A method for life-cycle management of artificial intelligence and machine learning models in wireless networks
AI and ML models in wireless networks enhance adaptive configuration by reducing resource intensity and computation needs, improving beam management and CSI feedback efficiency.
Patent Information
- Application Number
- PCT/CN2024/077076
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- Filing Date
- 2024-02-08
- Publication Date
- 2025-08-14
AI Technical Summary
Existing wireless communication systems face inefficiencies in determining adaptive network configurations due to lengthy and resource-intensive measurement/reporting processes and significant computation power requirements, particularly in beam management, channel state information feedback, and wireless terminal positioning.
Implementing AI and ML functionalities and models on terminal devices and network nodes to facilitate collaborative procedures for efficient configuration and provisioning, including testing, activating, or deactivating these models through a sequence of triggered messages.
Reduces the need for extensive measurements and computation by leveraging learned correlations, enabling more agile and efficient network configurations, such as improved beam management, CSI feedback, and UE positioning.
Smart Images

Figure CN2024077076_14082025_PF_FP_ABST
Abstract
Description
A METHOD FOR LIFE-CYCLE MANAGEMENT OF ARTIFICIAL INTELLIGENCE AND MACHINE LEARNING MODELS IN WIRELESS NETWORKSTECHNICAL FIELD
[0001] This disclosure is directed generally to wireless communication networks and particularly to configuration and provisioning of Artificial Intelligence (AI) and / or Machine Learning (ML) functionalities and models in both terminal devices and network nodes in wireless communication networks.BACKGROUND
[0002] In a wireless communication system, determination of adaptive network configuration of communication resources particularly within an over-the-air communication interface may require lengthy and resource-intensive measurement / reporting processes and / or significant amounts of computation power. Such types of configurations of communication resources may include but are not limited to beam management, channel state information (CSI) feedback compression and decompression, wireless terminal positioning / orientation, and the like. Correlation between various network conditions and these adaptive resource configurations may be learned via artificial intelligence (AI) techniques and models and utilized to assist in provisioning of the wireless communication resources. It may thus be desirable to provide an efficient mechanism for configuration, selection, and provisioning the various AI models, deployed either on terminal devices or on wireless network nodes, based on capabilities of the terminal devices and the radio network conditions between these terminal devices and the wireless network.SUMMARY
[0003] This disclosure is directed generally to wireless communication networks and particularly to configuration and provisioning of Artificial Intelligence (AI) and / or Machine Learning (ML) functionalities and models in both terminal devices and network nodes in wireless communication networks. For example, such AI / ML functionalities and models may reside in either the wireless terminal side or in the wireless network side. The wireless terminal and the network may perform a collaborative procedure to determine, configure, activate, or deactivate a set of AI / ML functionalities and models for adaptive prediction and inference of network provisioning via a sequence of triggered messages.
[0004] In one example implementation, a method performed by a user equipment (UE) in communication with a network (NW) in a wireless communication network is disclosed. The method may include performing a testing procedure for Artificial Intelligence or Machine Learning (AIML) capabilities of the UE in collaboration with the NW; receiving from the NW an AIML configuration for one or more AIML features, functionalities, or models selected by the NW from AIML capabilities of the UE based at least partially on the testing procedure; and utilizing the one or more AIML features, functionalities, or models for inference according to the AIML configuration.
[0005] In the example implementation above, the testing procedure comprises: receiving a request message from the NW for testing the AIML features, functionalities, or models of the UE; performing inference using at least one AIML feature, functionality, or model to generate an inference result; and transmitting a response message to the NW, the response message comprising at least the inference result.
[0006] In any one of the example implementations above, the request message comprises at least one of: an indicator for indicating AIML features, functionalities, or models for testing; and reference input data to the indicated applicable AIML features, functions, or models for testing.
[0007] In any one of the example implementations above, performing the inference comprises processing the reference input data using the indicated AIML features, functionalities, or models to generate the inference result.
[0008] In any one of the example implementations above, the response message comprises at least one of: the indicator for indicating the applicable AIML features, functionalities, or models for testing; or an output data list comprising the inference result.
[0009] In any one of the example implementations above, the response message further comprises a set of UE additional conditions, the UE additional conditions and the inference result being used by the NW to evaluate the one or more AIML features, functionalities, or models.
[0010] In any one of the example implementations above, the AIML configuration comprises an RRC configuration structure.
[0011] In any one of the example implementations above, the AIML configuration is included in a serving cell configuration structure.
[0012] In any one of the example implementations above, the AIML configuration comprises a list of configurations for the one or more AIML features, functionalities, or models.
[0013] In any one of the example implementations above, the AIML configuration comprises a hierarchical listing of configurations for the one or more AIML features, functionalities, or models.
[0014] In any one of the example implementations above, each of the list of configurations for an AIML functionality comprises at least one of: an index or ID of the AIML functionality; an inference related configuration for the AIML functionality; a performance monitoring related configuration for the AIML functionality; or a legacy configuration for a non-AI functionality corresponding to the AIML functionality.
[0015] In any one of the example implementations above, the inference related configuration comprises at least one of: a measurement configuration for obtaining input data for inference for the AIML functionality; an inference reporting configuration for performing inference of output data associated with the AIML functionality; a performance monitoring configuration for inference performance monitoring of the AIML functionality; or a performance reporting configuration for reporting the inference performance monitoring of the AIML functionality.
[0016] In any one of the example implementations above, the AIML configuration is included in a Channel State Information (CSI) report configuraiton.
[0017] In any one of the example implementations above, the CSI report configuration comprises a list of identifiers for the one or more AIML features, functionalities, or models.
[0018] In any one of the example implementations above, the CSI report configuration comprises a report quantity configuration, the report quantity configuration comprising both a set of inference configurations for the one or more AIML features, functionalities, or models.
[0019] In any one of the example implementations above, the CSI report configuration comprises a timing configuration for AIML inference report.
[0020] In any one of the example implementations above, the timing configuration comprises at list one of:periodic reporting; semi persistence reporting on Physical Uplink Shared Channel (PUSCH) ; semi persistent reporting on PUCCH; or aperiodic.
[0021] In any one of the example implementations above, the CSI report configuration comprises a measurement resource configuration for inference for the one or more AIML features, functionalities, or models.
[0022] In any one of the example implementations above, the CSI report configuration comprises a measurement resource configuration for monitoring inference performance of the one or more AIML features, functionalities, or models.
[0023] In any one of the example implementations above, utilizing the one or more AIML features, functionalities, or models comprises activating or deactivating the one or more AIML features, functionalities, or models.
[0024] In any one of the example implementations above, activating or deactivating the one or more AIML features, functionalities, or models is implicitly triggered by an activation or deactivation of a semi-persistent CSI reporting.
[0025] In any one of the example implementations above, utilizing the one or more AIML features, functionalities, or models comprises activating or deactivating at least one of the one or more AIML features, functionalities, or models.
[0026] In any one of the example implementations above, activating or deactivating at least one of the one or more AIML features is triggered by the UE receiving a signaling from the NW.
[0027] In any one of the example implementations above, wherein the signaling comprises a DL MAC CE structure.
[0028] In any one of the example implementations above, the DL MAC CE structure comprises: a first octet with plurality of bits indicating whether to activate or deactivate associated AIML functionalities associated with corresponding serving cell; a second octet having a plurality of bits indicating a number of subsequent octets in the DL MAC CE structure for the corresponding serving cells; and the subsequent octets indicating AIML features and functionalities to be Activated or deactivated for corresponding serving cells.
[0029] In any one of the example implementations above, the method further comprises indicating to lower layers information regarding AI / ML activation or deactivation.
[0030] In any one of the example implementations above, the signaling comprises a semi-persistent CSI reporting activation / deactivation signaling.
[0031] In any one of the example implementations above, the method further comprises indicating to lower layers information regarding AI / ML activation or deactivation.
[0032] In any one of the example implementations above, the method further includes reporting the inference to the NW.
[0033] In any one of the example implementations above, the inference comprises at least one of: top N beams with corresponding N Reference Signal Received Power (RSRP) values, where N ≥1; top N beams with RSRP values of M beams among the top N beams, where N>M≥1; top N beams whose RSRP values are greater than a predefined RSRP threshold value; or a flag for indicating that an RSRP value, if present for a present beam ID, is inferred to measured.
[0034] In another example implementation, a method performed by a network (NW) in communication with a user equipment (UE) in a wireless communication network is disclosed. The method may include: performing a testing procedure for Artificial Intelligence or Machine Learning (AIML) capabilities of the UE in collaboration with the UE; transmitting to the UE an AIML configuration for one or more AIML features, functionalities, or models selected by the NW from AIML capabilities of the UE based at least partially on the testing procedure; and receiving an inference and / or measurement report from the UE, the inference and / or measurement report being generated by the UE based on the AIML configuration.
[0035] In the example implementation above, the testing procedure comprises: transmitting a request message to the UE for testing the AIML features, functionalities, or models of the UE; and receiving a response message from the UE, the response message comprising at least an inference result generated by the UE using at least one AIML feature, functionality, or model.
[0036] In any one of the example implementations above, the request message comprises at least one of: an indicator for indicating AIML features, functionalities, or models for testing; and reference input data to the indicated applicable AIML features, functions, or models for testing.
[0037] In any one of the example implementations above, the response message comprises at least one of: the indicator for indicating the applicable AIML features, functionalities, or models for testing; or an output data list comprising the inference result.
[0038] In any one of the example implementations above, the response message further comprises a set of UE additional conditions, the UE additional conditions and the inference result being used by the NW to evaluate the one or more AIML features, functionalities, or models.
[0039] In any one of the example implementations above, the AIML configuration comprises an RRC configuration structure.
[0040] In any one of the example implementations above, the AIML configuration is included in a serving cell configuration structure.
[0041] In any one of the example implementations above, the AIML configuration comprises a list of configurations for the one or more AIML features, functionalities, or models.
[0042] In any one of the example implementations above, the AIML configuration comprises a hierarchical listing of configurations for the one or more AIML features, functionalities, or models.
[0043] In any one of the example implementations above, each of the list of configurations for an AIML functionality comprises at least one of: an index or ID of the AIML functionality; an inference related configuration for the AIML functionality; a performance monitoring related configuration for the AIML functionality; or a legacy configuration for a non-AI functionality corresponding to the AIML functionality.
[0044] In any one of the example implementations above, the inference related configuration comprises at least one of: a measurement configuration for obtaining input data for inference for the AIML functionality; an inference reporting configuration for performing inference of output data associated with the AIML functionality; a performance monitoring configuration for inference performance monitoring of the AIML functionality; or a performance reporting configuration for reporting the inference performance monitoring of the AIML functionality.
[0045] In any one of the example implementations above, the AIML configuration is included in a Channel State Information (CSI) report configuraiton.
[0046] In any one of the example implementations above, the CSI report configuration comprises a list of identifiers for the one or more AIML features, functionalities, or models.
[0047] In any one of the example implementations above, the CSI report configuration comprises a report quantity configuration, the report quantity configuration comprising both a set of inference configurations for the one or more AIML features, functionalities, or models.
[0048] In any one of the example implementations above, the CSI report configuration comprises a timing configuration for AIML inference report.
[0049] In any one of the example implementations above, the timing configuration comprises at list one of:periodic reporting; semi persistence reporting on Physical Uplink Shared Channel (PUSCH) ; semi persistent reporting on PUCCH; or aperiodic.
[0050] In any one of the example implementations above, the CSI report configuration comprises a measurement resource configuration for inference for the one or more AIML features, functionalities, or models.
[0051] In any one of the example implementations above, the CSI report configuration comprises a measurement resource configuration for monitoring inference performance of the one or more AIML features, functionalities, or models.
[0052] In any one of the example implementations above, wherein the method further includes transmitting a signaling to the UE to activate or deactivate at least one of the one or more AIML features, functionalities, or models.
[0053] In any one of the example implementations above, the signaling comprises a DL MAC CE structure.
[0054] In any one of the example implementations above, wherein the DL MAC CE structure comprises: a first octet with plurality of bits indicating whether to activate or deactivate associated AIML functionalities associated with corresponding serving cell; a second octet having a plurality of bits indicating a number of subsequent octets in the DL MAC CE structure for the corresponding serving cells; and the subsequent octets indicating AIML features and functionalities to be Activated or deactivated for corresponding serving cells.
[0055] In any one of the example implementations above, the signaling comprises a semi-persistent CSI reporting activation / deactivation signaling.
[0056] In any one of the example implementations above, the inference comprises at least one of: top N beams with corresponding N Reference Signal Received Power (RSRP) values, where N ≥1; top N beams with RSRP values of M beams among the top N beams, where N>M≥1; top N beams whose RSRP values are greater than a predefined RSRP threshold value; or a flag for indicating that an RSRP value, if present for a present beam ID, is inferred to measured.
[0057] The UE or the NW of any one of the methods above is disclosed. The UE or the NW may include a processor and a memory, wherein the processor is configured to read computer code from the memory to cause the UE or the NW to perform the method of any one of the methods above.
[0058] A non-transitory computer-readable program medium with computer code stored thereupon is further disclosed. The computer code, when executed by a processor of the UE or the NW of any one of the methods above, is configured to cause the processor to implement any one of the methods above.
[0059] The above embodiments and other aspects and alternatives of their implementations are described in greater detail in the drawings, the descriptions, and the claims below.BRIEF DESCRIPTION OF THE DRAWINGS
[0060] FIG. 1 illustrates an example wireless communication network including a wireless access network, a core network, and data networks.
[0061] FIG. 2 illustrates an example wireless access network including a plurality of mobile stations / terminals or User Equipments (UEs) and a wireless access network node in communication with one another via an over-the-air radio communication interface.
[0062] FIG. 3 shows an example radio access network (RAN) architecture.
[0063] FIG. 4 shows an example communication protocol stack in a wireless access network node or wireless terminal device including various network layers.
[0064] FIG. 5 illustrates an example general procedure for AI / ML life cycle management of UE side AI models.
[0065] FIG. 6 illustrates an example procedure for UE triggered AI / ML capability reporting.
[0066] FIG. 7 illustrates an example uplink MAC CE for reporting available UE AI / ML functionalities.
[0067] FIG. 8 illustrates an example procedure for testing applicable AI / ML functionalities.
[0068] FIGS. 9A-9E illustrate example RRC configuration structures for AI / ML functionalities.
[0069] FIG. 10 illustrates an example DL MAC CE for signaling activation / deactivation of AI / ML functionalities.
