Method and apparatus for reference symbol pattern adaptation

Through machine learning technology, dynamically adjusting the density of the reference signal, the problem of low RS density configuration efficiency in 5G mobile communication systems is solved, and flexible channel management and bandwidth optimization are achieved.

CN119096505APending Publication Date: 2024-12-06SAMSUNG ELECTRONICS CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202380029196.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Priority Date
2023-01-25
Filing Date
2023-03-23
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The prior art is difficult to effectively manage the density and configuration of reference signals (RS) in 5G mobile communication systems, resulting in low bandwidth usage efficiency and inability to flexibly cope with different channel conditions.

Method used

Using machine learning (ML) technology to adaptively adjust the time, frequency, or spatial density of the reference signal, dynamically adjust the configuration of the RS by exchanging information elements between the user equipment (UE) and the base station (BS) to adapt to the underlying wireless Channel changes.

Benefits of technology

It realizes that the number of antenna ports or the density of the reference signal is reduced without losing throughput, improves channel utilization efficiency, adapts to different channel conditions, and reduces system complexity.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119096505A_ABST
    Figure CN119096505A_ABST
Patent Text Reader

Abstract

Machine learning (ML) adaptation of any one of reference signal (RS) time density, RS frequency density, RS spatial density, or the number of transmit / receive points (TRPs) of transmitting RS provides a configuration of lower RS density or less TRPs of transmitting RS without significant throughput loss if appropriate. The determination of switching from high-density transmission to low-density transmission, reducing the number of antenna ports or TRPs that transmit RSs, or fallback to high-density transmission may be made by an ML model, optionally with UE assistance information.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present disclosure relates generally to applying machine learning models to selection of reference signal transmission modes, and more particularly to reducing bandwidth usage for reference signal transmission when possible. Background Art

[0002] 5G mobile communication technology defines a wide frequency band so that high transmission rates and new services are possible, and can be realized not only in "below 6 GHz" frequency bands such as 3.5 GHz, but also in "above 6 GHz" frequency bands called mmWave (millimeter wave) including 28 GHz and 39 GHz. In addition, the implementation of the sixth generation (6G) mobile communication technology (called a super 5G system) in the terahertz frequency band (e.g., 95 GHz to 3 THz frequency band) has been considered in order to achieve a transmission rate fifty times faster than that of 5G mobile communication technology and an ultra-low latency of one tenth of that of 5G mobile communication technology.

[0003] At the beginning of the development of 5G mobile communication technology, in order to support services and meet performance requirements related to enhanced mobile broadband (eMBB), ultra-reliable low latency communication (URLLC), and massive machine type communication (mMTC), there has been ongoing standardization on beamforming and massive MIMO for mitigating radio wave path loss in mmWave and increasing radio wave transmission distance in mmWave, parameter sets supporting dynamic operation for efficient utilization of mmWave resources and time slot formats (e.g., operating multiple subcarrier spacings), initial access technology for supporting multi-beam transmission and broadband, definition and operation of BWP (bandwidth part), new channel coding methods such as LDPC (low-density parity check) codes for large-volume data transmission and polar codes for highly reliable transmission of control information, L2 preprocessing, and network slicing for providing dedicated networks dedicated to specific services.

[0004] Currently, in view of the services supported by 5G mobile communication technology, discussions are ongoing on improvements and performance enhancements of initial 5G mobile communication technology, and there is already physical layer standardization on technologies such as V2X (Vehicle to Everything) for assisting driving decisions made by autonomous vehicles based on information about the location and status of the vehicle transmitted by the vehicle and for enhancing user convenience, NR-U (New Radio Unlicensed) for system operation intended to comply with various regulatory-related requirements in unlicensed bands, NR UE power saving, non-terrestrial network (NTN) (which is UE-satellite direct communication for providing coverage in areas where communication with terrestrial networks is unavailable), and positioning.

