) and read-only memory (ROM) . The at least one memory 1630 may store computer-readable, computer-executable, or processor-executable code, such as the code 1635. The code 1635 may include instructions that, when executed by the at least one processor 1640, cause the device 1605 to perform various functions described herein. The code 1635 may be stored in a non-transitory computer-readable medium such as system memory or another type of memory. In some cases, the code 1635 may not be directly executable by the at least one processor 1640 but may cause a computer (e.g., when compiled and executed) to perform functions described herein. In some cases, the at least one memory 1630 may include, among other things, a basic I / O system (BIOS) which may control basic hardware or software operation such as the interaction with peripheral components or devices.
[0279] The at least one processor 1640 may include one or more intelligent hardware devices (e.g., one or more general-purpose processors, one or more DSPs, one or more CPUs, one or more graphics processing units (GPUs) , one or more neural processing units (NPUs) (also referred to as neural network processors or deep learning processors (DLPs) ) , one or more microcontrollers, one or more ASICs, one or more FPGAs, one or more programmable logic devices, discrete gate or transistor logic, one or more discrete hardware components, or any combination thereof) . In some cases, the at least one processor 1640 may be configured to operate a memory array using a memory controller. In some other cases, a memory controller may be integrated into the at least one processor 1640. The at least one processor 1640 may be configured to execute computer-readable instructions stored in a memory (e.g., the at least one memory 1630) to cause the device 1605 to perform various functions (e.g., functions or tasks supporting model training techniques for channel characteristic prediction procedures) . For example, the device 1605 or a component of the device 1605 may include at least one processor 1640 and at least one memory 1630 coupled with or to the at least one processor 1640, the at least one processor 1640 and the at least one memory 1630 configured to perform various functions described herein.
[0280] In some examples, the at least one processor 1640 may include multiple processors and the at least one memory 1630 may include multiple memories. One or more of the multiple processors may be coupled with one or more of the multiple memories, which may, individually or collectively, be configured to perform various functions described herein. In some examples, the at least one processor 1640 may be a component of a processing system, which may refer to a system (such as a series) of machines, circuitry (including, for example, one or both of processor circuitry (which may include the at least one processor 1640) and memory circuitry (which may include the at least one memory 1630) ) , or components, that receives or obtains inputs and processes the inputs to produce, generate, or obtain a set of outputs. The processing system may be configured to perform one or more of the functions described herein. For example, the at least one processor 1640 or a processing system including the at least one processor 1640 may be configured to, configurable to, or operable to cause the device 1605 to perform one or more of the functions described herein. Further, as described herein, being “configured to, ” being “configurable to, ” and being “operable to” may be used interchangeably and may be associated with a capability, when executing code 1635 (e.g., processor-executable code) stored in the at least one memory 1630 or otherwise, to perform one or more of the functions described herein.
[0281] The communications manager 1620 may support wireless communications in accordance with examples as disclosed herein. For example, the communications manager 1620 is capable of, configured to, or operable to support a means for receiving one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The communications manager 1620 is capable of, configured to, or operable to support a means for receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The communications manager 1620 is capable of, configured to, or operable to support a means for training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results. The communications manager 1620 is capable of, configured to, or operable to support a means for obtaining one or more predicted uplink channel characteristics associated with uplink transmission based on training the one or more machine learning models.
[0282] By including or configuring the communications manager 1620 in accordance with examples as described herein, the device 1605 may support techniques for reliable and efficient use of AI / ML models to predict channel characteristics, which in turn may provide for enhanced throughput (e.g., based on improved beam prediction accuracy) , reduced latency (e.g., based on improved beam prediction accuracy) , enhanced communications reliability, reduced power consumption, and efficient use of communications resources, including SRS resources and associated signaling.
[0283] In some examples, the communications manager 1620 may be configured to perform various operations (e.g., receiving, monitoring, transmitting) using or otherwise in cooperation with the transceiver 1615, the one or more antennas 1625, or any combination thereof. Although the communications manager 1620 is illustrated as a separate component, in some examples, one or more functions described with reference to the communications manager 1620 may be supported by or performed by the at least one processor 1640, the at least one memory 1630, the code 1635, or any combination thereof. For example, the code 1635 may include instructions executable by the at least one processor 1640 to cause the device 1605 to perform various aspects of model training techniques for channel characteristic prediction procedures as described herein, or the at least one processor 1640 and the at least one memory 1630 may be otherwise configured to, individually or collectively, perform or support such operations.
[0284] FIG. 17 shows a block diagram 1700 of a device 1705 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The device 1705 may be an example of aspects of a network entity 105 as described herein. The device 1705 may include a receiver 1710, a transmitter 1715, and a communications manager 1720. The device 1705, or one or more components of the device 1705 (e.g., the receiver 1710, the transmitter 1715, the communications manager 1720) , may include at least one processor, which may be coupled with at least one memory, to, individually or collectively, support or enable the described techniques. Each of these components may be in communication with one another (e.g., via one or more buses) .
[0285] The receiver 1710 may provide a means for obtaining (e.g., receiving, determining, identifying) information such as user data, control information, or any combination thereof (e.g., I / Q samples, symbols, packets, protocol data units, service data units) associated with various channels (e.g., control channels, data channels, information channels, channels associated with a protocol stack) . Information may be passed on to other components of the device 1705. In some examples, the receiver 1710 may support obtaining information by receiving signals via one or more antennas. Additionally, or alternatively, the receiver 1710 may support obtaining information by receiving signals via one or more wired (e.g., electrical, fiber optic) interfaces, wireless interfaces, or any combination thereof.