[0070] FIG. 11 illustrates an example procedure for UE additional change reporting.DETAILED DESCRIPTION
[0071] The present disclosure will now be described in detail hereinafter with reference to the accompanied drawings, which form a part of the present disclosure, and which show, by way of illustration, specific examples of embodiments. The present disclosure may, however, be embodied in a variety of different forms and, therefore, the covered or claimed subject matter is intended to be construed as not being limited to any of the embodiments to be set forth below.
[0072] Throughout the specification and claims, terms may have nuanced meanings suggested or implied in context beyond an explicitly stated meaning. Likewise, the phrase “in one embodiment” or “in some embodiments” as used herein does not necessarily refer to the same embodiment and the phrase “in another embodiment” or “in other embodiments” as used herein does not necessarily refer to a different embodiment. The phrase “in one implementation” or “in some implementations” as used herein does not necessarily refer to the same implementation and the phrase “in another implementation” or “in other implementations” as used herein does not necessarily refer to a different implementation. It is intended, for example, that claimed subject matter includes combinations of exemplary embodiments or implementations in whole or in part.
[0073] In general, terminology may be understood at least in part from usage in context. For example, terms, such as “and” , “or” , or “and / or, ” as used herein may include a variety of meanings that may depend at least in part upon the context in which such terms are used. Typically, “or” if used to associate a list, such as A, B or C, is intended to mean A, B, and C, here used in the inclusive sense, as well as A, B or C, here used in the exclusive sense. In addition, the term “one or more” or “at least one” as used herein, depending at least in part upon context, may be used to describe any feature, structure, or characteristic in a singular sense or may be used to describe combinations of features, structures or characteristics in a plural sense. Similarly, terms, such as “a” , “an” , or “the” , again, may be understood to convey a singular usage or to convey a plural usage, depending at least in part upon context. In addition, the term “based on” or “determined by” may be understood as not necessarily intended to convey an exclusive set of factors and may, instead, allow for existence of additional factors not necessarily expressly described, again, depending at least in part on context.
[0074] This disclosure is directed generally to wireless communication networks and particularly to configuration and provisioning of Artificial Intelligence (AI) and / or Machine Learning (ML) functionalities and models in both terminal devices and network nodes in wireless communication networks. For example, such AI / ML functionalities and models may reside in either the wireless terminal side or in the wireless network side. The wireless terminal and the network may perform a collaborative procedure to determine, configure, activate, or deactivate a set of AI / ML functionalities and models for adaptive prediction and inference of network provisioning via a sequence of triggered messages.
[0075] Wireless Network Overview
[0076] An example wireless communication network, shown as 100 in FIG. 1, may include wireless terminal devices or user equipment (UE) 110, 111, and 112, a carrier network 102, various service applications 140, and other data networks 150. The wireless terminal devices or UEs, may be alternatively referred to as wireless terminals. The carrier network 102, for example, may include access network nodes 120 and 121, and a core network 130. The carrier network 110 may be configured to transmit voice, data, and other information (collectively referred to as data traffic) among UEs 110, 111, and 112, between the UEs and the service applications 140, or between the UEs and the other data networks 150. The access network nodes 120 and 121 may be configured as various wireless access network nodes (WANNs, alternatively referred to as wireless base stations) to interact with the UEs on one side of a communication session and the core network 130 on the other. The term “access network” may be used more broadly to refer a combination of the wireless terminal devices 110, 111, and 112 and the access network nodes 120 and 121. A wireless access network may be alternatively referred to as Radio Access Network (RAN) . The core network 130 may include various network nodes configured to control communication sessions and perform network access management and traffic routing. The service applications 140 may be hosted by various application servers deployed outside of but connected to the core network 130. Likewise, the other data networks 150 may also be connected to the core network 130.
[0077] In the example wireless communication network of 100 of FIG. 1, the UEs may communicate with one another via the wireless access network. For example, UE 110 and 112 may be connected to and communicate via the same access network node 120. The UEs may communicate with one another via both the access networks and the core network. For example, UE 110 may be connected to the access network node 120 whereas UE 111 may be connected to the access network node 121, and as such, the UE 110 and UE 111 may communicate to one another via the access network nodes 120 and 121, and the core network 130. The UEs may further communicate with the service applications 140 and the data networks 150 via the core network 130. Further, the UEs may communicate to one another directly via side link communications, as shown by 113.
[0078] FIG. 2 further shows an example system diagram of the wireless access network 120 including a WANN 202 serving UEs 110 and 112 via the over-the-air interface 204. The wireless transmission resources for the over-the-air interface 204 include a combination of frequency, time, and / or spatial resource. Each of the UEs 110 and 112 may be a mobile or fixed terminal device installed with mobile access units such as SIM / USIM modules for accessing the wireless communication network 100. The UEs 110 and 112 may each be implemented as a terminal device including but not limited to a mobile phone, a smartphone, a tablet, a laptop computer, a vehicle on-board communication equipment, a roadside communication equipment, a sensor device, a smart appliance (such as a television, a refrigerator, and an oven) , or other devices that are capable of communicating wirelessly over a network. As shown in FIG. 2, each of the UEs such as UE 112 may include transceiver circuitry 206 coupled to one or more antennas 208 to effectuate wireless communication with the WANN 120 or with another UE such as UE 110. The transceiver circuitry 206 may also be coupled to a processor 210, which may also be coupled to a memory 212 or other storage devices. The memory 212 may be transitory or non-transitory and may store therein computer instructions or code which, when read and executed by the processor 210, cause the processor 210 to implement various ones of the methods described herein.
[0079] Similarly, the WANN 120 may include a wireless base station or other wireless network access point capable of communicating wirelessly via the over-the-air interface 204 with one or more UEs and communicating with the core network 130. For example, the WANN 120 may be implemented, without being limited, in the form of a 2G base station, a 3G nodeB, an LTE eNB, a 4G LTE base station, a 5G NR base station of a 5G gNB, a 5G central-unit base station, a 5G distributed-unit base station, or 6G base stations. Each type of these WANNs may be configured to perform a corresponding set of wireless network functions. The WANN 202 may include transceiver circuitry 214 coupled to one or more antennas 216, which may include an antenna tower 218 in various forms, to effectuate wireless communications with the UEs 110 and 112. The transceiver circuitry 214 may be coupled to one or more processors 220, which may further be coupled to a memory 222 or other storage devices. The memory 222 may be transitory or non-transitory and may store therein instructions or code that, when read and executed by the one or more processors 220, cause the one or more processors 220 to implement various functions of the WANN 120 described herein.
[0080] Data packets in a wireless access network such as the example described in FIG. 2 may be transmitted as protocol data units (PDUs) . The data included therein may be packaged as PDUs at various network layers wrapped with nested and / or hierarchical protocol headers. The PDUs may be communicated between a transmitting device or transmitting end (these two terms are used interchangeably) and a receiving device or receiving end (these two terms are also used interchangeably) once a connection (e.g., a radio link control (RRC) connection) is established between the transmitting and receiving ends. Any of the transmitting device or receiving device may be either a wireless terminal device such as device 110 and 120 of FIG. 2 or a wireless access network node such as node 202 of FIG. 2. Each device may both be a transmitting device and receiving device for bi-directional communications.
[0081] The core network 130 of FIG. 1 may include various network nodes geographically distributed and interconnected to provide network coverage of a service region of the carrier network 102. These network nodes may be implemented as dedicated hardware network nodes. Alternatively, these network nodes may be virtualized and implemented as virtual machines or as software entities. These network nodes may each be configured with one or more types of network functions which collectively provide the provisioning and routing functionalities of the core network 130.
[0082] Returning to wireless radio access network (RAN) , FIG. 3 illustrates an example RAN 340 in communication with a core network 310 and wireless terminals UE1 to UE7. The RAN 340 may include one or more various types of wireless base station or WANNs 320 and 321 which may include but are not limited to gNB, eNodeB, NodeB, or other type of base stations. The RAN 340 may be backhauled to the core network 310. The WANNs 320, for example, may further include multiple separate access network nodes in the form of a Central Unit (CU) 322 and one or more Distributed Unit (DU) 324 and 326. The CU 322 is connected with DU1 324 and DU2 326 via various interfaces, for example, an F1 interface. The F1 interface, for example, may further include an F1-C interface and an F1-U interface, which may be used to carry control plane information and user plane data, respectively. In some embodiments, the CU may be a gNB Central Unit (gNB-CU) , and the DU may be a gNB Distributed Unit (gNB-DU) . While the various implementations described below are provided in the context of a 5G cellular wireless network, the underlying principles described herein are applicable to other types of radio access networks including but not limited to other current or future generations of cellular network such as 4G LTE and 6G network, as well as Wi-Fi, Bluetooth, ZigBee, and WiMax networks.
[0083] The UEs may be connected to the network via the WANNs 320 over an air interface. The UEs may be served by at least one cell. Each cell is associated with a coverage area. These cells may be alternatively referred to as serving cells. The coverage areas between cells may partially overlap. Each UE may be actively communicating with at least one cell while may be potentially connected or connectable to more than one cell. In the example of FIG. 1, UE1, UE2, and UE3 may be served by cell1 330 of the DU1, whereas UE4 and UE5 may be served by cell2 332 of the DU1, and UE6 and UE7 may be served by cell3 associated with DU2. In some implementations, a UE may be served simultaneously by two or more cells. Each of the UE may be mobile and the signal strength and quality from the various cells at the UE may depend on the UE location and mobility.
[0084] In some example implementations, the cells shown in FIG. 3 may be alternatively referred to as serving cells. The serving cells may be grouped into serving cell groups (CGs) . A serving cell group may be either a Master CG (MCG) or Secondary CG (SCG) . Within each type of cell groups, there may be one primary cell and one or more secondary cells. A primary cell in a MSG, for example, may be referred to as a PCell, whereas a primary cell in a SCG may be referred to as PScell. Secondary cells in either an MCG or an SCG may be all referred to as SCell. The primary cells including PCell and PScell may be collectively referred to as spCell (special Cell) . All these cells may be referred to as serving cells or cells. The term “cell” and “serving cell” may be used interchangeably in a general manner unless specifically differentiated. The term “serving cell” may refer to a cell that is serving, will serve, or may serve the UE. In other words, a “serving cell” may not be currently serving the UE. While the various embodiment described below may at times be referred to one of the types of serving cells above, the underlying principles apply to all types of serving cells in both types of serving cell groups.
[0085] FIG. 4 further illustrates a simplified view of the various network layers involved in transmitting user-plane PDUs from a transmitting device 402 to a receiving device 404 in the example wireless access network of FIGS. 1-3. FIG. 4 is not intended to be inclusive of all essential device components or network layers for handling the transmission of the PDUs. FIG. 4 illustrates that the data packaged by upper network layers 420 at the transmitting device 402 may be transmitted to corresponding upper layer 430 (such as radio resource control or RRC layer) at the receiving device 304 via Packet Data Convergence Protocol layer (PDCP layer, not shown in FIG. 4) and radio link control (RLC) layer 422 and of the transmitting device, the physical (PHY) layers of the transmitting and receiving devices and the radio interface, as shown as 406, and the media access control (MAC) layer 434 and RLC layer 432 of the receiving device. Various network entities in each of these layers may be configured to handle the transmission and retransmission of the PDUs.
[0086] In FIG. 4, the upper layers 420 may be referred as layer-3 or L3, whereas the intermediate layers such as the RLC layer and / or the MAC layer and / or the PDCP layer (not shown in FIG. 4) may be collectively referred to as layer-2, or L2, and the term layer-1 is used to refer to layers such as the physical layer and the radio interface-associated layers. In some instances, the term “low layer” may be used to refer to a collection of L1 and L2, whereas the term “high layer” may be used to refer to layer-3. In some situations, the term “lower layer” may be used to refer to a layer among L1, L2, and L3 that are lower than a current reference layer. Control signaling may be initiated and triggered at each of L1 through L3 and within the various network layers therein. These signaling messages may be encapsulated and cascaded into lower layer packages and transmitted via allocated control or data over-the-air radio resources and interfaces. The term “layer” generally includes various corresponding entities thereof. For example, a MAC layer encompasses corresponding MAC entities that may be created. The layer-1, for example, encompasses PHY entities. The layer-2, for another example encompasses MAC layers / entities, RLC layers / entities, service data adaptation protocol (SDAP) layers and / or PDCP layers / entities.
[0087] AI / ML Assisted Wireless Network Provisioning and Configuration
[0088] AI and ML (referred to AL generally) may facilitate more efficient configuration and provisioning in wireless networks. At the core of a general AI framework are various AI models. An AI model generally contains a large number of model parameters that are determined through a training process where correlations in a set of training data are learned and embedded in the trained model parameters. The trained model parameters may thus be used to generate inference or predictions from a set of input dataset. AI models are particularly suitable for situations where there is few trackable deterministic or analytical derivation paths between input data and output but correlations in the data may be identified from historical data and may be embedded into the AI model via training processes.
[0089] In a wireless communication system such as the ones described above, determination of adaptive network configuration may rely on empirical characteristics and may further require lengthy measurement processes and / or significant amounts of computation power. Such types of configurations may include but are not limited to over-the-air interface beam management, channel state information (CSI) feedback compression and decompression, and wireless terminal positioning. Correlation between various network conditions and these adaptive configurations may be learned via AI techniques. The use of AI models for assisting in network configuraiton may thus help reduce the amount of measurements and computation requirement, providing a more agile and more efficient network configuration. Accordingly, it may thus be desirable to provide a mechanism for provisioning of AI models according to AI capabilities of terminal devices and network conditions in assisting in adaptively determining these network configurations.
[0090] For example, AI technology may be applied to beam management in the over-the-air communication interface. In current implementations, beam management typically relies on the exhaustive searching beam sweeping and measurements. In other words, the network (NW) may perform a full sweep of the beams by sending sufficient number of reference signals. A UE may be configured to monitor and measure each reference signal and then report the measurement result to NW for the NW to decide the best beam for the UE to switch to. This process, however, is resource and power intensive and time consuming. With trained AI models that embed learned correlation between various network condition parameters, fewer measurements (or fewer reference signals) may be needed in order to infer the best beams with reasonable accuracy, enabling the so-called as the AI / ML based spatial beam management. In some implementations, AI model may help identify inference of best candidate beams in a future time instance using other network conditions and then only sweep and measure the candidate beams to select the beam for use in the future time instance, enabling the so-called AI / ML based temporal beam management. Additionally, as beam configuration is closely tied to a location of the UE, AI technology may further be used for inferring or predicting UE trajectory or location, thereby indirectly help selection of best beams.