[0005] In addition, there is ongoing standardization of air interface architecture / protocol aspects of technologies such as Industrial Internet of Things (IIoT) for supporting new services through interworking and integration with other industries, IAB (Integrated Access and Backhaul) for providing nodes for network service area expansion by supporting wireless backhaul links and access links in an integrated manner, mobility enhancements including conditional switching and DAPS (Dual Active Protocol Stack) switching, and two-step random access (two-step RACH for NR) for simplifying the random access process. Standardization is also ongoing on the 5G baseline architecture (e.g., service-based architecture or service-based interface) for combining network function virtualization (NFV) and software defined network (SDN) technologies, and the system architecture / services of mobile edge computing (MEC) for receiving services based on UE location.

[0006] As 5G mobile communication systems are commercialized, the number of connected devices, which has been growing exponentially, will be connected to the communication network, and it is therefore expected that enhanced functions and performance of the 5G mobile communication systems and the integrated operation of connected devices will be necessary. To this end, new research related to extended reality (XR) is arranged for effectively supporting AR (augmented reality), VR (virtual reality), MR (mixed reality), etc., improving 5G performance and reducing complexity by utilizing artificial intelligence (AI) and machine learning (ML), AI service support, metaverse service support, and drone communication.

[0007] Furthermore, such developments in 5G mobile communication systems will serve as a foundation not only for the development of new waveforms that provide terahertz band coverage for 6G mobile communication technology, multi-antenna transmission technologies such as full-dimensional MIMO (FD-MIMO), array antennas, and massive antennas, metamaterial-based lenses and antennas for improving terahertz band signal coverage, high-dimensional spatial multiplexing technology using OAM (orbital angular momentum), and RIS (reconfigurable smart surface), but also for the development of full-duplex technology that improves the frequency efficiency of 6G mobile communication technology and improves system networks, AI-based communication technology for achieving system optimization by leveraging satellites and AI (artificial intelligence) from the design stage and internalizing end-to-end AI support functions, and next-generation distributed computing technology for achieving services at a complexity level that exceeds the operational capability limits of UEs by utilizing ultra-high-performance communication and computing resources. Summary of the invention

[0008] Technical issues

[0009] Aspects of the present disclosure are to address at least the above-mentioned problems and / or disadvantages and to provide at least the advantages described below. Therefore, one aspect of the present disclosure is to provide an enhanced communication method.

[0010] Technical Solutions

[0011] Machine learning (ML) adaptation of any one of reference signal (RS) time density, RS frequency density, RS spatial density, or the number of transmit / receive points (TRPs) from which RS is transmitted provides configurations with lower RS ​​density or fewer TRPs without significant throughput loss where appropriate. The determination to switch from high-density transmission to low-density transmission, to reduce the number of antenna ports or TRPs from which RS is transmitted, or to fall back to high-density transmission can be made by an ML model, optionally with UE assistance information.

[0012] In a first embodiment, a method includes information indicating the ability of a UE to support machine learning (ML) adaptation of a reference signal (RS) density in a domain. The method also includes receiving first configuration information from a BS, the first configuration information indicating one or more of: enabling or disabling ML adaptation of RS density in a domain, an ML model for RS density adaptation in a domain, an updated model parameter of an ML model, or whether a model parameter received from a UE will be used for ML adaptation of RS density in a domain. The method also includes receiving second configuration information from a BS, the second configuration information indicating RS density in a first domain, receiving an RS having a density in the first domain, and sending auxiliary information to the BS. In response to sending the auxiliary information, the method includes: receiving third configuration information indicating RS density in a second domain from the BS, wherein the RS density in the first domain is greater than the RS density in the second domain. Finally, the method includes receiving an RS having an RS density in a second domain.

[0013] In a second embodiment, a user equipment (UE) includes a processor and a transceiver operably coupled to the processor. The transceiver is configured to send information indicating the ability of the UE to support machine learning (ML) adaptation of reference signal (RS) density in a domain to a base station (BS). The transceiver is also configured to receive first configuration information from the BS, the first configuration information indicating one or more of the following: enabling or disabling ML adaptation of RS density in the domain, an ML model for RS density adaptation in the domain, an updated model parameter of the ML model, or whether the model parameter received from the UE will be used for ML adaptation of RS density in the domain. The transceiver is also configured to receive second configuration information from the BS, the second configuration information indicating RS density in the first domain, receiving RS with RS density in the first domain, and sending auxiliary information to the BS. In response to sending the auxiliary information, the transceiver is configured to receive third configuration information indicating RS density in the second domain from the BS, wherein the RS density in the first domain is greater than the RS density in the second domain. The transceiver is also configured to receive RS with RS density in the second domain.