[0286] The transmitter 1715 may provide a means for outputting (e.g., transmitting, providing, conveying, sending) information generated by other components of the device 1705. For example, the transmitter 1715 may output information such as user data, control information, or any combination thereof (e.g., I / Q samples, symbols, packets, protocol data units, service data units) associated with various channels (e.g., control channels, data channels, information channels, channels associated with a protocol stack) . In some examples, the transmitter 1715 may support outputting information by transmitting signals via one or more antennas. Additionally, or alternatively, the transmitter 1715 may support outputting information by transmitting signals via one or more wired (e.g., electrical, fiber optic) interfaces, wireless interfaces, or any combination thereof. In some examples, the transmitter 1715 and the receiver 1710 may be co-located in a transceiver, which may include or be coupled with a modem.
[0287] The communications manager 1720, the receiver 1710, the transmitter 1715, or various combinations or components thereof may be examples of means for performing various aspects of model training techniques for channel characteristic prediction procedures as described herein. For example, the communications manager 1720, the receiver 1710, the transmitter 1715, or various combinations or components thereof may be capable of performing one or more of the functions described herein.
[0288] In some examples, the communications manager 1720, the receiver 1710, the transmitter 1715, or various combinations or components thereof may be implemented in hardware (e.g., in communications management circuitry) . The hardware may include at least one of a processor, a DSP, a CPU, an ASIC, an FPGA or other programmable logic device, a microcontroller, discrete gate or transistor logic, discrete hardware components, or any combination thereof configured as or otherwise supporting, individually or collectively, a means for performing the functions described in the present disclosure. In some examples, at least one processor and at least one memory coupled with the at least one processor may be configured to perform one or more of the functions described herein (e.g., by one or more processors, individually or collectively, executing instructions stored in the at least one memory) .
[0289] Additionally, or alternatively, the communications manager 1720, the receiver 1710, the transmitter 1715, or various combinations or components thereof may be implemented in code (e.g., as communications management software or firmware) executed by at least one processor (e.g., referred to as a processor-executable code) . If implemented in code executed by at least one processor, the functions of the communications manager 1720, the receiver 1710, the transmitter 1715, or various combinations or components thereof may be performed by a general-purpose processor, a DSP, a CPU, an ASIC, an FPGA, a microcontroller, or any combination of these or other programmable logic devices (e.g., configured as or otherwise supporting, individually or collectively, a means for performing the functions described in the present disclosure) .
[0290] In some examples, the communications manager 1720 may be configured to perform various operations (e.g., receiving, obtaining, monitoring, outputting, transmitting) using or otherwise in cooperation with the receiver 1710, the transmitter 1715, or both. For example, the communications manager 1720 may receive information from the receiver 1710, send information to the transmitter 1715, or be integrated in combination with the receiver 1710, the transmitter 1715, or both to obtain information, output information, or perform various other operations as described herein.
[0291] The communications manager 1720 may support wireless communications in accordance with examples as disclosed herein. For example, the communications manager 1720 is capable of, configured to, or operable to support a means for outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The communications manager 1720 is capable of, configured to, or operable to support a means for outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The communications manager 1720 is capable of, configured to, or operable to support a means for performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results. By including or configuring the communications manager 1720 in accordance with examples as described herein, the device 1705 (e.g., at least one processor controlling or otherwise coupled with the receiver 1710, the transmitter 1715, the communications manager 1720, or a combination thereof) may support techniques for reduced processing, reduced latency, reduced power consumption, more efficient utilization of communication resources, including for SRS and associated resources.
[0292] FIG. 18 shows a block diagram 1800 of a device 1805 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The device 1805 may be an example of aspects of a device 1705 or a network entity 105 as described herein. The device 1805 may include a receiver 1810, a transmitter 1815, and a communications manager 1820. The device 1805, or one or more components of the device 1805 (e.g., the receiver 1810, the transmitter 1815, the communications manager 1820) , may include at least one processor, which may be coupled with at least one memory, to support the described techniques. Each of these components may be in communication with one another (e.g., via one or more buses) .
[0293] The receiver 1810 may provide a means for obtaining (e.g., receiving, determining, identifying) information such as user data, control information, or any combination thereof (e.g., I / Q samples, symbols, packets, protocol data units, service data units) associated with various channels (e.g., control channels, data channels, information channels, channels associated with a protocol stack) . Information may be passed on to other components of the device 1805. In some examples, the receiver 1810 may support obtaining information by receiving signals via one or more antennas. Additionally, or alternatively, the receiver 1810 may support obtaining information by receiving signals via one or more wired (e.g., electrical, fiber optic) interfaces, wireless interfaces, or any combination thereof.
[0294] The transmitter 1815 may provide a means for outputting (e.g., transmitting, providing, conveying, sending) information generated by other components of the device 1805. For example, the transmitter 1815 may output information such as user data, control information, or any combination thereof (e.g., I / Q samples, symbols, packets, protocol data units, service data units) associated with various channels (e.g., control channels, data channels, information channels, channels associated with a protocol stack) . In some examples, the transmitter 1815 may support outputting information by transmitting signals via one or more antennas. Additionally, or alternatively, the transmitter 1815 may support outputting information by transmitting signals via one or more wired (e.g., electrical, fiber optic) interfaces, wireless interfaces, or any combination thereof. In some examples, the transmitter 1815 and the receiver 1810 may be co-located in a transceiver, which may include or be coupled with a modem.
[0295] The device 1805, or various components thereof, may be an example of means for performing various aspects of model training techniques for channel characteristic prediction procedures as described herein. For example, the communications manager 1820 may include an SRS scheduling component 1825, a measurement indication component 1830, an AI / ML management component 1835, or any combination thereof. The communications manager 1820 may be an example of aspects of a communications manager 1720 as described herein. In some examples, the communications manager 1820, or various components thereof, may be configured to perform various operations (e.g., receiving, obtaining, monitoring, outputting, transmitting) using or otherwise in cooperation with the receiver 1810, the transmitter 1815, or both. For example, the communications manager 1820 may receive information from the receiver 1810, send information to the transmitter 1815, or be integrated in combination with the receiver 1810, the transmitter 1815, or both to obtain information, output information, or perform various other operations as described herein.