[0091] For another example, AI technologies may be applied to channel state information (CSI) feedback. Traditionally, the CSI feedback may be implemented using a codebook known by UE and NW. The UE may measure the CSI and obtain a measurement result, and then map the measurement result to a closest vector of the codebook, and transmit the index of that vector to the NW in order to save the air-interface resource consumption. However, because the codebook is not unlimited or dynamic changeable over time, there would be always mismatch so some degree, thereby causing un-controlled CSI feedback errors as the wireless environment varies. AI thus may be applied to, for example, compression-decompression for CSI feedback. Specifically, a CSI report may be compressed by a UE-side AI model and decompressed by a corresponding NW-side AI model. Such AI models may be initially trained and continuously developed over time and accumulation of network conditions. In some example implementations, AI model may help identify inference of future CSI feedback value using other network conditions and then only measure the current reference signaling for CSI, enabling the so-called as the AI / ML based CSI prediction.
[0092] For yet another example, AI technology may be applied to UE positioning. Traditional approaches for UE positioning depend on Positioning Reference Signal (PRS) or Sounding Reference Signal (SRS) . e.g., DL Positional Reference Signal and uplink Sounding Reference Signal. Regardless of the alternative approaches, the Line-Of-Sight (LOS) beams are the key beams to identify in order to generate the most precise location estimation by triangulation at the NW side. However, in most cases, it is difficult to identify the LOS beams from other Non-Line-Of-Sight (NLOS) beams, thereby providing inaccurate UE positioning. A trained AI model, on the other hand, may identify various pattern and correlation in the PRS and SRS for extracting LOS information and providing more accurate UE positioning.
[0093] In some example implementations, AI models may reside on the wireless terminal side. In some other implementations, AI model may instead reside on the network (NW) side. AI models may be provided as a service. AI models may be retrained and updated as needed. The UE or the NW may determine what models to retrieve or retain and how they are updated and configured for assisting in the wireless communications, including but not limited to the beam management, CSI feedback, UE positioning, and the like. The selection, retrieval, storage, and execution of the AI models may be dependent on the capabilities of the terminal device and / or the NW. The terminal and the NW may be configured to communicate such capabilities as well as network conditions in order to provision the selection, configuration, and usage of the AI models over time, generally referred to as AI / ML Life Cycle Management (LCM) .
[0094] AI / ML Life Cycle Management (LCM) For UE Side AI Feature / Functionality / Model
[0095] The AI / ML LCM may be performed at various different levels or granularities. In one example implementation, the AI / ML LCM may be performed in a level of AI / ML based feature groups, for example, AI / ML based beam management group (e.g., Feature Group) including AI / ML based spatial beam management aspects (e.g., features) and AI / ML based temporal beam management aspects (e.g., features) , AI / ML base CSI feedback group including AI / ML based CSI feedback enhancement aspects (e.g., features) and AI / ML based CSI prediction aspects (e.g., features) . In another example implementation, the AI / ML LCM may be performed in a level of AI / ML based features. In another example implementation, the AI / ML LCM may be performed at a level of AI / ML functionalities. In yet another example implementation, the AI / ML LCM may be performed at a level of AI / ML models. The terms feature group, feature, functionality, and model for AI / ML are merely used to represent the various levels at which AI / ML can be configured and applied. They may be hierarchically related. They may overlap in some situations. They may be delineated in any suitable manner in order to facilitate the configuration and management of the usage of AI / ML. Each of these levels may be itself hierarchical. For example, an AI / ML model may include lower level AI / MI models as internal components. An AI / ML feature may include lower-level sub-features, and likewise, an AI / ML functionality may include lower-level sub-functions.
[0096] For example, an AI / ML model may refer to a specific trained algorithm that process one or more input to generate a prediction as an output. An AI / MI model may include utilize components such as various types of neural networks, regression algorithms, support vector machine algorithms, K-Nearest Neighbors algorithms, random forest, K-means clustering, principle component analysis, Bayesian networks, and the like.
[0097] For example, an AI / ML model may be trained to support a particular AI feature or AI functionality. An AI / ML feature or AI / ML functionality may be achieved by different AI models which may differ in their internal architecture, hyper parameters, trained model parameters, inputs, computation resource requirements, complexity, prediction accuracy, and the like. For example, an AI / ML functionality may encompass one or one set of AI / ML models.
[0098] For example, an AI / ML based feature may encompass one or more AI / MI functionalities. In some example implementations, AI / ML feature may be synonymous to AI / MI functionality (e.g., an AI / ML functionality may represent a corresponding AI / ML based feature) . For another example, an AI / ML based feature group may include one or more AI / ML based features to form a category of features.
[0099] Merely as an example, AI / ML features may refer to a delineation of AI / ML based spatial beam prediction, AI / ML based temporal beam prediction, AI / ML based CSI feedback compressing and decompression, AI / ML based CSI prediction, AI / ML based temporal / spatial cell measurement result prediction for mobility, the AI / ML based temporal / spatial beam prediction for mobility, and the like.
[0100] In some example implementations, an AI / ML LCM may include at least one of the following non-limiting aspects:
[0101] ● AI / ML feature (feature group) / functionality / model identification.
[0102] ● Applicable AI / ML feature / functionality / model reporting between wireless terminals and NWs.
[0103] ● AI / ML feature / functionality / model control (including selection, activation, deactivation, fallback) .
[0104] ● AI / ML feature / functionality / model performance monitoring.
[0105] AI / ML features / functionalities / models may reside on either the UE side or on the NW side. LCM of UE side AI / MI features / functionalities / models may depend on UE capabilities.
[0106] UE-Side AI / ML LCM -General
[0107] An example implementation for the above LCM aspects with wireless terminal side (UE side) AI / ML models involving the UE 502 and the NW 504 is shown in FIG. 5, including the following general steps:
[0108] ● STEP 1: A UE capability reporting procedure may be performed between the UE 502 and the NW 504 for one or more AI / ML based features / functionalities / models.
[0109] ● STEP 2: The NW 504 and the UE 502 may additionally perform applicable AI / ML features / functionalities / models reporting procedure.
[0110] ● STEP 3: The NW 504 and the UE 502 may perform a preparation stage to test the AI / ML features / functionalities / models.
[0111] ● STEP 4: The NW 504 may configure RRC configuration of AI / ML features / functionalities / models to the UE 502 according to the UE capability and / or applicable features / functionalities / models reporting.
[0112] ● STEP 5: The NW 504 may send a message to The UE 502 to activate / deactivate an AI / ML feature / functionality / model.
[0113] ● STEP 6: The UE 502 may accordingly activate or deactivate an AI / ML feature / functionality / model and perform predictions and inferences.
[0114] ● STEP 7: The UE 502 may then perform prediction / reference or actual measurements and transmit predictions and inferences or actual measurements to the NW 504 via one or more inference reports and / or measurement reports.
[0115] Thus, in the general example procedure above, the UE may first report to the NW set of supportable AI / ML based features / functionalities / models based on UE capabilities. The UE and the NW may then collaboratively determine (1) applicable or suitable features / functionalities / models among the supportable AI / ML based features / functionalities / models according to capabilities and network conditions and a variety of other factors for the UE to activate, and / or (2) features / functionalities / models among the supportable features / functionalities / models that are no longer applicable or suitable for the UE to deactivate.
[0116] UE-Side AI / ML LCM –UE AI / ML Capability Reporting
[0117] In STEP 1 of FIG. 5 above, the capability reporting procedure for one or more AI / ML based features / functionalities / models may be triggered either by the UE or by the NW, as shown exemplarily in FIG. 6. The step 608 is illustrated for NW triggered AI / ML capability reporting, whereas step 606 would precede 608 for UE triggered AI / ML capability reporting.
[0118] Specifically, in one example implementation, the UE 602 may proactively trigger an AI / ML capability reporting by transmitting a triggering message in 606 to the NW (e.g., a base station or core network node) . Then the NW may accept the request and initiate the UE AI / ML capability reporting procedure as shown in 608. In some other example implementation, the UE AI / ML capability reporting may be triggered by the NW and then executed between the UE and the NW as shown in 608, without the UE triggering step of 606.
[0119] In some example implementations, the UE may support a superset of AI / ML features / functionalities / models (supportable features / functionalities / models) . In some cases, the UE may only download / store locally a subset of the superset models depending on its locations and other factors (available features / functionalities / models) . While the superset supportable features / functionalities / models may be stable over time, it may nevertheless change. Even if the superset features / functionalities / models may be stable, the available features / functionalities / models may be more fluid and may change more frequently over time. The UE thus may determine to trigger the UE AI / ML capability reporting in 606 under at least one of the following cases or conditions:
[0120] ● The superset supported AI / ML features / functionalities / models has changed at the UE, which may include but is not limited to:
[0121] ○ At least one of the supported AI / ML features / functionalities / models has been updated since the latest UE capability reporting.
[0122] ○ At least one new AI / ML feature / functionality / model have become supported since the latest UE capability reporting.
[0123] ○ At least one AI / ML feature / functionality / model have been removed from the superset since the latest UE capability reporting.
[0124] ○ An overall computation resources for AI / ML feature / functionality / model in the superset has changed since the latest UE capability reporting.
[0125] ● The actually stored AI / ML features / functionalities / models at the UE have changed, which may include and is not limited to:
[0126] ○ At least one of the AI / ML features / functionalities / models stored at UE has been updated since the latest UE capability reporting.
[0127] ○ At least one new AI / ML feature / functionality / model has been obtained by UE since the latest UE capability reporting.
[0128] ○ At least one stored feature / functionality / model has been removed from UE since the latest UE capability reporting.
[0129] ○ The overall computation resources for the available AI / ML feature / functionality / model has changed since the latest UE capability reporting.
[0130] ● At least one of the AI / ML features / functionalities / models has been configured by the NW for the UE.
[0131] In some example implementations, the triggering message of 606 by the UE may be transmitted using one of the following formats: an uplink (UL) MAC Control Element (MAC CE) ; or a UL RRC signaling (e.g., a UE Assistance Information (UAI) message) or a protocol signaling terminated between an AI / ML logical layer and an NW logical entity / unit.
[0132] In some example implementation, such triggering message of 606 by the UE may contain or indicate at least one of the following information or information items:
[0133] ● One or more AI / ML feature / functionality / model indications: to indicate AI / ML features / functionalities / models that have changed.
[0134] ● The change type indication, where the types of change may include, for example, 1) an addition of AI / ML feature / functionality / model; 2) an update of the stored AI / ML feature / functionality / model; 3) a release of the AI / ML feature / functionality / model from the supported AI / ML features / functionalities / models.
[0135] The information about in the UE triggering message of 606 may be used by the NW to determine whether and when to actually request for UE AI / ML capability report in 608 of FIG. 6.
[0136] In 608 of FIG. 6, the NW may transmit an enquiry message to the UE for UE AI / ML capability report (either as a triggering message for UE AI / ML capability report by the NW, or as a response to the triggering message from the UE in 606 when the NW determine to request the report) . The enquiry message may be referred to as UECapabityEnquiring, as indicated in FIG. 6.
[0137] In response, the UE may then transmit the UE AI / ML capability report to the NW. The message that contains the report may be referred to as UECapability, as indicated in FIG. 6. In some example implementation, the message of UECapability may include or indicate at least one of the following information or information items:
[0138] ● A supported AI / ML based feature or feature group (list) : to indicate the list of AI / ML based feature that is supported by UE (these features, while supported, however, may or may not be available at the UE) .
[0139] ● One of more supported AI / ML based functionality lists: to indicate UE supported AI / ML based functionalities list for an AI / ML based feature (these functionalities, while supported, however, may or may not be available at the UE) .
[0140] ● One of more supported AI / ML model lists: to indicate UE supported AI / ML model lists for AI / ML based features or functionalities.
[0141] ● Bands or band list indication: to indicate supported frequency bands for each of the AI / ML based features / functionalities / models.
[0142] ● Indicator or indicators for computation resource consumption for each AI / ML feature / functionality / model: to indicate a quantification value that an AI / ML based feature / functionality / model would consume in terms of computing resources, if activated.
[0143] ● An overall computation resource indicator for AI / ML: to indicate the maximum quantification of computation resources the UE can support for AI / ML.
[0144] ● Indicator or indicators for preparation time for each AI / ML feature / functionality / model: to indicate the maximum or minimum preparation time from the reception of activation signaling for an AI / ML feature / functionality / model by UE to until when the AI / ML feature / functionality / model inference can actually be executed.
[0145] ● Training configuration (e.g., UE settings, gNB settings) , scenario, conditions indication: to indicate the configurations, scenario, conditions under which the AI / ML based models within the UE supported AI / ML based feature / functionality / model are trained, including but not limited to:
[0146] ○ For AI / ML based spatial beam prediction and AI / ML based temporal beam prediction:
[0147] ■ Information items related to the scenario aspect:
[0148] ● Inter-Site Distance (ISD) to indicate cell range associated with the AI / ML training, the supported value of ISD may be, for example, UMa, UMi, 200 m, 500 m.
[0149] ● Antenna height of base station of the cells associated with the AI / ML training, the supported value of antenna height of based station may be 1 m, 2 m, 5 m, 10 m, etc.
[0150] ● NLOS probability for NLOS radio propagation associated with the AI / ML training.
[0151] ● Indoor / outdoor indicators for indicating an indoor or outdoor scenario and / or indoor / outdoor ratio for the AI / ML training.
[0152] ■ Information items for gNB settings:
[0153] ● DL Tx beam codebook indication to indicate the DL TX beam codebook that is used for AI / ML model training.
[0154] ● Indication for gNB antenna array dimensions associated with the AI / ML model training.
[0155] ● A down tilt of the gNB antenna associated with the AI / ML model training.
[0156] ■ Information items for training beam Conditions and beam set (Set A may refer to group of beams where the best K beams that are predicted by the AI / ML model / functionality, Set B may refer to the beams which are measured and the corresponding measurements and / or associated beam Id may be used as input of the AI / ML model / functionality , whereas a full beam set may refer to all beams, e.g., all 64 beams in certain configuration) :
[0157] ● Beam pattern of set B indication: to indicate a beam pattern of set B compare to the Set A. For example, set B for training may be a 1 / 4 subset of the set A in a manner of even distribution. For another example, set B may not be a subset of the set A. For yet another example, set B may be SSB and Set A may be CSI-RS, or vice versa, etc.