[0014] In a third embodiment, a base station (BS) includes a processor and a transceiver operably coupled to the processor. The transceiver is configured to receive information from a user equipment (UE) indicating the ability of the UE to support machine learning (ML) adaptation of reference signal (RS) density in a domain. The transceiver is also configured to send a first configuration information to the UE, the first configuration information indicating one or more of the following: enabling or disabling ML adaptation of RS density in a domain, an ML model for RS density adaptation in a domain, an updated model parameter of the ML model, or whether the model parameter received from the UE will be used for ML adaptation of RS density in a domain. The transceiver is also configured to send a second configuration information to the UE, the second configuration information indicating the RS density in the first domain, sending an RS with RS density in the first domain, and receiving auxiliary information from the UE. In response to receiving the auxiliary information, the transceiver is configured to send a third configuration information indicating the RS density in the second domain to the UE, wherein the RS density in the first domain is greater than the RS density in the second domain. The transceiver is also configured to send an RS with RS density in the second domain.

[0015] In any of the aforementioned embodiments, the domain may be one of a time domain, a frequency domain, or a spatial domain.

[0016] In any of the aforementioned embodiments, the backoff request sent to the BS may indicate an RS density in a domain, the fourth configuration sent by the BS may indicate an RS density in a third domain, and then an RS having an RS density in the third domain may be received.

[0017] In the previous embodiment, the RS density in the third domain may be greater than the RS density in the second domain.

[0018] In any of the foregoing embodiments, the received information element may indicate one of an RS frequency or time density value of a continuous resource and an RS frequency or time density value of some resources mapped to antenna ports configured for the RS, and then the RS may be received in the resource by switching between the RS density in the first domain and the RS density in the second domain based on the information element.

[0019] In any of the aforementioned embodiments, the auxiliary information may include one or more of a block error rate, a UE speed, a UE acceleration, and a recommended RS density in the domain.

[0020] In any of the foregoing embodiments, a configuration of RS density in a forbidden domain may be received from a BS.

[0021] Other technical features may be apparent to those skilled in the art from the following drawings, descriptions and claims.

[0022] Before the following detailed description, it may be advantageous to set forth the definitions of certain words and phrases used throughout this patent document. The term "coupling" and its derivatives refer to any direct or indirect communication between two or more elements, whether these elements are in physical contact with each other. The terms "send", "receive" and "communication" and their derivatives include direct and indirect communication. The terms "include" and "comprise" and their derivatives mean including but not limited to this. The term "or" is inclusive, meaning and / or. The phrase "associated with..." and its derivatives mean including, included in, interconnected with, included in, connected to or connected with, coupled to or coupled with, can communicate with, collaborate with, interweave, juxtapose, be close to, be bound to or bound with, have, have the nature of, have a relationship with, etc. The term "controller" means any device, system or part thereof that controls at least one operation. Such a controller can be implemented in hardware or a combination of hardware and software and / or firmware. The functions associated with any particular controller can be centralized or distributed, whether local or remote. When used with a list of items, the phrase "at least one of" means that different combinations of one or more of the listed items can be used, and that only one of the items in the list may be required. For example, "at least one of A, B, and C" includes any of the following combinations: A, B, C, A and B, A and C, B and C, and A and B and C. Thus, a collection of items can be a single item or a collection of two or more items.

[0023] In addition, the various functions described below can be implemented or supported by one or more computer programs, each of which is formed by a computer-readable program code and embodied in a computer-readable medium. The terms "application" and "program" refer to one or more computer programs, software components, instruction sets, processes, functions, objects, classes, instances, related data or a portion thereof suitable for implementation in a suitable computer-readable program code. The phrase "computer-readable program code" includes any type of computer code, including source code, object code and executable code. The phrase "computer-readable medium" includes any type of medium that can be accessed by a computer, such as a read-only memory (ROM), a random access memory (RAM), a hard drive, a compact disk (CD), a digital video disc (DVD) or any other type of memory. "Non-transitory" computer-readable media does not include wired, wireless, optical or other communication links that transmit temporary electrical signals or other signals. Non-transitory computer-readable media include media that can permanently store data and media that can store data and rewrite data later, such as rewritable optical disks or erasable memory devices.