[0296] The communications manager 1820 may support wireless communications in accordance with examples as disclosed herein. The SRS scheduling component 1825 is capable of, configured to, or operable to support a means for outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The measurement indication component 1830 is capable of, configured to, or operable to support a means for outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The AI / ML management component 1835 is capable of, configured to, or operable to support a means for performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0297] FIG. 19 shows a block diagram 1900 of a communications manager 1920 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The communications manager 1920 may be an example of aspects of a communications manager 1720, a communications manager 1820, or both, as described herein. The communications manager 1920, or various components thereof, may be an example of means for performing various aspects of model training techniques for channel characteristic prediction procedures as described herein. For example, the communications manager 1920 may include an SRS scheduling component 1925, a measurement indication component 1930, an AI / ML management component 1935, an associated ID application component 1940, or any combination thereof. Each of these components, or components or subcomponents thereof (e.g., one or more processors, one or more memories) , may communicate, directly or indirectly, with one another (e.g., via one or more buses) . The communications may include communications within a protocol layer of a protocol stack, communications associated with a logical channel of a protocol stack (e.g., between protocol layers of a protocol stack, within a device, component, or virtualized component associated with a network entity 105, between devices, components, or virtualized components associated with a network entity 105) , or any combination thereof.
[0298] The communications manager 1920 may support wireless communications in accordance with examples as disclosed herein. The SRS scheduling component 1925 is capable of, configured to, or operable to support a means for outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The AI / ML management component 1935 is capable of, configured to, or operable to support a means for performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0299] In some examples, one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest RSRP or a highest selection probability, one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set, or one or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set. In some examples, the one or more SRS resource sets include at least a first SRS resource set and a second SRS resource set. In some examples, the set of measurement results corresponding to the first SRS resource set includes a set of RSRP measurements for each SRS resource included in the first SRS resource set, and the set of measurement results corresponding to the second SRS resource set includes a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.
[0300] In some examples, the associated ID application component 1940 is capable of, configured to, or operable to support a means for outputting signaling including an associated identifier (ID) that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink TRP, where the set of measurement results are based on one or more SRS resource sets. In some examples, consistency of one or more network conditions between training of the one or more machine learning models and an inference phase of the one or more machine learning models is based on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.
[0301] In some examples, a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase, a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics, a quantity of SRS resources or SRS resource sets, a periodicity of the one or more SRS resource sets, or one or more types of measurements included in the set of measurement results.
[0302] In some examples, to support outputting the set of measurement results, the measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, where a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink TRP, to an uplink channel characteristic, or both.
[0303] In some examples, the measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting one or more CSI reference signal identifiers, one or more SSB reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models, where the one or more machine learning models are based on a set of reference signal measurements corresponding to the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof. In some examples, to support outputting the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof, the measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting an indication of a CSI reference signal resource, an indication of an SSB reference signal resource, or any combination thereof, where the CSI reference signal resource, the SSB reference signal resource, or any combination thereof are associated with corresponding downlink channel characteristic identifiers.
[0304] In some examples, the measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting assistance information signaling including at least one of TRP location information including a location of a downlink TRP, a location of an uplink TRP, a location of the downlink TRP relative to the location of the uplink TRP, or spatial direction information for a quantity of downlink reference signals received by the UE.
[0305] In some examples, the AI / ML management component 1935 is capable of, configured to, or operable to support a means for obtaining one or more UE capability reports indicating content requested to be included in the set of measurement results. In some examples, the measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting, for training of the one or more machine learning models indicated by an associated identifier (ID) , signaling indicative of an indication of a quantity of candidate SRS resource sets of the one or more SRS resource sets, an indication of a quantity of candidate SRS resources corresponding to the quantity of candidate SRS resource sets, an indication of one or more types of measurements included in the set of measurement results, or any combination thereof.
[0306] In some examples, the AI / ML management component 1935 is capable of, configured to, or operable to support a means for obtaining one or more messages indicative of one or more preferences selected from the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof. In some examples, the AI / ML management component 1935 is capable of, configured to, or operable to support a means for obtaining one or more UE capability reports indicating support for training and inference of the one or more machine learning models using the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, where the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, are each indicated using the associated ID.
[0307] In some examples, the AI / ML management component 1935 is capable of, configured to, or operable to support a means for obtaining a request for a different quantity of candidate SRS resource sets, a different quantity of candidate SRS resources, or a different set of measurements included in the set of measurement results, or any combination thereof. In some examples, the measurement indication component 1930 is capable of, configured to, or operable to support a means for outputting, via RRC reconfiguration signaling, an indication of the different quantity of candidate SRS resource sets, the different quantity of candidate SRS resources, or the different set of measurements included in the set of measurement results, or any combination thereof, based on the request. In some examples, the one or more parameters associated with the respective downlink reference signals include TCI state information, spatial relation information, or both.
[0308] FIG. 20 shows a diagram of a system 2000 including a device 2005 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The device 2005 may be an example of or include components of a device 1705, a device 1805, or a network entity 105 as described herein. The device 2005 may communicate with other network devices or network equipment such as one or more of the network entities 105, UEs 115, or any combination thereof. The communications may include communications over one or more wired interfaces, over one or more wireless interfaces, or any combination thereof. The device 2005 may include components that support outputting and obtaining communications, such as a communications manager 2020, a transceiver 2010, one or more antennas 2015, at least one memory 2025, code 2030, and at least one processor 2035. These components may be in electronic communication or otherwise coupled (e.g., operatively, communicatively, functionally, electronically, electrically) via one or more buses (e.g., a bus 2040) .