[0158] ● Beam pattern of the Set A: to indicate a beam pattern of set A compare to the full beam set (e.g., 64 beams) . For example, beam set A may be a 1 / 4 subset of the full beam set in a manner of even distribution, or beam set A may be a full beam set etc.
[0159] ■ Information items for UE settings associated with training of the model, such as:
[0160] ● UE speed: to indicate the UE speed information related to the AI / ML feature / functionality / model training
[0161] ● UE orientation: to indicate the UE orientation related to the AI / ML feature / functionality / model training. In one implementation, it indicates a maximum value of UE orientation change for the associated AI / ML feature / functionality / model, e.g., 45, 90, 120, 180 degrees, etc.
[0162] ● UE RX codebook.
[0163] ● UE antenna array dimensions.
[0164] ■ A parameter (set) explicitly including one or more above mentioned information items above with an ID.
[0165] ■ A parameter set implicitly indicating one or more information above with an ID, for example, a cell identifier (e.g., CGI, PCI, etc. ) which may include the information items regarding NW specific configuration / setting, Cell scenario, etc.
[0166] ○ For AI / ML based CSI feedback compression / decompression (e.g., two-side model) :
[0167] ■ A model ID or model ID list: to indicate model IDs for a list of UE supported UE portion of two-side AI / ML models.
[0168] ■ Training dataset ID or training data ID list: to indicate dataset IDs for a list of datasets used for training the UE portion of two-side AI models.
[0169] ■ UE setting indication: to indicate a transceiver unit (TxRU) mapping pattern, e.g., [2, 8, 2] , [4, 4, 2] , etc.
[0170] ■ Scenario indication: to indicate training scenarios for the UE supported UE portion of two-side functionality / model, which may include at least one of:
[0171] ● Inter-Site Distance (ISD) to indicate cell range associated with the AI / ML training, the supported value of ISD may be, for example, UMa, UMi, 200 m, 500 m.
[0172] ● The outdoor / indoor indication to indicate indoor, outdoor, the and indoor / outdoor ration for the training.
[0173] ● Frequency range or band for the training.
[0174] ○ For AI / ML based temporal CSI prediction:
[0175] ■ Scenario aspect information items similar to the list above for the AI / ML based temporal beam management.
[0176] ■ gNB setting information items similar to the list above for the AI / ML based temporal beam management.
[0177] ■ UE setting information items similar to the list above for the AI / ML based temporal beam management.
[0178] In some example implementations, a signaling structure of the UE capability report of 608 for AI / ML based features / functionalities / models may be carried in an RRC information element of UE-NR-Capability or supportedBandListNR in RF-Parameters. In some alternative example implementations, the UE capability report for the features, functionalities, and models may be separated or bifurcated. Merely as one example, the information of the report related to AI / ML based features or functionality may be carried in RRC information element of UE-NR-Capability whereas the AI / ML functionalities corresponding to the AI / ML based feature or AI / ML models corresponding to the AI / ML functionality may be carried and transmitted in the RRC information element of supportedBandListNR in RF-parameters.
[0179] In some example implementations, a signaling structure of the UE capability report of 608 for AI / ML based features / functionalities / models may be carried in a protocol signaling terminated between an AI / ML logical layer and an NW logical entity / unit.
[0180] In one example implementation of coordinating the computation power for AI / ML between MN and SN in DC (e.g., in dual connectivity) scenario, one inter-node information between MN and SN regarding the computation power for AI / ML may be introduced. The SN / MN may send an information to the MN / SN with the current occupied computation power for activated AI / ML features / functionalities / models in order not to overwhelm the maximum computation power reported in the UE capability. In one example implementation, such information from MN to SN may be a parameter that describes the quantification values of computation power that has been consumed by the MN. In one example implementation, such information from SN to MN may be a parameter that describes the quantification values of computation power that has been consumed by the MN. In another example implementation, such information from MN to SN may be a parameter that describes the maximum quantification values of computation power for AI / ML that can be consumed by the SN.
[0181] In another example implementation of coordinating the computation power for AI / ML between MN and SN in a DC scenario, one UL MAC CE may be introduced. The UL MAC CE may be triggered by a MAC entity with at least one of the following conditions:
[0182] ● An AI / ML feature / functionality / model is activated or deactivated for the MAC entity.
[0183] ● The total computation power quantification values for the activated AI / ML features / functionalities / models in the MAC entity is greater or equal to a pre-configured / pre-defined maximum value.
[0184] In one example implementation of generating the UL MAC CE, at least one of the following conditions may need to be met:
[0185] ● A UL grant has been received for a MAC entity other than the MAC entity where the UL MAC CE is triggered.
[0186] ● the UL grant have an ability of accommodating the UL MAC CE according to the LCP procedure.
[0187] ● A UL grant has been received for any one of the MAC entities.
[0188] In one example implementation, the UL MAC CE may include at least one of the following information:
[0189] ● The occupied quantification value of the computation power for AI / ML;
[0190] ● The maximum quantification value of the computation power for AI / ML;
[0191] ● The remaining quantification value of the computation power for AI / ML.
[0192] In one example implementation of coordinating the computation power for AI / ML between the first NW and the second NW in Multiple SIM scenario, the UAI (e.g., UE assistance information) message from UE to NW may be used.
[0193] In one example implementation of configuring the UAI by NW to the UE, the configuration of the UAI may include at least one of the following information:
[0194] ● A minimum quantification value of the computation power for AI / ML: to indicate the minimum quantification value of remaining computation power.
[0195] ● A maximum quantification value of the computation power for AI / ML: to indicate the maximum quantification value of remaining computation power.
[0196] ● A prohibit timer: a time length of a timer to prohibit the triggering of the UAI when it is running.
[0197] In one example implementation of triggering the UAI, at least one of the following conditions may need to be met:
[0198] ● The remaining quantification value of computation power for AI / ML is less than or equal to the pre-configured minimum quantification value.
[0199] ● No UAI has been sent since the remaining quantification value of computation power has changed to be less than a pre-configured threshold value.
[0200] ● The remaining quantification value of computation power for AI / ML is greater than or equal to the pre-configured maximum quantification value.
[0201] ● No UAI has been sent since the remaining quantification value of computation power has been changed to be greater than or equal to a pre-configured threshold value
[0202] ● The prohibit timer is not running.
[0203] ● In one example implementation, the contents of the UAI may include the remaining quantification value of the computation power for UE side AI / ML
[0204] In one example implementation of transmitting the UAI, the prohibit timer may be started / restarted. UE-Side AI / ML LCM –Applicable AI / ML Feature / Functionality / Model Determination and Reporting
[0205] Returning FIG. 5, STEP 2 may be performed for determining applicable UE AI / ML features / functionalities / models among supported UE AI / ML features / functionalities / models. Such a procedure may be implemented for NW to further determine the applicable UE AI / ML features / functionalities / models according to network conditions and / or actually available AI / ML features / functionalities / models those has been stored at the UE (rather than all supported UE AI / ML features / functionalities / models) . Such a procedure would generally involve the NW transmitting a set of additional conditions (e.g., network conditions) and / or the request of applicable UE AI / ML features / functionalities / models reporting to the UE and the UE determining a set of applicable UE AI / ML features / functionalities / models according to the additional conditions and / or the available AI / ML features / functionalities / models at the UE, and reporting the determined set of applicable UE AI / ML features / functionalities / models to the NW.
[0206] The procedure of the STEP 2 of FIG. 5 may be implemented using two example alternative solutions for messaging, which are indicated as Solution 1 and Solution 2 in FIG. 5.
[0207] In a first solution, for example, message (s) for handling the interaction between the UE and the NW for the determination and reporting of applicable UE AI / ML features / functionalities / models may be based on signaling or messaging formats including but not limited to (1) RRC message for transmission of conditions by the NW, and / or (2) a message response RRC message for the UE to transmit the applicable UE AI / ML feature / functionality / model report in response to the conditions or the request sent by the NW, as shown as Solution 1-1 and 1-2 in FIG. 5. In a second alternative solution, for example, message (s) for handling the interaction between the UE and the NW for the determination and reporting of applicable UE AI / ML features / functionalities / models may be based on RRC reconfiguration messages and corresponding RRC reconfiguration response messages, shown as Solution 2-1 and 2-2 in FIG. 5.
[0208] In some example implementations of the transmission of conditions of the first solution, the conditions may be transmitted by the NW to the UE via message formats including but not limited to (1) RRC system information message for transmission of conditions and / or request of applicable feature / functionality / model reporting by the NW, and / or (2) a message referred to as RequestApplicablilityReporting for the transmitting the conditions from the NW to the UE.
[0209] In some example implementations, the RRC system information messaging above may include or indicate at least one of the following information or information items when providing additional conditions for the UE to determine applicable UE AI / M features / functionalities / models:
[0210] ● AI / ML based feature group indication: to indicate the AI / ML based feature group for which the NW additional conditions may be provided or for which the applicable feature / functionality / model reporting are requested. In one example implementation, the AI / ML based feature group may be hard coded. For example, the AI / ML based feature may be represented by some IDs. Merely as examples, AI / ML based feature or with Id=0 may indicate AI / ML based beam management feature group. AI / ML based feature with Id =1 may indicate AI / ML based CSI feature group, etc.
[0211] ● In one implementation, the AI / ML based feature may be indicated by an enumerate type parameter with example values of {beam management, CSI, CSI prediction, spare 1... } .
[0212] ● AI / ML based feature indication: to indicate the AI / ML based feature for which the NW additional conditions may be provided or for which the applicable functionality / model reporting are requested:
[0213] ○ In one implementation, the AI / ML based feature may be hard coded. For example, the AI / ML based feature may be represented by some IDs. Merely as examples, AI / ML based feature or with Id=0 may indicate AI / ML based spatial beam management, AI / ML based feature with Id =1 may indicate AI / ML based temporal beam management, etc.
[0214] ○ In one implementation, the AI / ML based feature may be indicated by an enumerate type parameter with example values of {spatial beam management, temporal beam management, CSI feedback compression / decompression, CSI prediction, spare 1... } .
[0215] ● NW additional conditions or indication of the NW additional conditions associated with an indicated AI / ML based feature or feature group to indicate the NW additional conditions regarding the AI / ML based feature / feature group. At least one of the following information items may be indicated in the system message as NW additional conditions:
[0216] ○ AI / ML based feature indication: to indicate the AI / ML based feature the message or the NW additional conditions are for.
[0217] ○ ISD to indicate cell range, for example, UMa, UMi, 200 m, 500 m.
[0218] ○ Antenna height to indicate the Antenna height of base station of the cell.
[0219] ○ NLOS probability: to indicate a probability of the NLOS radio propagation.
[0220] ○ Indoor / outdoor indication to indicate whether the current scenario is indoor or outdoor, or an indoor / outdoor ratio.
[0221] ○ DL Tx beam codebook indication: to indicate a DL TX beam codebook that is used for AI / ML models training.
[0222] ○ gNB antenna array dimensions indication.
[0223] ○ A down tilt of the gNB antenna.
[0224] ○ A beam pattern of set B indication to indicate the beam pattern of set B compare to the Set A (see above) . For example, set B may be a 1 / 4 subset of the set A, etc.
[0225] ○ The beam pattern of the Set A for the AI / ML based functionality to indicate the beam pattern of set A compare to the full beam set (see above) . For example, set A may be a 1 / 4 subset of full set, etc.
[0226] Likewise, in some example implementations of the first solution above for transmitting the conditions using the RequestApplicablilityReporting message, at least one of the following information or information items may be indicated / included in the example RequestApplicablilityReporting message transmitted from the NW to the UE:
[0227] ● AI / ML supported indication: to indicate the UE side AI / ML based feature / FG is supported by the NW.
[0228] ● AI / ML feature group indication: to indicate the AI / ML feature group (e.g., AI / ML based beam management, AI / ML based CSI, AI / ML based positioning) where the NW additional conditions may be provided or belonging to which the applicable AI / ML features are request.
[0229] ● AI / ML based feature indication: to indicate the AI / ML based feature for which the NW additional conditions may be provided or for which the applicable functionality / model reporting are requested.
[0230] ○ In one example implementation, the AI / ML based feature / feature group may be indicated via an AI / ML based feature index / identity which has been indicated by the UE capability.
[0231] ○ In one example implementation, the AI / ML based feature / feature group may be indicated via an AI / ML based feature index / identity according to the order of entries in a supported AI / ML feature List in the UE capability. For example, the AI / ML based feature / FG index=0 represents the first entry of the supported AI / ML based feature / FG list. The AI / ML based feature / FG index=1 represents the second entry of the supported AI / ML based feature / FG list, and so on.
[0232] ○ In one example implementation, the AI / ML based feature / feature group may be indicated by an enumerate type parameter with example values of {spatial beam management, temporal beam management, CSI feedback compression / decompression, CSI prediction, spare 1... } .
[0233] ○ In one example implementation, the AI / ML based feature / feature group may be indicated via a bit string type parameter according to the order of entries in a supported AI / ML feature List in the UE capability. For example, the leftmost or rightmost bit in the bit string represents the first entry of the supported AI / ML based feature / FG list, and the second leftmost or rightmost bit in the bit string represents the second entry of the supported AI / ML based feature / FG list, and so on. The applicable reporting corresponding to AI / ML based feature / FG is requested if the corresponding bit is set to 1.
[0234] ● NW additional conditions of indication of NW additional conditions associated with an indicated AI / ML based feature or feature group to indicate the NW additional conditions regarding the AI / ML based feature / feature group. at least one of the following information may be indicated in the RequestApplicablilityReporting message as NW additional conditions:
[0235] ○ AI / ML based feature indication: to indicate the AI / ML based feature the message or the NW additional conditions are for.
[0236] ○ ISD to indicate cell range, for example, UMa, UMi, 200 m, 500 m.
[0237] ○ Antenna height to indicate the Antenna height of base station of the cell.
[0238] ○ NLOS probability to indicate a probability of the NLOS radio propagation.
[0239] ○ Indoor / outdoor indication to indicate whether the current scenario is indoor or outdoor, or an indoor / outdoor ratio.
[0240] ○ DL Tx beam codebook indication to indicate a DL TX beam codebook that is used for AI / ML models training.
[0241] ○ gNB antenna array dimensions indication.
[0242] ○ A down tilt of the gNB antenna.