[0024] Definitions for certain other words and phrases are provided throughout this patent document. Those of ordinary skill in the art should understand that in many, if not most, instances, such definitions apply to prior, as well as future uses of such defined words and phrases.

[0025] Beneficial effects

[0026] This disclosure describes a framework for supporting AI / ML techniques for RS time density adaptation based on statistics of underlying randomly varying wireless channels. The corresponding signaling details are discussed in this disclosure. This disclosure solves the problem that RS time density adaptation currently depends on network implementation.

[0027] The present disclosure provides a method that a network can use to configure the time density of RS patterns using an AI / ML-based solution. The present disclosure also provides a framework for self-adapting the time density of RS patterns based on UE inference and information. Details about supporting AI / ML techniques for RS time density adaptation are disclosed, including information elements to be exchanged between a transmitter and a receiver. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] For a more complete understanding of the present disclosure and its advantages, reference is now made to the following description taken in conjunction with the accompanying drawings, in which:

[0029] Figure 1 An exemplary networking system utilizing reference signal time density configuration according to various embodiments of the present disclosure is shown;

[0030] Figure 2 An exemplary base station (BS) configured with reference signal time density according to various embodiments of the present disclosure is shown;

[0031] Figure 3 An exemplary electronic device for communicating in a networked computing system utilizing reference signal time density configuration according to various embodiments of the present disclosure is shown;

[0032] Figure 4 A high-level flow chart illustrating an example of BS operations for supporting RS time density adaptation according to various embodiments of the present disclosure;

[0033] Figure 5 A high-level flow chart illustrating an example of UE operation for supporting RS time density adaptation according to various embodiments of the present disclosure is shown;

[0034] Figure 6 A high-level flow chart illustrating an example of BS operations for supporting fallback to RS time high density configuration according to various embodiments of the present disclosure;

[0035] Figure 7A high-level flow chart illustrating an example of UE operation for supporting fallback to RS time high density configuration according to various embodiments of the present disclosure;

[0036] Figure 8 A high-level flow chart illustrating an example of BS operation for supporting RS time disabling according to various embodiments of the present disclosure;

[0037] Fig. 9 A high-level flow chart illustrating an example of UE operation for supporting RS time disabling according to various embodiments of the present disclosure;

[0038] Fig.10 An example of RS time density adaptation according to various embodiments of the present disclosure is shown;

[0039] Fig.11 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown;

[0040] Fig.12 An example of a new MAC CE for RS time high density backoff request according to various embodiments of the present disclosure is shown;

[0041] Fig.13 An example of an AI / ML model architecture supporting RS time density adaptation according to various embodiments of the present disclosure is shown;

[0042] Fig.14 A high-level flow chart illustrating an example of BS operations for supporting RS frequency density adaptation according to various embodiments of the present disclosure;

[0043] Fig.15 A high-level flow chart illustrating an example of UE operation for supporting RS frequency density adaptation according to various embodiments of the present disclosure;

[0044] Fig.16 A high-level flow chart illustrating an example of BS operations for supporting fallback to RS frequency high density configuration according to various embodiments of the present disclosure;

[0045] Fig.17 A high-level flow chart illustrating an example of UE operation for supporting fallback to RS frequency high density configuration according to various embodiments of the present disclosure;

[0046] Fig.18 A high-level flow chart illustrating an example of BS operation for supporting RS frequency disabling according to various embodiments of the present disclosure;

[0047] Fig.19 A high-level flow chart illustrating an example of UE operation for supporting RS frequency disabling according to various embodiments of the present disclosure;

[0048] Fig. 20 A high-level flow chart illustrating an example of BS operation for supporting RS frequency-based BWP switching according to various embodiments of the present disclosure;

[0049] Fig.21 A high-level flow chart illustrating an example of UE operation for supporting RS frequency-based BWP switching according to various embodiments of the present disclosure;