[0309] The transceiver 2010 may support bi-directional communications via wired links, wireless links, or both as described herein. In some examples, the transceiver 2010 may include a wired transceiver and may communicate bi-directionally with another wired transceiver. Additionally, or alternatively, in some examples, the transceiver 2010 may include a wireless transceiver and may communicate bi-directionally with another wireless transceiver. In some examples, the device 2005 may include one or more antennas 2015, which may be capable of transmitting or receiving wireless transmissions (e.g., concurrently) . The transceiver 2010 may also include a modem to modulate signals, to provide the modulated signals for transmission (e.g., by one or more antennas 2015, by a wired transmitter) , to receive modulated signals (e.g., from one or more antennas 2015, from a wired receiver) , and to demodulate signals. In some implementations, the transceiver 2010 may include one or more interfaces, such as one or more interfaces coupled with the one or more antennas 2015 that are configured to support various receiving or obtaining operations, or one or more interfaces coupled with the one or more antennas 2015 that are configured to support various transmitting or outputting operations, or a combination thereof. In some implementations, the transceiver 2010 may include or be configured for coupling with one or more processors or one or more memory components that are operable to perform or support operations based on received or obtained information or signals, or to generate information or other signals for transmission or other outputting, or any combination thereof. In some implementations, the transceiver 2010, or the transceiver 2010 and the one or more antennas 2015, or the transceiver 2010 and the one or more antennas 2015 and one or more processors or one or more memory components (e.g., the at least one processor 2035, the at least one memory 2025, or both) , may be included in a chip or chip assembly that is installed in the device 2005. In some examples, the transceiver 2010 may be operable to support communications via one or more communications links (e.g., communication link (s) 125, backhaul communication link (s) 120, a midhaul communication link 162, a fronthaul communication link 168) .
[0310] The at least one memory 2025 may include RAM, ROM, or any combination thereof. The at least one memory 2025 may store computer-readable, computer-executable, or processor-executable code, such as the code 2030. The code 2030 may include instructions that, when executed by one or more of the at least one processor 2035, cause the device 2005 to perform various functions described herein. The code 2030 may be stored in a non-transitory computer-readable medium such as system memory or another type of memory. In some cases, the code 2030 may not be directly executable by a processor of the at least one processor 2035 but may cause a computer (e.g., when compiled and executed) to perform functions described herein. In some cases, the at least one memory 2025 may include, among other things, a BIOS which may control basic hardware or software operation such as the interaction with peripheral components or devices. In some examples, the at least one processor 2035 may include multiple processors and the at least one memory 2025 may include multiple memories. One or more of the multiple processors may be coupled with one or more of the multiple memories which may, individually or collectively, be configured to perform various functions herein (for example, as part of a processing system) .
[0311] The at least one processor 2035 may include one or more intelligent hardware devices (e.g., one or more general-purpose processors, one or more DSPs, one or more CPUs, one or more graphics processing units (GPUs) , one or more neural processing units (NPUs) (also referred to as neural network processors or deep learning processors (DLPs) ) , one or more microcontrollers, one or more ASICs, one or more FPGAs, one or more programmable logic devices, discrete gate or transistor logic, one or more discrete hardware components, or any combination thereof) . In some cases, the at least one processor 2035 may be configured to operate a memory array using a memory controller. In some other cases, a memory controller may be integrated into one or more of the at least one processor 2035. The at least one processor 2035 may be configured to execute computer-readable instructions stored in a memory (e.g., one or more of the at least one memory 2025) to cause the device 2005 to perform various functions (e.g., functions or tasks supporting model training techniques for channel characteristic prediction procedures) . For example, the device 2005 or a component of the device 2005 may include at least one processor 2035 and at least one memory 2025 coupled with one or more of the at least one processor 2035, the at least one processor 2035 and the at least one memory 2025 configured to perform various functions described herein. The at least one processor 2035 may be an example of a cloud-computing platform (e.g., one or more physical nodes and supporting software such as operating systems, virtual machines, or container instances) that may host the functions (e.g., by executing code 2030) to perform the functions of the device 2005. The at least one processor 2035 may be any one or more suitable processors capable of executing scripts or instructions of one or more software programs stored in the device 2005 (such as within one or more of the at least one memory 2025) .
[0312] In some examples, the at least one processor 2035 may include multiple processors and the at least one memory 2025 may include multiple memories. One or more of the multiple processors may be coupled with one or more of the multiple memories, which may, individually or collectively, be configured to perform various functions herein. In some examples, the at least one processor 2035 may be a component of a processing system, which may refer to a system (such as a series) of machines, circuitry (including, for example, one or both of processor circuitry (which may include the at least one processor 2035) and memory circuitry (which may include the at least one memory 2025) ) , or components, that receives or obtains inputs and processes the inputs to produce, generate, or obtain a set of outputs. The processing system may be configured to perform one or more of the functions described herein. For example, the at least one processor 2035 or a processing system including the at least one processor 2035 may be configured to, configurable to, or operable to cause the device 2005 to perform one or more of the functions described herein. Further, as described herein, being “configured to, ” being “configurable to, ” and being “operable to” may be used interchangeably and may be associated with a capability, when executing code stored in the at least one memory 2025 or otherwise, to perform one or more of the functions described herein.
[0313] In some examples, a bus 2040 may support communications of (e.g., within) a protocol layer of a protocol stack. In some examples, a bus 2040 may support communications associated with a logical channel of a protocol stack (e.g., between protocol layers of a protocol stack) , which may include communications performed within a component of the device 2005, or between different components of the device 2005 that may be co-located or located in different locations (e.g., where the device 2005 may refer to a system in which one or more of the communications manager 2020, the transceiver 2010, the at least one memory 2025, the code 2030, and the at least one processor 2035 may be located in one of the different components or divided between different components) .