[0243] ○ A beam pattern of set B indication to indicate the beam pattern of set B compare to the Set A (see above) . For example, set B may be a 1 / 4 subset of the set A, etc.
[0244] ○ The beam pattern of the Set A for the AI / ML based functionality to indicate the beam pattern of set A compare to the full beam set. For example, set A may be a 1 / 4 subset of full set, etc.
[0245] In some example implementations, the reporting of the applicable UE features / functionalities / models by the UE to the NW in Solution 1-2 of FIG. 5 may be implemented in at least one of the following UL RRC messaging formats or MAC protocol signaling:
[0246] ● Option 1: A UE Assistant Information (UAI) message.
[0247] ● Option 2: A dedicated RRC message, referred to as UEApplicableFunctionalityReporting.
[0248] ● Option 3: UL MAC CE.
[0249] For the Option 1 above, for example, the UAI message for reporting applicable functionalities / models for an AI / ML based feature / feature group or for reporting applicable features may be triggered by meeting some predefined conditions, including but not limited to at least one or several following example triggering conditions:
[0250] ● Triggering Condition 1: The RRC configuration related to applicable functionality / model reporting for an AI / ML based feature / feature group has been configured and / or there are no applicable functionalities / models for the AI / ML based feature / feature group has been sent before.
[0251] ● Triggering Condition 2: At least one AI / ML functionality / models for the AI / ML based feature / feature group has been activated;
[0252] ● Triggering Condition 3: The AI / ML based features / functionalities / models for the configured AI / ML based feature / feature group stored in the UE have changed since the latest UE applicable functionality reporting. In one example implementation of the change of AI / ML based features / functionality / models that are stored in the UE, it may include an addition of an AI / ML based features / functionality / models, a removal an AI / ML based features / functionality / models, or modification of an AI / ML based features / functionality / models.
[0253] ● Triggering Condition 4: The SIB where the AI / ML based feature group is indicated as supported or the associated applicable feature / functionality / model is indicated as requested by NW has been received and no UAI related to applicable AI / ML features / functionalities / models reporting has been sent before.
[0254] Further for option 1, in some example implementations, the UAI message as triggered and transmitted to the NW may contain various information about the applicable functionalities / models for the AI / ML based feature and / or may contain various information about the applicable AI / ML based features. In some example implementations, the applicable functionality / models for the AI / ML based feature or the applicable AI / ML based features for AI / ML based feature group may be a subset or full set of supported applicable features / functionality / models reported in the UE capability. In some example implementations, the applicable functionality / models for the AI / ML based feature or the applicable AI / ML based features for AI / ML based feature group may be additional to the supported applicable features / functionality / models reported in the UE capability.
[0255] In the case that UAI message is triggered by the AI / ML based feature / functionality / model modification or update, the UAI message may contain at least one of the following information regarding the NW additional conditions for a modified / updated AI / ML based feature / functionality / model:
[0256] ● AI / ML based feature indication: to indicate the AI / ML based feature the message or the NW additional conditions are for.
[0257] ● ISD to indicate cell range, for example, UMa, UMi, 200 m, 500 m.
[0258] ● Antenna height to indicate the antenna height of base station of the cell.
[0259] ● NLOS probability to indicate a probability of the NLOS radio propagation.
[0260] ● Indoor / outdoor indication to indicate whether the current scenario is indoor or outdoor, or an indoor / outdoor ratio.
[0261] ● DL Tx beam codebook indication to indicate a DL TX beam codebook that is used for AI / ML models training.
[0262] ● gNB antenna array dimensions indication.
[0263] ● A down tilt of the gNB antenna.
[0264] ● A beam pattern of set B indication to indicate the beam pattern of set B compare to the Set A (see above) . For example, set B may be a 1 / 4 subset of the set A, etc.
[0265] ● The beam pattern of the Set A for the AI / ML based functionality to indicate the beam pattern of set A compare to the full beam set. For example, set A may be a 1 / 4 subset of full set, etc.
[0266] In the case that UAI message is triggered by AI / ML based feature / functionality / model addition, the UAI message may contain at least one of the following information regarding the NW additional conditions for the newly added AI / ML based feature / functionality / model:
[0267] ● AI / ML based feature indication: to indicate the AI / ML based feature the message or the NW additional conditions are for.
[0268] ● ISD to indicate cell range, for example, UMa, UMi, 200 m, 500 m.
[0269] ● Antenna height to indicate the Antenna height of base station of the cell.
[0270] ● NLOS probability to indicate a probability of the NLOS radio propagation.
[0271] ● Indoor / outdoor indication to indicate whether the current scenario is indoor or outdoor, or an indoor / outdoor ratio.
[0272] ● DL Tx beam codebook indication to indicate a DL TX beam codebook that is used for AI / ML models training.
[0273] ● gNB antenna array dimensions indication.
[0274] ● A down tilt of the gNB antenna.
[0275] ● A beam pattern of set B indication to indicate the beam pattern of set B compare to the Set A (see above) . For example, set B may be a 1 / 4 subset of the set A, etc.
[0276] ● The beam pattern of the Set A for the AI / ML based functionality to indicate the beam pattern of set A compare to the full beam set. For example, set A may be a 1 / 4 subset of full set, etc.
[0277] For the Option 2 above (dedicated RRC messaging) , the example UEApplicableFunctionalityReporting message may be a response message to the RequestApplicableFunctionality above where the NW additional conditions for at least one AI / ML based feature / feature group is present and / or the applicable AI / ML based features / functionalities / models reporting for at least one AI / ML based feature / feature group is requested. The example UEApplicableFunctionalityReporting message, for example, may indicate / include the applicable AI / ML functionalities / models for the AI / ML based features or the applicable AI / ML based features for the AI / ML based feature group In some example implementations, the applicable functionalities / models for the AI / ML based features or the applicable AI / ML based features for AI / ML based feature groups are a subset or full set of supported applicable features / functionality / models reported in the UE capability. In some example implementations, the applicable functionalities / models for the AI / ML based feature or the applicable AI / ML based features for AI / ML based feature group is additional to the supported applicable features / functionalities / models reported in the UE capability.
[0278] In the case of UEApplicableFunctionalityReporting containing the AI / ML feature / functionality / model which is updated or added compare to the supported AI / ML feature / functionality / model in the (latest) reported UE capability, the UEApplicableFunctionalityReporting may contain at least one of the following information regarding the NW additional conditions for the modified / updated / added AI / ML feature / functionality / model:
[0279] ○ AI / ML based feature indication: to indicate the AI / ML based feature the message or the NW additional conditions are for.
[0280] ○ ISD to indicate cell range, for example, UMa, UMi, 200 m, 500 m.
[0281] ○ Antenna height to indicate the Antenna height of base station of the cell.
[0282] ○ NLOS probability to indicate a probability of the NLOS radio propagation.
[0283] ○ Indoor / outdoor indication to indicate whether the current scenario is indoor or outdoor, or an indoor / outdoor ratio.
[0284] ○ DL Tx beam codebook indication to indicate a DL TX beam codebook that is used for AI / ML models training.
[0285] ○ gNB antenna array dimensions indication.
[0286] ○ A down tilt of the gNB antenna.
[0287] ○ A beam pattern of set B indication to indicate the beam pattern of set B compare to the Set A (see above) . For example, set B may be a 1 / 4 subset of the set A in a manner of even distribution, etc.
[0288] ○ The beam pattern of the Set A for the AI / ML based functionality to indicate the beam pattern of set A compare to the full beam set (see above) . For example, set A may be a 1 / 4 subset of full set in a manner of even distribution, etc.
[0289] In some example implementations, for both Option 1 and Option 2 above, the applicable AI / ML functionalities / models for an AI / ML based feature / feature group may be reported / indicated via a BIT STRING type parameter (or bit map type parameter) , or the applicable AI / ML features may be reported / indicated via a BIT STRING type parameter (or bit map type parameter) . As an example syntax, a bit string type parameter indicating applicable AI / ML functionalities for spatial beam management features is shown below:
[0290] AI-ML-SpatialBeamManagement BIT STRING (SIZE
[0291] (1.. maxNrofFunctionalities / Models) )
[0292] In such an example bit string or bit map, the first bit from the left (or right) may map to the first entry of the functionalities list for AI / ML based feature / feature group (e.g., AI / ML based spatial beam management) reported by UE capability, the second bit from the left (or right) may map to the second entry of the functionality list for AI / ML based feature / FG (e.g., AI / ML based spatial beam management) reported by UE capability, and so on. The AI / ML functionality mapped to a particular bit may be considered as applicable if that particular bit is set to 1 in the report. Otherwise, the AI / ML functionality mapped to and associated with that particular bit may not be considered as applicable
[0293] As another example syntax for applicable AI / ML based feature, a bit string type parameter indicating applicable AI / ML features for an AI / ML feature group for beam management is shown below:
[0294] AI-ML-BeamManagement BIT STRING (SIZE (2) )
[0295] In such an example, the value AI-ML-BeamManagement = 01 indicates that the feature of AI / ML spatial beam management / AI / ML temporal beam management is applicable. The value AI-ML-BeamManagement = 10 indicates that the feature of AI / ML temporal beam management / AI / ML spatial beam management is applicable. The value AI-ML-BeamManagement = 00 indicates that neither AI / ML temporal beam management nor AI / ML spatial beam management is applicable. The value AI-ML-BeamManagement = 11 indicates that both AI / ML temporal beam management and AI / ML spatial beam management are applicable.
[0296] In some example implementations, for both Option 1 and Option 2 above, the applicable AI / ML functionalities / models for an AI / ML based feature / feature group may be reported / indicated via a list of AI / ML functionality index / identity or AI / ML model index / identity. As an example syntax, a integer type parameter list indicating applicable AI / ML based spatial beam management is shown below:
[0297] ApplicableFuntionalities SEQUENCE (SIZE (0.. maxNrofAIMLfunctionalities) OF FunctionalityId
[0298] Or
[0299] ApplicableModels SEQUENCE (SIZE (0.. maxNrofAIMLModels) OF ModelId
[0300] In such an example AI / ML functionality / model Index or identity, the functionality or model index may be numbered in the order of the entry in the AI / ML functionality list or in model list for one AI / ML based feature or functionality in UE capability. For example, the first entry in the list is numbered with Index=0, the second entry in the list is numbered with index=1, and so on.
[0301] In such an example AI / ML functionality / model Index or identity, the functionality or model index / identity may be aligned with the functionality or model index / identity reported in UE capability.
[0302] In some example implementations, for both Option 1 and Option 2 above, the applicable AI / ML based features may be reported / indicated via a list of AI / ML based features. As an example syntax, a list parameter indicating applicable AI / ML based beam management is shown below:
[0303] ApplicableFeaturesForBeamManagement SEQUENCE (SIZE (0. . maxNrofAIMLfeaturesForBM) OF FeatureId
[0304] In such an example AI / ML feature Index or identity, the feature index or identity may be numbered by the order of the entry in the AI / ML feature list for one AI / ML based feature group in UE capability, for example, the first entry in the list is numbered with Index=0, the second entry in the list is numbered with index=1, and so on. In such an example AI / ML feature Index or identity, the feature may be aligned with the AI / ML based feature index / identity for a AI / ML based feature group reported in UE capability.
[0305] For the Option 3 above, a UL MAC CE (e.g., an Applicable AI / ML Reporting MAC CE) for reporting the applicable UE AI / ML features / functionalities / models may be triggered. In some example implementations, such an applicable Functionality Reporting MAC CE may be triggered by any one or more of the following conditions similar to Option 1 and Option 2:
[0306] ● Triggering Condition 1: The RRC configuration related to applicable feature / functionality / model reporting for an AI / ML based feature group / feature / functionality has been configured and no applicable feature / functionality / models reporting has been sent before.
[0307] ● Triggering Condition 2: At least one AI / ML feature / functionality / model for an AI / ML based feature group / feature / functionality has been activated;
[0308] ● Triggering Condition 3: The AI / ML based feature / functionalities / models for the AI / ML based feature group / feature / functionality stored in the UE have been changed since the latest transmission of UE applicable feature / functionality / model reporting. In one example implementation of the change of AI / ML based features / functionality / models those are stored in the UE, it may include additions of AI / ML based features / functionality / models, removals of AI / ML based features / functionality / models, or modification of AI / ML based features / functionalities / models.
[0309] In some example implementations, the UL MAC CE above may trigger a Scheduling Request (SR) if there has not been any available UL grant received for transmitting such UL MAC CE. In some example implementations, only one SR configuration may be applied in one cell group for the applicable functionality reporting.
[0310] In some example implementation, the UL MAC CE above may include at least one of the following information or information items:
[0311] ● Serving cell indication: to indicate the serving cell where the AI / ML based feature (s) is / are applied.
[0312] ● AI / ML based feature group indication: to indicate AI / ML based feature group (s) of concern for applicable AI / ML feature / functionalities / models.
[0313] ● AI / ML based feature indication: to indicate AI / ML based feature (s) of concern for applicable AI / ML features / functionalities / models.
[0314] ● AI / ML functionality indication: to indicate AI / ML based functionality or functionalities of concern for applicable AI / ML models.
[0315] ● AI / ML applicable feature / functionality / model indication: to indicate the current applicable AI / ML features / functionalities / models at the UE for the indicated AI / ML based feature (s) .
[0316] In some example implementations, the UL MAC CE above may follow the example structure / format illustrated in FIG. 7. In FIG. 7, “Fi” in the first octet of the UL MAC CE indicates whether an octet containing Fu i, j related to the i-th AI / ML based feature is present (or applicable) . The octet containing Fu i, j is present if Fi =1, otherwise the octet containing Fu i, j is absent. As such, the number Fu octets n would be equal to a number of 1’s of Fi in the first octet.
[0317] Further in the example of FIG. 7, “Fui, j” octet indicate applicable functionalities at UE for AI / ML based feature corresponding “Fi” if indicated a being present. Each of the bit in “Fui, j” corresponds to and indicates whether the j-th AI / ML functionality of the AI / ML feature is applicable in the supported functionality list reported by the UE capability (e.g., UE capability reported in STEP 1 of FIG. 5) . The j-th AI / ML functionality in an AI / ML based feature is applicable if “Fui, j” is set to 1. Otherwise, the AI / ML functionality is not applicable.
[0318] Returning to the second alternative Solution in STEP 2 for determining and reporting UE available AI / ML features / functionalities / models (Solution 2) of FIG. 5 using the RRC reconfiguration mechanism, the RRC reconfiguration procedure may including transmitting an RRCReconfiguration message from the NW to the UE (Solution 2-1 of FIG. 5) for providing the conditions described above, and a response message RRCReconfigurationComplete sent by the UE to the NW (Solution 2-2 of FIG. 5) for reporting applicable AI / ML features / functionalities / models.