[0050] Fig. 22 An example of RS frequency density adaptation according to various embodiments of the present disclosure is shown;

[0051] Fig.23 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown;

[0052] Fig.24 An example of a new MAC CE for RS frequency high density backoff request according to various embodiments of the present disclosure is shown;

[0053] Fig.25 An example of an AI / ML model architecture supporting RS frequency density adaptation according to various embodiments of the present disclosure is shown;

[0054] Fig.26 A high-level flow chart illustrating an example of BS operations for supporting RS spatial density adaptation according to various embodiments of the present disclosure;

[0055] Fig. 27 A high-level flow chart illustrating an example of UE operation for supporting RS spatial density adaptation according to various embodiments of the present disclosure;

[0056] Fig.28 A high-level flow chart illustrating an example of BS operation for supporting fallback to RS spatial high density configuration according to various embodiments of the present disclosure;

[0057] Fig.29 A high-level flow chart illustrating an example of UE operation for supporting fallback to RS spatial high density configuration according to various embodiments of the present disclosure;

[0058] Fig.30 A high-level flow chart illustrating an example of BS operation for supporting RS spatial disabling according to various embodiments of the present disclosure;

[0059] Fig.31 A high-level flow chart illustrating an example of UE operation for supporting RS spatial disabling according to various embodiments of the present disclosure;

[0060] Fig.32A high-level flow chart illustrating an example of BS operations for supporting RS space-based port switching according to various embodiments of the present disclosure;

[0061] Fig.33 A high-level flow chart illustrating an example of UE operation for supporting RS space-based port switching according to various embodiments of the present disclosure;

[0062] Fig.34 An example of RS spatial density adaptation according to various embodiments of the present disclosure is shown;

[0063] Fig.35 An example of a new MAC CE for BS assistance information reporting according to various embodiments of the present disclosure is shown;

[0064] Fig.36 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown;

[0065] Fig.37 An example of a new MAC CE for RS spatial high density backoff request according to various embodiments of the present disclosure is shown;

[0066] Fig.38 An example of an AI / ML model architecture supporting RS spatial density adaptation according to various embodiments of the present disclosure is shown;

[0067] Fig.39 A high-level flow chart illustrating an example of BS operation for supporting multi-TRP RS density adaptation according to various embodiments of the present disclosure;

[0068] Fig.40 A high-level flow chart illustrating an example of UE operation for supporting multi-TRP RS density adaptation according to various embodiments of the present disclosure;

[0069] Fig.41 A high-level flow chart illustrating an example of BS operation for supporting multi-TRP RS disabling according to various embodiments of the present disclosure;

[0070] Fig.42 A high-level flow chart illustrating an example of UE operation for supporting multi-TRP RS disabling according to various embodiments of the present disclosure;

[0071] Fig.43 A high-level flow chart illustrating an example of BS operation for supporting multi-TRP CSI information according to various embodiments of the present disclosure;

[0072] Fig.44A high-level flow chart illustrating an example of UE operation for supporting multi-TRP CSI information according to various embodiments of the present disclosure;

[0073] Fig.45 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown;

[0074] Fig.46 An example of a new MAC CE for DMRS disabling recommendation for a given TRP according to various embodiments of the present disclosure is shown;

[0075] Fig.47 An example of a new MAC CE for CSI information for a given TRP according to various embodiments of the present disclosure is shown; and

[0076] Fig.48 An example of an AI / ML model architecture supporting multi-TRP RS density adaptation according to various embodiments of the present disclosure is shown. DETAILED DESCRIPTION

[0077] The drawings and various embodiments used to describe the principles of the present disclosure herein are illustrative only and should not be interpreted in any way as limiting the scope of the present disclosure. In addition, those skilled in the art will appreciate that the principles of the present disclosure can be implemented in any appropriately arranged wireless communication system.