[0314] In some examples, the communications manager 2020 may manage aspects of communications with a core network 130 (e.g., via one or more wired or wireless backhaul links) . For example, the communications manager 2020 may manage the transfer of data communications for client devices, such as one or more UEs 115. In some examples, the communications manager 2020 may manage communications with one or more other network entities 105, and may include a controller or scheduler for controlling communications with UEs 115 (e.g., in cooperation with the one or more other network devices) . In some examples, the communications manager 2020 may support an X2 interface within an LTE / LTE-A wireless communications network technology to provide communication between network entities 105.
[0315] The communications manager 2020 may support wireless communications in accordance with examples as disclosed herein. For example, the communications manager 2020 is capable of, configured to, or operable to support a means for outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The communications manager 2020 is capable of, configured to, or operable to support a means for outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The communications manager 2020 is capable of, configured to, or operable to support a means for performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results.
[0316] By including or configuring the communications manager 2020 in accordance with examples as described herein, the device 2005 may support techniques for reliable and efficient use of AI / ML models to predict channel characteristics, which in turn may provide for enhanced throughput (e.g., based on improved beam prediction accuracy) , reduced latency (e.g., based on improved beam prediction accuracy) , enhanced communications reliability, reduced power consumption, and efficient use of communications resources, including SRS resources and associated signaling.
[0317] In some examples, the communications manager 2020 may be configured to perform various operations (e.g., receiving, obtaining, monitoring, outputting, transmitting) using or otherwise in cooperation with the transceiver 2010, the one or more antennas 2015 (e.g., where applicable) , or any combination thereof. Although the communications manager 2020 is illustrated as a separate component, in some examples, one or more functions described with reference to the communications manager 2020 may be supported by or performed by the transceiver 2010, one or more of the at least one processor 2035, one or more of the at least one memory 2025, the code 2030, or any combination thereof (for example, by a processing system including at least a portion of the at least one processor 2035, the at least one memory 2025, the code 2030, or any combination thereof) . For example, the code 2030 may include instructions executable by one or more of the at least one processor 2035 to cause the device 2005 to perform various aspects of model training techniques for channel characteristic prediction procedures as described herein, or the at least one processor 2035 and the at least one memory 2025 may be otherwise configured to, individually or collectively, perform or support such operations.
[0318] FIG. 21 shows a flowchart illustrating a method 2100 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The operations of the method 2100 may be implemented by a UE or its components as described herein. For example, the operations of the method 2100 may be performed by a UE 115 as described with reference to FIGs. 1 through 16. In some examples, a UE may execute a set of instructions to control the functional elements of the UE to perform the described functions. Additionally, or alternatively, the UE may perform aspects of the described functions using special-purpose hardware.
[0319] At 2105, the method may include receiving one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The operations of 2105 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2105 may be performed by an SRS scheduling component 1525 as described with reference to FIG. 15.
[0320] At 2110, the method may include receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The operations of 2110 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2110 may be performed by a measurement evaluation component 1530 as described with reference to FIG. 15.
[0321] At 2115, the method may include training one or more machine learning models based on the one or more SRS resource sets and the set of measurement results. The operations of 2115 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2115 may be performed by an AI / ML training component 1535 as described with reference to FIG. 15.
[0322] At 2120, the method may include obtaining one or more predicted uplink channel characteristics associated with uplink transmission based on training the one or more machine learning models. The operations of 2120 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2120 may be performed by an AI / ML inferencing component 1540 as described with reference to FIG. 15.
[0323] FIG. 22 shows a flowchart illustrating a method 2200 that supports model training techniques for channel characteristic prediction procedures in accordance with one or more aspects of the present disclosure. The operations of the method 2200 may be implemented by a network entity or its components as described herein. For example, the operations of the method 2200 may be performed by a network entity as described with reference to FIGs. 1 through 12 and 17 through 20. In some examples, a network entity may execute a set of instructions to control the functional elements of the network entity to perform the described functions. Additionally, or alternatively, the network entity may perform aspects of the described functions using special-purpose hardware.
[0324] At 2205, the method may include outputting one or more messages including scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals. The operations of 2205 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2205 may be performed by an SRS scheduling component 1925 as described with reference to FIG. 19.
[0325] At 2210, the method may include outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets. The operations of 2210 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2210 may be performed by a measurement indication component 1930 as described with reference to FIG. 19.
[0326] At 2215, the method may include performing one or more prediction management procedures with a UE based on one or more machine learning models that are based on the one or more SRS resource sets and the set of measurement results. The operations of 2215 may be performed in accordance with examples as disclosed herein. In some examples, aspects of the operations of 2215 may be performed by an AI / ML management component 1935 as described with reference to FIG. 19.
[0327] The following provides an overview of aspects of the present disclosure:
[0328] Aspect 1: A method for wireless communications at a UE, comprising: receiving one or more messages comprising scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals; receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets; training one or more machine learning models based at least in part on the one or more SRS resource sets and the set of measurement results; and obtaining one or more predicted uplink channel characteristics associated uplink transmission based at least in part on training the one or more machine learning models.
[0329] Aspect 2: The method of aspect 1, wherein the set of measurement results correspond to an SRS resource set of the one or more SRS resource sets, the set of measurement results comprising at least one of: one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest RSRP or a highest selection probability, one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set. or one or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.
[0330] Aspect 3: The method of any of aspects 1 through 2, wherein the one or more predicted uplink channel characteristics comprise one or more temporal predictions, and wherein obtaining the one or more predicted uplink channel characteristics comprises: obtaining the one or more temporal predictions based at least in part on the training of the one or more machine learning models using one or more previously received SRS resource sets and corresponding sets of previously received measurement results, wherein the corresponding sets of previously received measurement results comprise one or more RSRP measurements associated with each SRS resource of the one or more previously received SRS resource sets.