[0319] Specifically, the NW may request the UE to report the applicable AI / ML functionalities for an AI / ML based feature / feature group via an RRCReconfiguration message, and UE may correspondingly respond to the NW with the applicable AI / ML functionalities for an AI / ML based feature / feature group via an RRCReconfigurationComplete message.
[0320] In some example implementations for the RRCReconfiguration message for requesting applicable UE AI / ML functionalities / models for an AI / ML based feature, parameters, for example, an ENUMERATED type parameter per each AI / ML based feature / feature group may be introduced or included as exemplarily illustrated below:
[0321] RequestForAI-ML-SpatialBeamManagement ENUMERATED {TRUE}
[0322] OPTIONAL --Need N
[0323] RequestForAI-ML-TemporalBeamManagement ENUMERATED {TRUE}
[0324] OPTIONAL --Need N
[0325] RequestForAI-ML-CSI-RS-Compression ENUMERATED {TRUE}
[0326] OPTIONAL, --Need N
[0327] RequestForAI-ML-CSI-RS-Prediction ENUMERATED {TRUE}
[0328] OPTIONAL, --Need N
[0329] In some example implementations for the RRCReconfiguration message for requesting applicable UE AI / ML based features / functionalities / models for an AIML base feature group, parameters, for example, an ENUMERATED type parameter per each AI / ML based feature / feature group may be introduced or included, as exemplarily illustrated below:
[0330] RequestForAI-ML-BeamManagement ENUMERATED {TRUE}
[0331] OPTIONAL --Need N
[0332] RequestForAI-ML-CSI ENUMERATED {TRUE}
[0333] OPTIONAL, --Need N
[0334] In one example implementation of above parameters, these parameters may be provided per UE, per cell group, per band, or per cell.
[0335] In an example implementation of the RRCReconfiguration message for the request of applicable feature / functionality / model, one or more information elements for the request for reporting applicable AI / ML feature / functionality / model may be introduced, which may include at least one of the following information items about NW additional conditions for consideration by the UE for reporting AI / ML feature / functionality / model:
[0336] ● AI / ML based feature indication: to indicate the AI / ML based feature the message or the NW additional conditions are for.
[0337] ● ISD to indicate cell range, for example, UMa, UMi, 200 m, 500 m.
[0338] ● Antenna height to indicate the Antenna height of base station of the cell.
[0339] ● NLOS probability: to indicate a probability of the NLOS radio propagation.
[0340] ● Indoor / outdoor indication to indicate whether the current scenario is indoor or outdoor, or an indoor / outdoor ratio.
[0341] ● DL Tx beam codebook indication: to indicate a DL TX beam codebook that is used for AI / ML models training.
[0342] ● gNB antenna array dimensions indication.
[0343] ● A down tilt of the gNB antenna.
[0344] ● A beam pattern of set B indication to indicate the beam pattern of set B compare to the Set A (see above) . For example, set B may be a 1 / 4 subset of the set A, etc.
[0345] ● The beam pattern of the Set A for the AI / ML based functionality to indicate the beam pattern of set A compare to the full beam set (see above) . For example, set A may be a 1 / 4 subset of full set, etc.
[0346] In some example implementations of the parameter (s) contained in RRCReconfigurationComplete by the UE in response to the request in the RRCReconfiguration above, the applicable AI / ML based feature / functionality / model for an AI / ML based feature group / feature may be indicated via an example BIT STRING or bitmap type parameter. In such an example bit string or bit map, the first bit from the left (or right) may map to the first entry of the feature / functionality / model list for AI / ML based feature group / feature (e.g., AI / ML based beam management or AI / ML based spatial beam management) reported by UE capability, the second bit from the left (or right) may map to the second entry of the feature / functionality / model list for AI / ML based FG / feature (e.g., AI / ML based beam management or AI / ML based spatial beam management) reported by UE capability, and so on. The AI / ML feature / functionality / model mapped to a particular bit may be considered as applicable if that particular bit is set to 1 in the report. Otherwise, the AI / ML feature / functionality / model mapped to and associated with that particular bit may not be considered as applicable. An example bit string is shown blow:
[0347] AI-ML-SpatialBeamManagement BIT STRING (SIZE (1. . maxNrofFunctionalities) )
[0348] OPTIONAL
[0349] In one example implementation of above parameters of applicable AI / ML feature / functionality / model, these parameters may be provided per UE, per cell group, per band, or per cell.
[0350] UE-Side AI / ML LCM –Testing of Applicable UE AI / ML Features / Functionalities / Models
[0351] Turning to STEP 3 of FIG. 5 with respect to preparation for AI / ML feature / functionality / model activation which may be after the UE report of applicable AI / ML functionalities in STEP 2, a testing for the applicable AI / ML features / functionalities / models may be performed their adoption and activation is further considered. For example, the conditions provided by the NW to the UE in STEP 2 above may be limited / restricted in order for the NW to avoid as much as possible an exposure of sensitive NW additional conditions (e.g., the conditions that may risk being analyzed to expose sensitive user information should not be disseminated) , the determination by the UE of the available AI / ML functionalities as reported to the NW in STEP 2 above may be performed with limited / restricted conditions provided by the NW. As such, it may be desirable in STEP 3 of FIG. 5 for the UE and the NW to test of applicable AI / ML functionalities as reported by the UE to determine their predication performance / accuracy in order to further determine whether they can be adopted / activated in predicting, e.g., settings of gNB, the deployment scenario, etc.
[0352] An example for such testing procedure by the UE 802 and the NW 804 for UE side AI / ML functionalities is illustrated in FIG. 8, which includes the following example steps:
[0353] ● STEP 3-1: The NW 804 may send a Message A for testing one or more applicable AI / ML features / functionalities / models.
[0354] ● STEP 3-2: The UE 802 may input a reference data list received from Message A into the one or more applicable AI / ML features / functionalities / models to obtain the corresponding inference outputs or values.
[0355] ● STEP 3-3: The UE 802 may send the inference outputs or values via a Msg B to the NW 804 for the NW to compare the inference outputs or values to ground truths known by the NW 804.
[0356] In some example implementations of STEP 3-1 above, Message A may be implemented as a DL RRC message, DL MAC CE or PDCCH signaling. In another example implementation, Message A may be a protocol signaling terminated between the AI / ML logical layer and NW logical entity / unit.
[0357] Further in some example implementations of STEP 3-1 above, the Message A may include / indicate at least one of the following information or information items:
[0358] ● An applicable AI / ML feature / functionality / model indication: to indicate the applicable AI / ML features / functionalities / models that need to be tested. In some example implementation, the applicable AI / ML features / functionalities / models may be from the applicable AI / ML reporting in above STEP 2 of FIG. 5.
[0359] ● A supported AI / ML feature / functionality / model indication: to indicate the supported AI / ML features / functionalities / models that need to be tested. In some example implementation, the supported AI / ML features / functionalities / models may be from the (latest) reported UE capability in above STEP 1 of FIG. 5.
[0360] ● Reference data list (Input data list for AI / ML functionalities to be tested) : a list of data, which is to be used as input data for the AI / ML features / functionalities / models those need to be tested.
[0361] ● Reference data list (the benchmark value list) : a list of data, which is to be used as benchmark data for evaluating the output data from the AI / ML features / functionalities / models those need to be tested.
[0362] In some example implementations of STEP 3-2 above, the UE may process the input reference data list and / or benchmark reference data list received from Message A using corresponding one or more applicable AI / ML features / functionalities / models to obtain the corresponding inference outputs values that may be used by the NW to check the performance and validity of these AI / ML features / functionalities / models or performance metrics that may be used by the UE to check the performance and validity of these AI / ML features / functionalities / models.
[0363] In some example implementations of STEP 3-3 above, the protocol format of Message B may be implemented as a UL RRC message (e.g. UAI) , UL MAC CE, or via a PUCCH signaling.
[0364] Further in some example implementations of STEP 3-3 above, the Message B may contain / indicate at least one of the following information or information items:
[0365] ● An applicable AI / ML feature / functionality / model indication: to indicate the applicable AI / ML features / functionalities / models that have been tested.
[0366] ● The output data list: to include the output data obtained from the inference of each applicable AI / ML feature / functionality / model.
[0367] ● UE additional conditions: to include UE additional conditions to assist the NW to determine validity or performance of the tested AI / ML features / functionalities / models.
[0368] ● An available AI / ML feature / functionality / model indication: to indicate the valid AI / ML features / functionalities / models that have been tested.
[0369] ● KPI value indication: to indicate the KPI value to each AI / ML features / functionalities / models. In some example implementations, it is used to indicate the KPI value to each AI / ML feature / functionality / model that are considered as available.
[0370] In some example implementations of another alternative of testing of Applicable UE AI / ML Features / Functionalities / Models, the following steps may be performed:
[0371] ● STEP 1: the NW may configure an RRC signaling to the UE for applicable UE side AI / ML Features / Functionalities / Models.
[0372] ● STEP 2: The UE may send one message for reporting the available AI / ML Features / Functionalities / Models to the NW.
[0373] In some example implementation of the RRC signaling for STEP 1, the RRC signaling may contain an RRC configuration which may be one or more pieces of CSI-ResourceConfig for one or more applicable AI / ML Features / Functionalities / Models reported in Step 2 (e.g., applicable AI / ML functionality / models reporting) . the RRC configuration may be one or more pieces of CSI-ResourceConfig for one or more supported AI / ML Features / Functionalities / Models reported in Step 1 (e.g., UE capability) . In some example implementation, the CSI-ResourceConfig may be a configuration of the reference signal which is used for UE to perform measurement to obtain the input of the AI / ML functionality / models and / or to obtain the benchmark value (e.g., measurement result) to be used for calculating the performance metric by comparing the actual measurement result and inferred value (e.g., output of the tested AI / ML functionality / models) .
[0374] In some example implementation of the RRC signaling for STEP 1, the RRC signaling may contain an RRC configuration which may be one or more pieces RRC configuration to indicate the AI / ML features / functionalities / models those need to be tested. In some example implementation of, the AI / ML feature / functionality / model needing testing may be one of the applicable AI / ML features / functionalities / models reported by UE as in STEP 2 of FIG. 5. In some example implementation d, the AI / ML feature / functionality / model needing testing may be one of the supported AI / ML features / functionalities / models in UE capability as in STEP 1 of FIG. 5.
[0375] In some example implementations, the message above may be an UL RRC signaling, an UL MAC CE, or a UL RRC message / UL MAC CE. At least one of the following information may be include / indicated:
[0376] ● An available AI / ML based feature indication.
[0377] ● An available AI / ML based functionality indication.
[0378] ● An available AI / ML based model indication.
[0379] ● A Performance KPI value for each available AI / ML based feature / functionality / model.
[0380] UE-Side AI / ML LCM –UE AI / ML Configuration
[0381] Turning now to STEP 4 of FIG. 5, the NW may now determine configurations for AI / ML functionality based on the available UE AI / ML functionalities, the testing output and or UE additional conditions received from the UE in STEP 3, and further communicate such configuration to the UE. For example, the NW may determine which of the applicable AI / ML models being tested provide acceptable prediction accuracy or performance and configurations thereof.
[0382] In one example implementation of the STEP 4 above, the configuration as determined by the NW for the UE AI / ML functionalities may be transmitted to the UE via several alternative RRC configuration structures.
[0383] In a first optional or alternative example, an RRC configuration structure for the UE AI / ML related configuration as shown in FIGS. 9A-9E may be constructed and transmitted from the NW to the UE. As shown in the example of FIGS. 9A-9E:
[0384] ● 1: The AL / ML (AIML) related configuration ( “ALMLConfig” ) may be configured in the serving cell configuration ( “ServingCellConfig” ) , as shown in FIG. 9A In another example implementation, the AI / ML (AIML) related configuration ( “AIMLConfig” ) may be configured in the CellGroupConfig, as shown in FIG. 9B or PhysicalCellGroupConfig, as shown in FIG. 9C, or MAC-CellGroupConfig, as shown in FIG. 9D. In another implementation, the AI / ML (AIML) related configuration ( “AIMLConfig” ) may be configured per UE, that is, the information element AIMLConfig has a same level with IE CellGroupConfig, as shown in FIG. 9E.
[0385] ● 2: The AIML related configuration (e.g. AIMLConfig) may contain a list of configurations of AI / ML based feature (e.g., “AIMLBasedFeatureConfig#1” through “AIMLBasedFeatureConfig#4” ) .
[0386] ● 3: Each AI / ML based feature configuration may contain a list of configurations of AI / ML functionalities (e.g., “AIMLFunctionalityAddModList” including “AIMLFunctionalityConfig#1” through “AIMLFunctionalityConfig#3” ) and / or an index / identity of the AI / ML based feature configuration, and / or an indication of the AI / ML based feature for such configuration.
[0387] ● 4: An AIML functionality configuration may contain at least one of the following information items:
[0388] ○ An index / identity of the AI / ML functionality or AI / ML functionality Id (e.g., “AIMLFunctionalityID” ) .
[0389] ○ An AI / ML model list of the AI / ML functionality (e.g. “AIMLModelToAddModList” ) . In each AI / ML model in the list, one model Id may be included to represent the AI / ML model, and / or the quantification value of computation power consumption when the corresponding AI / ML model is activated.
[0390] ○ Inference related configuration (list) (e.g., “InferenceConfig” or “InferenceConfigToAddModList” ) .
[0391] ○ Performance Monitoring related configuration (list) (e.g., “MonitoringConfig” or “MonitoringToAddModList” ) .
[0392] ○ The legacy configuration for enabling or disabling corresponding non-AI based feature.
[0393] ○ The quantification value of computation power consumption when the corresponding AI / ML feature / functionality / model is activated.
[0394] Further in this example RRC configuration structure of FIGS. 9A-9E, all AI / ML functionalities / features / models can be configured together so as to facilitate future AI development, where adding new AI / ML functions / features / models and their configurations may be convenient achieved. In this example, the AI / ML related functions are configured per cell, or per cell group, or per UE.