[0078] References:

[0079] [1]3GPP, TS 38.211, 5G; NR; Physical channels and modulation

[0080] [2]3GPP, TS 38.331, 5G; NR; Radio Resource Control (RRC); Protocol specification

[0081] [3]3GPP, TS 38.321, 5G; NR; Medium Access Control (MAC); Protocol specification

[0082] [4]https: / / www.tensorflow.org / api_docs / python / tf / keras / layers / ConvLSTM2D

[0083] [5]N.Ballas, L.Yao, C.Pal, and A.Courville, "Delving deeper into convolutional networks for learning video representations," https: / / arxiv.org / pdf / 1511.06432.pdf

[0084] [6] Y.Zhou and K.Hauser, "Incorporating side-channel information into convolutional neural networks for robotic tasks," http: / / motion.pratt.duke.edu / sidechannel / pdf / icra2017incorporating.pdf

[0085] [7]3GPP, TS 38.214, 5G; NR; Physical layer procedures for data

[0086] [8] J.Figueroa Barraza, E.López Drogett, and MRMartins, "Towards interpretable deep learning: Feature selection framework for prognostics and health management using deep neural networks", Sensors, vol.21, no.17, pp.5888.

[0087] The above references are incorporated herein by reference.

[0088] abbreviation:

[0089] 3GPP Third Generation Partnership Project

[0090] 4G Fourth Generation

[0091] 5G Fifth Generation

[0092] AI

[0093] BS Base Station

[0094] BWP Bandwidth Part

[0095] C-RNTI Cell-Radio Network Temporary Identifier

[0096] CDM Code Division Multiplexing

[0097] CE Control Elements

[0098] CNN Convolutional Neural Network

[0099] CORESET Control Resource Set

[0100] CRNN Convolutional Recurrent Neural Network

[0101] CSI Channel State Information

[0102] CQI Channel Quality Indicator

[0103] DCI Downlink Control Information

[0104] DL Downlink

[0105] DMRS Demodulation Reference Signal

[0106] gNB gNodeB(BS)

[0107] GRU Gated Recurrent Unit

[0108] LSTM Long Short-Term Memory

[0109] MAC Media Access Control

[0110] ML Machine Learning

[0111] NC-JT Non-coherent joint transmission

[0112] NR New Radio

[0113] OFDM Orthogonal Frequency Division Multiplexing

[0114] PDSCH Physical Downlink Shared Channel

[0115] PMI Precoding matrix indicator

[0116] PUCCH Physical Uplink Control Channel

[0117] PUSCH Physical Uplink Shared Channel

[0118] QCL Quasi-Co-sited

[0119] RE Resource Element

[0120] RRC Radio Resource Control

[0121] RS reference signal

[0122] SB Subband

[0123] SRS Sounding Reference Signal

[0124] TCI Transmission Configuration Indicator

[0125] TRP Send / Receive Point

[0126] UE User Equipment

[0127] UL Uplink

[0128] In 5G NR, several types of RS have been defined. For example, CSI-RS is used for DL ​​communication between gNB and UE, where the UE uses the received CSI-RS to measure DL CSI and report those measurements to the gNB. In addition, DMRS is used by the receiver (for DL ​​or UL communication) to estimate CSI; this estimate is used to demodulate the received data.

[0129] Before transmitting RSs such as CSI-RS and DMRS, a time-frequency mapping function is applied to the RSs, resulting in a specific RS pattern. The RS pattern depends on parameters such as the transmit antenna port, the CDM type, and whether frequency hopping is enabled.

[0130] It may be advantageous to vary the temporal density of the RS pattern based on the statistics of the underlying randomly varying wireless channel. For example, if the channel selectivity degrades over time, reducing the temporal density of the RS pattern may have a negligible effect on the CSI estimation error—while reducing the signaling overhead. As another example, if the channel is static, RS signaling may be disabled (at least temporarily), assuming that the receiver maintains the current channel state in its memory.

[0131] 5G NR supports flexibility in selecting RS patterns, where the selection of RS patterns is based on the statistics of the underlying randomly varying radio channel. For example, the parameter dmrs-AdditionalPosition can be used to increase the number of DMRS in a given time slot in high mobility scenarios. As another example, the parameters periodicityAndOffset-p and periodicityAndOffset-sp can be used to change the periodicity (and time slot offset) of the SRS. The details of the algorithm used to select the RS pattern are generally left to the network.