[0331] Aspect 4: The method of any of aspects 1 through 3, wherein obtaining the one or more predicted uplink channel characteristics comprise one or more spatial predictions, and wherein obtaining the one or more predicted uplink channel characteristics comprises: obtaining the one or more spatial predictions using a set of candidate channel characteristics obtained during the training of the one or more machine learning models and corresponding one or more measurement results, wherein corresponding sets of previously received measurement results comprise one or more RSRP measurements associated with each SRS resource of one or more previously received SRS resource sets, and wherein the one or more spatial predictions indicate a quantity of channel characteristics that is less than a quantity of the set of candidate channel characteristics.
[0332] Aspect 5: The method of any of aspects 1 through 4, wherein the one or more SRS resource sets comprise at least a first SRS resource set and a second SRS resource set, and the set of measurement results corresponding to the first SRS resource set comprises a set of RSRP measurements for each SRS resource included in the first SRS resource set, and the set of measurement results corresponding to the second SRS resource set comprises a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.
[0333] Aspect 6: The method of any of aspects 1 through 5, further comprising: receiving signaling comprising an associated ID that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink TRP, wherein the set of measurement results are based at least in part on one or more SRS resource sets.
[0334] Aspect 7: The method of aspect 6, wherein consistency of one or more network conditions between the training of the one or more machine learning models and an inference phase of the one or more machine learning models is based at least in part on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.
[0335] Aspect 8: The method of aspect 7, wherein the one or more network conditions comprise at least one of a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase; a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics; a quantity of SRS resources or SRS resource sets; a periodicity of the one or more SRS resource sets; or one or more types of measurements included in the set of measurement results.
[0336] Aspect 9: The method of any of aspects 1 through 8, wherein receiving the set of measurement results comprises: receiving an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, wherein a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink TRP, to an uplink channel characteristic, or both.
[0337] Aspect 10: The method of aspect 9, further comprising: predicting, using the one or more machine learning models, a set of uplink TRP identifiers for future selection, a set of uplink channel characteristic identifiers for future selection, or both, based at least in part on the group of measurement results.
[0338] Aspect 11: The method of any of aspects 1 through 10, further comprising: receiving one or more CSI-RS identifiers, one or more SSB reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models; obtaining a set of reference signal measurements corresponding to the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof, wherein training the one or more machine learning models comprises: training the one or more machine learning models based at least in part on the set of reference signal measurements.
[0339] Aspect 12: The method of aspect 11, wherein receiving the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof, comprises: receiving an indication of a CSI-RS resource, an indication of an SSB reference signal resource, or any combination thereof, wherein the CSI-RS resource, the SSB reference signal resource, or any combination thereof are associated with corresponding downlink channel characteristics.
[0340] Aspect 13: The method of any of aspects 1 through 12, further comprising: receiving assistance information signaling comprising at least one of TRP location information comprising a location of a downlink TRP, a location of an uplink TRP, a location of the downlink TRP relative to the location of the uplink TRP, or spatial direction information for a quantity of downlink reference signals received by the UE.
[0341] Aspect 14: The method of any of aspects 1 through 13, further comprising: transmitting one or more UE capability reports indicating content requested to be included in the set of measurement results.
[0342] Aspect 15: The method of any of aspects 1 through 14, further comprising: receiving, for the training of the one or more machine learning models indicated by an associated ID, signaling indicative of: an indication of a quantity of candidate SRS resource sets of the one or more SRS resource sets, an indication of a quantity of candidate SRS resources corresponding to the quantity of candidate SRS resource sets, an indication of one or more types of measurements included in the set of measurement results, or any combination thereof.
[0343] Aspect 16: The method of aspect 15, further comprising: transmitting one or more messages indicative of one or more preferences selected from the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof.
[0344] Aspect 17: The method of any of aspects 15 through 16, further comprising: transmitting one or more UE capability reports indicating support for training and inference of the one or more machine learning models using the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, wherein the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, are each indicated using the associated ID.
[0345] Aspect 18: The method of any of aspects 15 through 17, further comprising: transmitting a request for a different quantity of candidate SRS resource sets, a different quantity of candidate SRS resources, or a different set of measurements included in the set of measurement results, or any combination thereof; and receiving, via RRC reconfiguration signaling, an indication of the different quantity of candidate SRS resource sets, the different quantity of candidate SRS resources, or the different set of measurements included in the set of measurement results, or any combination thereof, based at least in part on the request.
[0346] Aspect 19: The method of any of aspects 1 through 18, wherein the one or more parameters associated with the respective downlink reference signals comprise transmission configuration indicator (TCI) state information, spatial relation information, or both.
[0347] Aspect 20: A method for wireless communications at a network entity, comprising: outputting one or more messages comprising scheduling information for one or more SRS resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals; outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets; and performing one or more prediction management procedures with a UE based at least in part on one or more machine learning models that are based at least in part on the one or more SRS resource sets and the set of measurement results.
[0348] Aspect 21: The method of aspect 20, wherein the set of measurement results correspond to an SRS resource set of the one or more SRS resource sets, the set of measurement results comprising at least one of one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest RSRP or a highest selection probability, one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set, or one or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.
[0349] Aspect 22: The method of any of aspects 20 through 21, wherein the one or more SRS resource sets comprise at least a first SRS resource set and a second SRS resource set, and the set of measurement results corresponding to the first SRS resource set comprises a set of RSRP measurements for each SRS resource included in the first SRS resource set, and the set of measurement results corresponding to the second SRS resource set comprises a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.
[0350] Aspect 23: The method of any of aspects 20 through 22, further comprising: outputting signaling comprising an associated ID that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink TRP, wherein the set of measurement results are based at least in part on one or more SRS resource sets.
[0351] Aspect 24: The method of aspect 23, wherein consistency of one or more network conditions between training of the one or more machine learning models and an inference phase of the one or more machine learning models is based at least in part on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.