[0395] In some example implementations, the AI / ML functionalities may be indexed in order to align the UE and NW’s understanding for correctly identifying the AI / ML functionalities / models. For example:
[0396] ● In some example implementations, the AI / ML functionality indexes / identities may be constructed in accordance with locations of AI / ML functionalities for an AI / ML based feature in the supported AI / ML functionality list in UE capability (as reported in STEP 1 of FIG. 5) . For example, the AI / ML functionality index=0 for an AI / ML based feature in RRC configuration may be to represent the first entry of the supported AI / ML functionality list for the AI / ML based feature in the UE capability, the AI / ML functionality index=1 for an AI / ML based feature in RRC configuration may be to indicate the second entry of the supported AI / ML functionality list for the AI / ML based feature in the UE capability, and so on.
[0397] ● In some example implementations, an AI / ML functionality represented by an AI / ML functionality index / identity in the RRC configuration above may be constructed with the same AI / ML functionality index / identity in UE capability (as reported in STEP 1 of FIG. 5) .
[0398] ● In some example implementations, the AI / ML functionality index / identity for an AI / ML functionality in RRC Configuration may be constructed in accordance with the location of the AI / ML functionality in applicable functionality list in applicable functionality reporting (as reported in STEP 2 of FIG. 5) .
[0399] ● In some other implementations, an AI / ML functionality represented by an AI / ML functionality index / identity in the RRC configuration may be the same as AI / ML functionality index / identity in applicable functionality reporting (as reported in STEP 2 of FIG. 5) . For example, the AI / ML functionality index=0 for an AI / ML based feature in RRC configuration may be to represent the first entry of the supported AI / ML functionality list for the AI / ML based feature in the applicable AI / ML functionality reporting, and the AI / ML functionality index=1 for an AI / ML based feature in RRC configuration may be to indicate the second entry of the supported AI / ML functionality list for the AI / ML based feature in the AI / ML functionality reporting, and so on.
[0400] In some example implementations, the InferenceConfigurations of FIGS. 9A-9E above may include at least one of the following information or information items:
[0401] ● Measurement Configuration (list) for obtaining input data to Inference, for example, MeasurementForInferenceConfig or MeasurementForInferenceConfigToAddmodList.
[0402] ● Output data, inference result, reporting configuration (list) (e.g: inferenceReporting or inferenceReportingToAddModList) . For example, in one implementation, the InferenceReporting may contain an indication about the contents to be reported. For example, In one implementation of information of reporting contents, for beam management, it may be at least one of the following information about the contents:
[0403] ○ The top N DL beams according to the inference result, N>0;
[0404] ○ The top N DL beams according to the inference result with corresponding RSRP values, N>0;
[0405] ○ The top N DL beams according to the inference result with the RSRP values of the best M DL beams among the N DL beams, N>M>=0;
[0406] ● In one implementation, the InferenceReporting may contain an indication about the report type:
[0407] ○ Semi-persistent periodic
[0408] ○ Periodic
[0409] ○ Event triggered
[0410] ■ The TOP 1 beam have been changed according to the inference result.
[0411] ■ The current DL TX beam have been out of the top N beams according to the inference result
[0412] ■ The current DL TX beam with a RSRP value lower than a threshold value according to the inference result.
[0413] ■ The current DL TX beam with a RSRP value lower than a threshold value and at least one beam with a RSRP value higher than the threshold value according to the inference result
[0414] In some example implementations of performance monitoring related configuration, it may contain at least one of the following information:
[0415] ● Measurement Configuration: The measurement configuration for performance monitoring
[0416] ● Measurement Reporting Configuration: The reporting configuration for performance metrics or performance result reporting.
[0417] ● In one implementation of legacy related configuration, it may contain at least one of the following information:
[0418] ● CSI-ReportingConfigId: To indicate the CSI reporting configuration for Non-AI based beam management. In one implementation, the report type configured in CSI-ReportingConfig indicated by such CSI-ReportingConfigId is Semi-Persistent on PUSCH or Semi-Persistent on PUCCH, or Aperioidc.
[0419] In an implementation of option 1, assuming that the AI / ML model (list) were present, each entry of the InferenceToAddModList is associated with each entry of the AI / ML model (list) in a manner of ascending order, and / or each entry of the MonitoringToAddModList is associated with each entry of the AI / ML model (list) in a manner of ascending order.
[0420] In a second optional or alternative example for RRC configuration structure for the AI / ML feature / functionality / model, an AI / ML based feature and / or AI / ML based functionality / model may be associated with CSI-ReportConfig. An example CSI-ReportConfig structure including the AI / ML based feature / functionality / model configurations is shown below:
[0421] With above example RRC structure, the AI / ML functionality / feature / model can be implicitly activated / deactivated by the activation / deactivation of semi-persistent CSI reporting. In some example implementations for configuring the AI / ML feature / functionality / model to the UE, the AI / ML related information may be configured in an RRC configuration for CSI reporting (e.g., CSI-ReportConfig) , as shown above.
[0422] In some example implementations, the RRC configuration for AI / ML feature / functionality / model may include and / or indicate at least one of the following information or information items:
[0423] ● An AI / ML feature / functionality / model indication: to indicate the AI / ML feature / functionality / model that the CSI reporting configuration corresponds to.
[0424] ● A measurement resource configuration for inference indication: to indicate the measurement resources configuration for AI / ML functionality / model inference input which is related to this CSI report configuration. The indication may be via a CSI-ResouceConfigId.
[0425] ● A measurement resource configuration for monitoring indication: to indicate the measurement resources configuration for AI / ML functionality / model performance monitoring which is related to this CSI report configuration, the indication may be a CSI-ResouceConfigId. In one example implementation, the measurement resource configuration for monitoring indication may be exclusive with the measurement resource configuration for inference indication.
[0426] ● An AI / ML reporting indication: to indicate the report information for the related AI / ML feature / functionality / model. In one example implementation, it may be included in the information element ReportQuantity in the CSI report configuration.
[0427] ● An AI / ML reporting type indication: to indicate the report type for the related AI / ML feature / functionality / model. In one example implementation, to introduce a event triggered type reporting, the detail would be described in below.
[0428] In some example implementations, for the CSI reporting configuration for configurating an AI / ML based feature / functionality / model, the CSI reporting type may be configured as semi-persistent on PUCCH or semi-persistent on PUSCH, or event triggered CSI reporting if the CSI reporting configuration is related to the measurement configuration for inference. The CSI reporting type may be configured as semi-persistent CSI reporting on PUCCH or semi-persistent CSI reporting on PUSCH, or as aperiodic CSI reporting, or as event triggered CSI reporting if it is related to the measurement configuration for monitoring.
[0429] Turning now to STEP 5 of FIG. 5, the UE may perform, for example, activation and / or deactivation of one or more AI / ML feature / functionality / model may be instructed from the NW to the UE.
[0430] In one example implementation for STEP 5, when an example RRC configuration structure of FIGS. 9A-9E is used, the activation of an AI / ML functionality may be effectuated via a separate signaling from the NW to the UE. Such a signaling may contain at least one of the following information or information items:
[0431] ● Serving cell indication: to indicate the serving cell the AI / ML feature / functionality / model is activated / deactivated for.
[0432] ● AI / ML based feature indication: to indicate the AI / ML based feature the AI / ML functionality belongs to.
[0433] ● AI / ML functionality indication: to indicate the AI / ML functionality which need to be activated / deactivated.
[0434] ● AI / ML functionality indication: to indicate the AI / ML model which need to be activated / deactivated.
[0435] ● The activated / deactivated indication: to indicate the activation / deactivation for the indicated AI / ML feature / functionality / model.
[0436] Such a signaling, for example, may be implemented as a DL MAC CE. An example DL MAC CE for signaling activation / deactivation of AI / ML functionalities is shown in FIG. 10.
[0437] In the example DL MAC CE of FIG. 10, “Ci” in the first octet indicates the serving cell with a servingCellId=i where the AI / ML functionality shall be activated / deactivated with an octet being present (when Ci=1) . Ci can be 0 or 1. “Fi” in the second octet indicates how many octets would be present following the second octet. For example, if Ci =1, Fi=1 would mean that there would be two octets being present for the serving cell represent by Ci (e.g., for AI / ML based spatial beam management, AI / ML based temporal beam management) , while Ci=1, Fi=0 would mean that there would be 1 octet being present for the serving cell present by Ci. “Feat IDi, j” indicates the AI / ML based feature ID to which the j-th functionality to be activated / deactivated belongs for the serving cell with a servingCellId=i. “Functionality IDi, j” indicates the AI / ML functionality that is indicated to be activated / deactivated for the serving cell with a servingCellId=i. The R bits indicated in FIG. 10 represent reserved bits.
[0438] In another example implementation for STEP 5 of FIG. 5, when the example CSI-ReportConfig RRC structure is used, the signaling of the activation / deactivation of AI / ML features / functionalities / models can be done via the related semi-persistent CSI reporting activation / deactivation whose configuration may be indicated as inference reporting (e.g. ReportQuantity is set to Inference-SSB-Index-RSRP) . For example, a DL MAC CE may be used with at least one of the following information or information items:
[0439] ● 1: Serving cell Id: to indicate the serving cell where CSI reporting is reported.
[0440] ● 2: BWP Id: to indicate the UL BWP where the CSI reporting is reported.
[0441] ● 3: CSI-report configuration indication: to indicate the activation / deactivation of the CSI reporting configuration for AI / ML feature / functionality / model.
[0442] UE-Side AI / ML LCM –UE AI / ML Activation / Deactivation
[0443] Turning to STEP 6 of FIG. 5, the UE may perform AI / ML feature / functionality / model activation and / or deactivation based on the instruction / signaling from the NW in STEP 5 above. In some example implementations, the UE may perform one or more of the following AI / ML feature / functionality / model activation operations upon receiving an DL MAC CE signaling, or in another example implementation, a protocol signaling terminated between the AI / ML logical layer at UE and AI / ML logical entity / unit at NW of STEP 5:
[0444] ● Indicating to lower layers the information regarding the AI / ML functionality activation / deactivation MAC CE.
[0445] ● Indicating to lower layers to deactivate the CSI-ReportConfig for legacy beam management related to the activated AI / ML Functionality.
[0446] In some example implementations, the UE may perform one or more the following AI / ML functionality deactivation operations:
[0447] ● Indicating to lower layers the information regarding the AI / ML functionality activation / deactivation MAC CE.
[0448] ● Indicate to lower layers to activate the CSI-ReportConfig for legacy beam management related to the deactivated AI / ML Functionality
[0449] UE-Side AI / ML LCM –UE AI / ML Inference Reporting
[0450] Turning now to STEP 7 of FIG. 5, inference and / or measurement reporting may be provided from the UE to the NW after the UE performs the activation and / or deactivation of AI / ML functionalities and or actual measurements. Such inference / measurement reporting may contain the predictions or inference by the UE based on activated AI / ML functionalities and models and / or actual measurements, which the NW may use for the configuration and provisioning of the network.
[0451] In some example implementations of STEP 7, the inference / measurement reporting may be implemented as a PUCCH signaling or PUSCH signaling when the associated AI / ML functionality / model is activated.
[0452] For example, for AI / ML based spatial beam management, the inference / measurement reporting may include at least one of the following information or information items:
[0453] ● 1: TOP N beam Ids with the corresponding N values for Reference Signal Received Power (RSRP) , where N>=1;
[0454] ● 2: TOP N beam Ids with RSRP values of M beams among the N beams, where N>M>=0;
[0455] ● 3: TOP N beam Ids whose RSRP values are greater than a pre-defined RSRP threshold value.
[0456] ● 4: A flag to indicate the RSRP value for beam. If the flag is present, its value may indicate whether the RSRP value of a corresponding beam with a beam Id is inferred via an AI / ML functionality / model or is actually measured.
[0457] Further, such example inference / measurement reporting may be triggered by at least one of the following events:
[0458] ● When the UE determines that the current DL TX beam is not among the TOP N beams according to the inference result so that there is a need to switch beam.
[0459] ● When the UE determines that an RSRP value of the current DL TX beam is lower than a predefined threshold RSRP value, according to the inference result, and that there is at least one DL TX beam with an RSRP value better than a pre-defined threshold RSRP value. The two pre-defined threshold value may be the same or different.
[0460] For another example, for AI / ML based temporal beam management, the inference / measurement reporting above in STEP 7 may be reported as the measurement result or inference result in each configured report occasion. Such reporting may include and / or indicate one or more of the following information or information items:
[0461] ● N beams Ids with the best N RSRP values in each inference report, where N>=1.
[0462] ● N beams Ids with the best M RSRP values in each inference report, where N>M>=1.
[0463] ● An indication to indicate whether the reporting contains the inference result or actual measurement result.
[0464] NW-Side AI / ML LCM –General
[0465] In some example implementations, AI / ML functionalities may reside on the NW side rather than the UE side. For AI / ML LCM of NW-side AI / ML functionality, while the NW may be able to control the determination of selection, activation, and deactivation of the AI / ML functionalities and models, it may still need UE assistance in providing some parameters and information in order for the NW to apply suitable AI / ML functionalities. Such UE assistance information may be provided to the NW as one or more UE Additional Conditions Report. A general LCM procedure performed by the UE 1102 and the NW 1104 for NW-side AI / ML functionalities are illustrated in FIG. 11, which includes the following example steps:
[0466] ● STEP 1: The NW 1104 may send to the UE 1102 a request message for UE additional condition reporting.
[0467] ● STEP 2: The UE 1102 may determine whether to trigger the UE additional conditions reporting based on one or more criteria. If the one or more criteria are met, the UE 1102 proceeds to STEP 3 below. Otherwise, the procedure ends.
[0468] ● STEP 3: The UE 1102 may then send the UE additional condition reporting to the NW 1104.
[0469] In some example implementations, the request message of STEP 1 above (e.g., RequestUEAdditionalConditions) of the UE additional condition reporting may be a DL RRC message (e.g., RRCReconfiguration) , in another example implementation, a protocol signaling terminated between the AI / ML logical layer and NW logical entity / unit which may include and / or indicate at least one of the following information or information items:
[0470] ● UE speed. For example, the UE speed may be provided as an enumerated type parameter with group values {15, 30, 60, 90, 120, etc. } . For another example, the UE speed may be provided as an enumerated type parameter with group values {static, low, mediate, high, etc. } , each of which may represent a speed range. For example, the value “static” may represent a speed range of [0, 5 km / h) , the value “low” may represent a speed range of [5 km / h, 15 km / h) , the value “mediate” may represent a speed range of [15 km / h, 60 km / h) , and the value “high” may represent a speed range of [60 km / h, 120 km / h) .