[0132] This disclosure describes a framework for supporting AI / ML techniques for RS time density adaptation based on statistics of underlying randomly varying wireless channels. The corresponding signaling details are discussed in this disclosure. This disclosure solves the problem that RS time density adaptation currently depends on network implementation.

[0133] The present disclosure provides a method that a network can use to configure the time density of RS patterns using an AI / ML-based solution. The present disclosure also provides a framework for self-adapting the time density of RS patterns based on UE inference and information. Details about supporting AI / ML techniques for RS time density adaptation are disclosed, including information elements to be exchanged between a transmitter and a receiver.

[0134] The following Figure 1-Figure 3 Various embodiments are described for implementing and using Orthogonal Frequency Division Multiplexing (OFDM) or Orthogonal Frequency Division Multiple Access (OFDMA) communication techniques in a wireless communication system. Figure 1-Figure 3 The description is not meant to imply physical or architectural limitations to the manner in which different embodiments may be implemented. Different embodiments of the present disclosure may be implemented in any suitably arranged communications system.

[0135] Figure 1 An exemplary networking system utilizing reference signal time density configuration according to various embodiments of the present disclosure is shown. Figure 1 The embodiment of the wireless network shown is for illustration only. Other embodiments of the wireless network 100 may be used without departing from the scope of the present disclosure.

[0136] like Figure 1 As shown, the wireless network includes gNB 101 (e.g., base station, BS), gNB 102, and gNB 103. gNB 101 communicates with gNB 102 and gNB 103. gNB 101 also communicates with at least one network 130 such as the Internet, a proprietary Internet Protocol (IP) network, or other data network.

[0137] The gNB 102 provides wireless broadband access to a network 130 for a plurality of first user equipments (UEs) within a coverage area 120 of the gNB 102. The plurality of first UEs include a UE 111 that may be located at a small business (SB); a UE 112 that may be located at an enterprise (E); a UE 113 that may be located at a WiFi hotspot (HS); a UE 114 that may be located at a first residence (R); a UE 115 that may be located at a second residence (R); and a UE 116 that may be a mobile device (M) such as a cellular phone, a wireless laptop, a wireless personal digital assistant (PDA), etc. The gNB 103 provides wireless broadband access to the network 130 for a plurality of second UEs within a coverage area 125 of the gNB 103. The plurality of second UEs include the UE 115. In some embodiments, one or more of gNBs 101-103 may communicate with each other and with UEs 111-116 using 5G / NR, Long Term Evolution (LTE), Long Term Evolution-Advanced (LTE-A), WiMAX, WiFi, or other wireless communication technologies.

[0138] Depending on the network type, the term "base station" or "BS" may refer to any component (or collection of components) configured to provide wireless access to a network, such as a transmission point (TP), a transmission-reception point (TRP), an enhanced base station (eNodeB or eNB), a 5G / NR base station (gNB), a macro cell, a femto cell, a WiFi access point (AP), or other wireless-capable devices. The base station may provide wireless access according to one or more wireless communication protocols, for example, 5G / NR Third Generation Partnership Project (3GPP) NR, Long Term Evolution (LTE), Advanced LTE (LTE-A), High Speed ​​Packet Access (HSPA), Wi-Fi 802.11a / b / g / n / ac, etc. For convenience, the terms "BS" and "TRP" may be used interchangeably in this disclosure to refer to network infrastructure components that provide wireless access to remote terminals. In addition, depending on the network type, the term "user equipment" or "UE" may refer to any component, such as a "mobile station", "user station", "remote terminal", "wireless terminal", "reception point" or "user equipment". For convenience, the terms "user equipment" and "UE" are used in this patent document to refer to a remote wireless device that wirelessly accesses a BS, whether the UE is a mobile device (such as a mobile phone or smartphone) or is generally considered a fixed device (such as a desktop computer or vending machine).

[0139] The dashed lines illustrate the approximate extents of coverage areas 120 and 125, which are shown as generally circular for purposes of illustration and explanation only. It should be clearly understood that coverage areas associated with gNBs, such as coverage areas 120 and 125, may have other shapes, including irregular shapes, depending on the configuration of the gNB and variations in the radio environment associated with natural and man-made obstacles.