[0352] Aspect 25: The method of aspect 24, wherein the one or more network conditions comprise at least one of a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase, a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics, a quantity of SRS resources or SRS resource sets, a periodicity of the one or more SRS resource sets, or one or more types of measurements included in the set of measurement results.
[0353] Aspect 26: The method of any of aspects 20 through 25, wherein outputting the set of measurement results comprises: outputting an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, wherein a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink TRP, to an uplink channel characteristic, or both.
[0354] Aspect 27: The method of any of aspects 20 through 26, further comprising: outputting one or more CSI-RS identifiers, one or more SSB reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models, wherein the one or more machine learning models are based at least in part on a set of reference signal measurements corresponding to the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof.
[0355] Aspect 28: The method of aspect 27, wherein outputting the one or more CSI-RS identifiers, the one or more SSB reference signal identifiers, or any combination thereof, comprises: outputting an indication of a CSI-RS resource, an indication of an SSB reference signal resource, or any combination thereof, wherein the CSI-RS resource, the SSB reference signal resource, or any combination thereof are associated with corresponding downlink channel characteristic identifiers.
[0356] Aspect 29: The method of any of aspects 20 through 28, further comprising: outputting assistance information signaling comprising at least one of TRP location information comprising a location of a downlink TRP, a location of an uplink TRP, a location of the downlink ...
Claims
1.A user equipment (UE) , comprising:one or more memories storing processor-executable code; andone or more processors coupled with the one or more memories and individually or collectively operable to execute the code to cause the UE to:receive one or more messages comprising scheduling information for one or more sounding reference signal (SRS) resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals;receive, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets;train one or more machine learning models based at least in part on the one or more SRS resource sets and the set of measurement results; andobtain one or more predicted uplink channel characteristics associated with uplink transmission based at least in part on training the one or more machine learning models.2.The UE of claim 1, the set of measurement results correspond to an SRS resource set of the one or more SRS resource sets, the set of measurement results comprising at least one of:one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest reference signal receive power (RSRP) or a highest selection probability;one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set; orone or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.3.The UE of claim 1, wherein the one or more predicted uplink channel characteristics comprise one or more temporal predictions, and wherein, to obtain the one or more predicted uplink channel characteristics, the one or more processors are individually or collectively operable to execute the code to cause the UE to:obtain the one or more temporal predictions based at least in part on the training of the one or more machine learning models using one or more previously received SRS resource sets and corresponding sets of previously received measurement results, wherein the corresponding sets of previously received measurement results comprise one or more reference signal receive power (RSRP) measurements associated with each SRS resource of the one or more previously received SRS resource sets.4.The UE of claim 1, wherein the one or more predicted uplink channel characteristics comprise one or more spatial predictions, and wherein, to obtain the one or more predicted uplink channel characteristics, the one or more processors are individually or collectively operable to execute the code to cause the UE to:obtain the one or more spatial predictions using a set of candidate channel characteristics obtained during the training of the one or more machine learning models and corresponding one or more measurement results, wherein corresponding sets of previously received measurement results comprise one or more reference signal receive power (RSRP) measurements associated with each SRS resource of one or more previously received SRS resource sets, and wherein the one or more spatial predictions indicate a quantity of channel characteristics that is less than a quantity of the set of candidate channel characteristics.5.The UE of claim 1, wherein the one or more SRS resource sets comprise at least a first SRS resource set and a second SRS resource set, and wherein:the set of measurement results corresponding to the first SRS resource set comprises a set of reference signal receive power (RSRP) measurements for each SRS resource included in the first SRS resource set; andthe set of measurement results corresponding to the second SRS resource set comprises a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.6.The UE of claim 1, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:receive signaling comprising an associated identifier (ID) that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink transmission reception point (TRP) , wherein the set of measurement results are based at least in part on one or more SRS resource sets.7.The UE of claim 6, wherein consistency of one or more network conditions between the training of the one or more machine learning models and an inference phase of the one or more machine learning models is based at least in part on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.8.The UE of claim 7, wherein the one or more network conditions comprise at least one of:a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase;a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics;a quantity of SRS resources or SRS resource sets;a periodicity of the one or more SRS resource sets; orone or more types of measurements included in the set of measurement results.9.The UE of claim 1, wherein, to receive the set of measurement results, the one or more processors are individually or collectively operable to execute the code to cause the UE to:receive an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, wherein a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink transmission reception point (TRP) , to an uplink channel characteristic, or both.10.The UE of claim 9, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:predict, using the one or more machine learning models, a set of uplink TRP identifiers for future selection, a set of uplink channel characteristic identifiers for future selection, or both, based at least in part on the group of measurement results.11.The UE of claim 1, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:receive one or more channel state information (CSI) reference signal identifiers, one or more synchronization signal block (SSB) reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models;obtain a set of reference signal measurements corresponding to the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof, wherein training the one or more machine learning models comprises:train the one or more machine learning models based at least in part on the set of reference signal measurements.12.The UE of claim 11, wherein, to receive the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof, the one or more processors are individually or collectively operable to execute the code to cause the UE to:receive an indication of a CSI reference signal resource, an indication of an SSB reference signal resource, or any combination thereof, wherein the CSI reference signal resource, the SSB reference signal resource, or any combination thereof are associated with corresponding downlink channel characteristics.13.The UE of claim 1, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:receive assistance information signaling comprising at least one of transmission reception point (TRP) location information comprising a location of a downlink TRP, a location of an uplink TRP, a location of the downlink TRP relative to the location of the uplink TRP, or spatial direction information for a quantity of downlink reference signals received by the UE.14.The UE of claim 1, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:transmit one or more UE capability reports indicating content requested to be included in the set of measurement