[0471] ● UE rotation, which, for example, may be an enumerated type parameter with a value among {30, 60, 90, 120, 180} .
[0472] IN some example implementations, a prohibit timer (with a predefined initial timer value) may be introduced for to controlling the reporting of the UE additional conditions. For example, such reporting may be prohibited and thus cannot be triggered if the prohibit timer is still running.
[0473] In some example implementations of the STEP 2 above, at least one of the following conditions may need to be met in order to trigger the UE additional condition reporting:
[0474] ● The prohibit Timer above for UE additional condition reporting is not running, if set or configured.
[0475] ● The UE additional conditions have changed since the latest UE additional condition reporting. For example:
[0476] ○ The UE speed has changed, for example, from static to low, from low to high, etc., since the last reporting.
[0477] ○ The UE speed has changed to be out of a configured speed range since the latest reporting. For example, the UE speed has reached to a value more than 15 km / h, or the UE speed has exceeded 30 km, etc.
[0478] ○ The UE rotation reaches a greater value than a threshold value. For example, the UE rotation exceeds 30 degrees since the latest report, or exceeds more than 60 degrees since the latest report, etc.
[0479] In some example implementations, a very first UE additional condition reporting to the NW may be conditioned on that no UE additional condition reporting has occurred since a last RRCReconfiguration message containing the RequestUEAdditionalConditions above.
[0480] In some example implementations of the STEP 3 above, the reporting may be included in an RRCReconfigurationComplete message, or in a protocol signaling terminated between the AI / ML logical layer and NW logical entity / unit which may contain at least one of the following information or information items regarding the UE additional conditions:
[0481] ● 1: UE speed information: to indicate the UE current speed information when the RRCReconfigurationComplete message is generated.
[0482] ● 2: UE codebook information: to indicate the UE’s codebook information regarding the RX beams.
[0483] ● 3: UE antenna array distribution information: to indicate the UE’s antenna array distributions.
[0484] In some other example implementations of STEP 3 above, the reporting may be included in a UEAssistanceInformation message or in a protocol signaling terminated between the AI / ML logical layer and NW logical entity / unit, which may likewise contain at least one of the following information or information items regarding the UE additional conditions:
[0485] ● 1: UE speed information: to indicate the UE current speed information when the UE additional reporting is generated.
[0486] ● 2: UE rotation information: to indicate the UE current rotation information compared to the UE’s orientation / directional information when receiving RequestUEAdditionalConditions.
[0487] The description and accompanying drawings above provide specific example embodiments and implementations. The described subject matter may, however, be embodied in a variety of different forms and, therefore, covered or claimed subject matter is intended to be construed as not being limited to any example embodiments set forth herein. A reasonably broad scope for claimed or covered subject matter is intended. Among other things, for example, subject matter may be embodied as methods, devices, components, systems, or non-transitory computer-readable media for storing computer codes. Accordingly, embodiments may, for example, take the form of hardware, software, firmware, storage media or any combination thereof. For example, the method embodiments described above may be implemented by components, devices, or systems including memory and processors by executing computer codes stored in the memory.
[0488] Throughout the specification and claims, terms may have nuanced meanings suggested or implied in context beyond an explicitly stated meaning. Likewise, the phrase “in one embodiment / implementation” as used herein does not necessarily refer to the same embodiment and the phrase “in another embodiment / implementation” as used herein does not necessarily refer to a different embodiment. It is intended, for example, that claimed subject matter includes combinations of example embodiments in whole or in part.
[0489] Reference throughout this specification to features, advantages, or similar language does not imply that all of the features and advantages that may be realized with the present solution should be or are included in any single implementation thereof. Rather, language referring to the features and advantages is understood to mean that a specific feature, advantage, or characteristic described in connection with an embodiment is included in at least one embodiment of the present solution. Thus, discussions of the features and advantages, and similar language, throughout the specification may, but do not necessarily, refer to the same embodiment.
[0490] Furthermore, the described features, advantages and characteristics of the present solution may be combined in any suitable manner in one or more embodiments. One of ordinary skill in the relevant art will recognize, in light of the description herein, that the present solution can be practiced without one or more of the specific features or advantages of a particular embodiment. In other instances, additional features and advantages may be recognized in certain embodiments that may not be present in all embodiments of the present solution.
Claims
1.A method performed by a user equipment (UE) in communication with a network (NW) in a wireless communication network, comprising:performing a testing procedure for Artificial Intelligence or Machine Learning (AIML) capabilities of the UE in collaboration with the NW;receiving from the NW an AIML configuration for one or more AIML features, functionalities, or models selected by the NW from AIML capabilities of the UE based at least partially on the testing procedure; andutilizing the one or more AIML features, functionalities, or models for inference according to the AIML configuration.2.The method of claim 1, wherein the testing procedure comprises:receiving a request message from the NW for testing the AIML features, functionalities, or models of the UE;performing inference using at least one AIML feature, functionality, or model to generate an inference result; andtransmitting a response message to the NW, the response message comprising at least the inference result.3.The method of claim 2, wherein the request message comprises at least one of:an indicator for indicating AIML features, functionalities, or models for testing; andreference input data to the indicated applicable AIML features, functions, or models for testing.4.The method of claim 3, wherein performing the inference comprises processing the reference input data using the indicated AIML features, functionalities, or models to generate the inference result.5.The method of claim 4, wherein the response message comprises at least one of:the indicator for indicating the applicable AIML features, functionalities, or models for testing; oran output data list comprising the inference result.6.The method of claim 5, wherein the response message further comprises a set of UE additional conditions, the UE additional conditions and the inference result being used by the NW to evaluate the one or more AIML features, functionalities, or models.7.The method of claim 1, wherein the AIML configuration comprises an RRC configuration structure.8.The method of claim 7, wherein the AIML configuration is included in a serving cell configuration structure.9.The method of claim 8, wherein the AIML configuration comprises a list of configurations for the one or more AIML features, functionalities, or models.10.The method of claim 9, wherein the AIML configuration comprises a hierarchical listing of configurations for the one or more AIML features, functionalities, or models.11.The method of claim 9, wherein each of the list of configurations for an AIML functionality comprises at least one of:an index or ID of the AIML functionality;an inference related configuration for the AIML functionality;a performance monitoring related configuration for the AIML functionality; ora legacy configuration for a non-AI functionality corresponding to the AIML functionality.12.The method of claim 11, wherein the inference related configuration comprises at least one of:a measurement configuration for obtaining input data for inference for the AIML functionality;an inference reporting configuration for performing inference of output data associated with the AIML functionality;a performance monitoring configuration for inference performance monitoring of the AIML functionality; ora performance reporting configuration for reporting the inference performance monitoring of the AIML functionality.13.The method of claim 7, wherein the AIML configuration is included in a Channel State Information (CSI) report configuraiton.14.The method of claim 13, wherein the CSI report configuration comprises a list of identifiers for the one or more AIML features, functionalities, or models.15.The method of claim 13, wherein the CSI report configuration comprises a report quantity configuration, the report quantity configuration comprising both a set of inference configurations for the one or more AIML features, functionalities, or models.16.The method of claim 13, wherein the CSI report configuration comprises a timing configuration for AIML inference report.17.The method of claim 16, wherein the timing configuration comprises at list one of:periodic reporting;semi persistence reporting on Physical Uplink Shared Channel (PUSCH) ;semi persistent reporting on PUCCH; oraperiodic.18.The method of claim 13, wherein the CSI report configuration comprises a measurement resource configuration for inference for the one or more AIML features, functionalities, or models.19.The method of claim 13, wherein the CSI report configuration comprises a measurement resource configuration for monitoring inference performance of the one or more AIML features, functionalities, or models.20.The method of claim 13, wherein utilizing the one or more AIML features, functionalities, or models comprises activating or deactivating the one or more AIML features, functionalities, or models.21.The method of claim 18, wherein activating or deactivating the one or more AIML features, functionalities, or models is implicitly triggered by an activation or deactivation of a semi-persistent CSI reporting.22.The method of claim 1, wherein utilizing the one or more AIML features, functionalities, or models comprises activating or deactivating at least one of the one or more AIML features, functionalities, or models.23.The method of claim 22, wherein activating or deactivating at least one of the one or more AIML features is triggered by the UE receiving a signaling from the NW.24.The method of claim 23, wherein the signaling comprises a DL MAC CE structure.25.The method of claim 24, wherein the DL MAC CE structure comprises:a first octet with plurality of bits indicating whether to activate or deactivate associated AIML functionalities associated with corresponding serving cell;a second octet having a plurality of bits indicating a number of subsequent octets in the DL MAC CE structure for the corresponding serving cells; andthe subsequent octets indicating AIML features and functionalities to be Activated or deactivated for corresponding serving cells.26.The method of claim 24, wherein the method further comprises indicating to lower layers information regarding AI / ML activation or deactivation.27.The method of claim 23, wherein the signaling comprises a semi-persistent CSI reporting activation / deactivation signaling.28.The method of claim 27, wherein the method further comprises indicating to lower layers information regarding AI / ML activation or deactivation.29.The method of claim 1, further comprising reporting the inference to the NW.30.The method of claim 29, wherein the inference comprises at least one of:top N beams with corresponding N Reference Signal Received Power (RSRP) values, where N ≥1;top N beams with RSRP values of M beams among the top N beams, where N>M≥1;top N beams whose RSRP values are greater than a predefined RSRP threshold value; ora flag for indicating that an RSRP value, if present for a present beam ID, is inferred to measured.31.A method performed by a network (NW) in communication with a user equipment (UE) in a wireless communication network, comprising:performing a testing procedure for Artificial Intelligence or Machine Learning (AIML) capabilities of the UE in collaboration with the UE;transmitting to the UE an AIML configuration for one or more AIML features, functionalities, or models selected by the NW from AIML capabilities of the UE based at least partially on the testing procedure; andreceiving an inference and / or measurement report from the UE, the inference and / or measurement report being generated by the UE based on the AIML configuration.32.The method of claim 31, wherein the testing procedure comprises:transmitting a request message to the UE for testing the AIML features, functionalities, or models of the UE;andreceiving a response message from the UE, the response message comprising at least an inference result generated by the UE using at least one AIML feature, functionality, or model.33.The method of claim 32, wherein the request message comprises at least one of:an indicator for indicating AIML features, functionalities, or models for testing; andreference input data to the indicated applicable AIML features, functions, or models for testing.34.The method of claim 33, wherein the response message comprises at least one of:the indicator for indicating the applicable AIML features, functionalities, or models for testing; oran output data list comprising the inference result.35.The method of claim 34, wherein the response message further comprises a set of UE additional conditions, the UE additional conditions and the inference result being used by the NW to evaluate the one or more AIML features, functionalities, or models.36.The method of claim 1, wherein the AIML configuration comprises an RRC configuration structure.37.The method of claim 36, wherein the AIML configuration is included in a serving cell configuration structure.38.The method of claim 37, wherein the AIML configuration comprises a list of configurations for the one or more AIML features, functionalities, or models.39.The method of claim 38, wherein the AIML configuration comprises a hierarchical listing of configurations for the one or more AIML features, functionalities, or models.40.The method of claim 38, wherein each of the list of configurations for an AIML functionality comprises at least one of:an index or ID of the AIML functionality;an inference related configuration for the AIML functionality;a performance monitoring related configuration for the AIML functionality; ora legacy configuration for a non-AI functionality corresponding to the AIML functionality.41.The method of claim 40, wherein the inference related configuration comprises at least one of:a measurement configuration for obtaining input data for inference for the AIML functionality;an inference reporting configuration for performing inference of output data associated with the AIML functionality;a performance monitoring configuration for inference performance monitoring of the AIML functionality; ora performance reporting configuration for reporting the inference performance monitoring of the AIML functionality.42.The method of claim 36, wherein the AIML configuration is included in a Channel State Information (CSI) report configuraiton.43.The method of claim 42, wherein the CSI report configuration comprises a list of identifiers for the one or more AIML features, functionalities, or models.44.The method of claim 42, wherein the CSI report configuration comprises a report quantity configuration, the report quantity configuration comprising both a set of inference configurations for the one or more AIML features, functionalities, or models.45.The method of claim 42, wherein the CSI report configuration comprises a timing configuration for AIML inference report.46.The method of claim 45, wherein the timing configuration comprises at list one of:periodic reporting;semi persistence reporting on Physical Uplink Shared Channel (PUSCH) ;Semi persistent reporting on PUCCH; oraperiodic.47.The method of claim 42, wherein the CSI report configuration comprises a measurement resource configuration for inference for the one or more AIML features, functionalities, or models.48.The method of claim 42, wherein the CSI report configuration comprises a measurement resource configuration for monitoring inference performance of the one or more AIML features, functionalities, or models.49.The method of claim 31, further comprising transmitting a signaling to the UE to activate or deactivate at least one of the one or more AIML features, functionalities, or models.50.The method of claim 49, wherein the signaling comprises a DL MAC CE structure.51.The method of claim 50, wherein the DL MAC CE structure comprises:a first octet with plurality of bits indicating whether to activate or deactivate associated AIML functionalities associated with corresponding serving cell;a second octet having a plurality of bits indicating a number of subsequent octets in the DL MAC CE structure for the corresponding serving cells; andthe subsequent octets indicating AIML features and functionalities to be Activated or deactivated for corresponding serving cells.52.The method of claim 49, wherein the signaling comprises a semi-persistent CSI reporting activation / deactivation signaling.53.The method of claim 31, wherein the inference comprises at least one of:top N beams with corresponding N Reference Signal Received Power (RSRP) values, where N ≥1;top N beams with RSRP values of M beams among the top N beams, where N>M≥1;top N beams whose RSRP values are greater than a predefined RSRP threshold value; ora flag for indicating that an RSRP value, if present for a present beam ID, is inferred to measured.54.The UE or the network node of any one of claims 1 to 53, the UE or the network node comprising a processor and a memory, wherein the processor is configured to read computer code from the memory to cause the UE or the network node to perform the method of any one of claims 1 to 53.55.A computer program product comprising a non-transitory computer-readable program medium with computer code stored thereupon, the computer code, when executed by a processor of the UE or the network node of any one of claims 1 to 53, causing the processor to implement the method of any one of claims 1 to 53.
Citation Information
Patent Citations
Artificial intelligence (AI) model training method and device in wireless network
CN116264684A
UE capability for ai / ml
US20220360973A1
Machine learning model management and assistance information
WO2023206501A1
Cited By
Performance test for functionality change
WO2026077600A1