results.15.The UE of claim 1, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:receive, for the training of the one or more machine learning models indicated by an associated identifier (ID) , signaling indicative of: an indication of a quantity of candidate SRS resource sets of the one or more SRS resource sets, an indication of a quantity of candidate SRS resources corresponding to the quantity of candidate SRS resource sets, an indication of one or more types of measurements included in the set of measurement results, or any combination thereof.16.The UE of claim 15, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:transmit one or more messages indicative of one or more preferences selected from the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof.17.The UE of claim 15, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:transmit one or more UE capability reports indicating support for training and inference of the one or more machine learning models using the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, wherein the candidate SRS resource sets, the candidate SRS resources, the one or more types of measurements, or any combination thereof, are each indicated using the associated ID.18.The UE of claim 15, wherein the one or more processors are individually or collectively further operable to execute the code to cause the UE to:transmit a request for a different quantity of candidate SRS resource sets, a different quantity of candidate SRS resources, or a different set of measurements included in the set of measurement results, or any combination thereof; andreceive, via radio resource control (RRC) reconfiguration signaling, an indication of the different quantity of candidate SRS resource sets, the different quantity of candidate SRS resources, or the different set of measurements included in the set of measurement results, or any combination thereof, based at least in part on the request.19.The UE of claim 1, wherein the one or more parameters associated with the respective downlink reference signals comprise transmission configuration indicator (TCI) state information, spatial relation information, or both.20.A network entity, comprising:one or more memories storing processor-executable code; andone or more processors coupled with the one or more memories and individually or collectively operable to execute the code to cause the network entity to:output one or more messages comprising scheduling information for one or more sounding reference signal (SRS) resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals;output a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets; andperform one or more prediction management procedures with a user equipment (UE) based at least in part on one or more machine learning models that are based at least in part on the one or more SRS resource sets and the set of measurement results.21.The network entity of claim 20, wherein the set of measurement results correspond to an SRS resource set of the one or more SRS resource sets, the set of measurement results comprising at least one of:one or more SRS resource identifiers corresponding to a set of SRS resources of the SRS resource set, the set of SRS resources having a strongest reference signal receive power (RSRP) or a highest selection probability;one or more RSRP measurements corresponding to the set of SRS resources of the SRS resource set; orone or more selection probability measurements corresponding to the set of SRS resources of the SRS resource set.22.The network entity of claim 20, wherein the one or more SRS resource sets comprise at least a first SRS resource set and a second SRS resource set, and wherein:the set of measurement results corresponding to the first SRS resource set comprises a set of reference signal receive power (RSRP) measurements for each SRS resource included in the first SRS resource set; andthe set of measurement results corresponding to the second SRS resource set comprises a single SRS resource identifier corresponding to a threshold RSRP measurement of a set of RSRP measurements for SRS resources included in the second SRS resource set.23.The network entity of claim 20, wherein the one or more processors are individually or collectively further operable to execute the code to cause the network entity to:output signaling comprising an associated identifier (ID) that indicates the set of measurement results associated with a set of uplink channel characteristics of an uplink transmission reception point (TRP) , wherein the set of measurement results are based at least in part on one or more SRS resource sets.24.The network entity of claim 23, wherein consistency of one or more network conditions between training of the one or more machine learning models and an inference phase of the one or more machine learning models is based at least in part on the associated ID received during the training being the same as a corresponding associated ID received during the inference phase.25.The network entity of claim 24, wherein the one or more network conditions comprise at least one of:a set of spatial filters associated with the set of uplink channel characteristics applied during the training and the inference phase;a quantity of uplink channel characteristics or an ordering of uplink channel characteristic IDs included in the set of uplink channel characteristics;a quantity of SRS resources or SRS resource sets;a periodicity of the one or more SRS resource sets; orone or more types of measurements included in the set of measurement results.26.The network entity of claim 20, wherein, to output the set of measurement results, the one or more processors are individually or collectively operable to execute the code to cause the network entity to:output an indication of one or more groups of measurement results corresponding to an SRS resource set of the one or more SRS resource sets, wherein a group of measurement results of the one or more groups of measurement results includes an identifier that associates the group of measurement results to an uplink transmission reception point (TRP) , to an uplink channel characteristic, or both.27.The network entity of claim 20, wherein the one or more processors are individually or collectively further operable to execute the code to cause the network entity to:output one or more channel state information (CSI) reference signal identifiers, one or more synchronization signal block (SSB) reference signal identifiers, or any combination thereof, indicated for training of the one or more machine learning models, wherein the one or more machine learning models are based at least in part on a set of reference signal measurements corresponding to the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof.28.The network entity of claim 27, wherein, to output the one or more CSI reference signal identifiers, the one or more SSB reference signal identifiers, or any combination thereof, the one or more processors are individually or collectively operable to execute the code to cause the network entity to:output an indication of a CSI reference signal resource, an indication of an SSB reference signal resource, or any combination thereof, wherein the CSI reference signal resource, the SSB reference signal resource, or any combination thereof are associated with corresponding downlink channel characteristic identifiers.29.A method for wireless communications at a user equipment (UE) , comprising:receiving one or more messages comprising scheduling information for one or more sounding reference signal (SRS) resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals;receiving, from a network entity, a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets;training one or more machine learning models based at least in part on the one or more SRS resource sets and the set of measurement results; andobtaining one or more predicted uplink channel characteristics associated with uplink transmission based at least in part on training the one or more machine learning models.30.A method for wireless communications at a network entity, comprising:outputting one or more messages comprising scheduling information for one or more sounding reference signal (SRS) resource sets, the one or more SRS resource sets excluding one or more parameters associated with respective downlink reference signals;outputting a set of measurement results corresponding to respective SRS resource sets of the one or more SRS resource sets; andperforming one or more prediction management procedures with a user equipment (UE) based at least in part on one or more machine learning models that are based at least in part on the one or more SRS resource sets and the set of measurement results.