[0140] although Figure 1 An example of a wireless network is shown, but Figure 1 Various changes may be made. For example, the wireless network may include any number of gNBs and any number of UEs in any suitable arrangement. In addition, gNB 101 may communicate directly with any number of UEs and provide wireless broadband access to network 130 for these UEs. Similarly, each gNB 102-103 may communicate directly with network 130 and provide direct wireless broadband access to network 130 to the UEs. In addition, gNBs 101, 102, and / or 103 may provide access to other or additional external networks, such as an external telephone network or other type of data network.

[0141] Figure 2 An exemplary base station (BS) utilizing reference signal time density configuration according to various embodiments of the present disclosure is shown. Figure 2 The embodiment of the gNB 102 shown in FIG. is for illustration only, and Figure 1 gNBs 101 and 103 in the embodiment may have the same or similar configurations. However, gNBs have various configurations. Figure 2 The scope of the present disclosure is not limited to any particular implementation of the gNB.

[0142] like Figure 2 As shown, the gNB 102 includes multiple antennas 205a-205n, multiple transceivers 210a-210n, a controller / processor 225, a memory 230, and a backhaul or network interface 235.

[0143] The transceivers 210a-210n receive incoming radio frequency (RF) signals from the antennas 205a-205n, such as signals sent by UEs in the network 100. The incoming RF signals of the transceivers 210a-210n are down-converted to generate IF or baseband signals. The IF or baseband signals are processed by receive (RX) processing circuits in the transceivers 210a-210n and / or the controller / processor 225 to generate processed baseband signals by filtering, decoding and / or digitizing the baseband or IF signals. The controller / processor 225 may further process the baseband signals.

[0144] The transmit (TX) processing circuitry in the transceivers 210a-210n and / or the controller / processor 225 receives analog or digital data (such as voice data, web data, email, or interactive video game data) from the controller / processor 225. The TX processing circuitry encodes, multiplexes, and / or digitizes the outgoing baseband data to generate a processed baseband or IF signal. The transceivers 210a-210n up-convert the baseband or IF signal to an RF signal that is transmitted via the antennas 205a-205n.

[0145] The controller / processor 225 may include one or more processors or other processing devices that control the overall operation of the gNB 102. For example, the controller / processor 225 may control the reception of UL channel signals and the transmission of DL channel signals by the transceivers 210a-210n in accordance with well-known principles. The controller / processor 225 may also support additional functionality, such as more advanced wireless communication functionality. For example, the controller / processor 225 may support beamforming or directional routing operations, in which outgoing / incoming signals from / to multiple antennas 205a-205n are weighted differently to effectively direct the outgoing signals in a desired direction. The controller / processor 225 may support any of a variety of other functions in the gNB 102.

[0146] Controller / processor 225 is also capable of executing programs and other processes, such as an OS, that reside in memory 230. Controller / processor 225 may move data into or out of memory 230 as required by the executing process.

[0147] The controller / processor 225 is also coupled to a backhaul or network interface 235. The backhaul or network interface 235 allows the gNB 102 to communicate with other devices or systems via a backhaul connection or via a network. The interface 235 may support communication over any suitable wired or wireless connection. For example, when the gNB 102 is implemented as part of a cellular communication system (such as a system supporting 5G / NR, LTE, or LTE-A), the interface 235 may allow the gNB 102 to communicate with other gNBs via a wired or wireless backhaul connection. When the gNB 102 is implemented as an access point, the interface 235 may allow the gNB 102 to communicate over a wired or wireless local area network or over a wired or wireless connection to a larger network (such as the Internet). The interface 235 includes any suitable structure that supports communication over a wired or wireless connection (such as Ethernet or a transceiver).

[0148] Memory 230 is coupled to controller / processor 225. A portion of memory 230 may include RAM, and another portion of memory 230 may include flash memory or other ROM.

[0149] although Figure 2 An example of a gNB 102 is shown, but the Figure 2 For example, gNB 102 may include any number of Figure 2 Each component shown in . In addition, Figure 2 The various components in may be combined, further subdivided, or omitted, and additional components may be added according to specific needs.

[0150] Figure 3 Exemplary electronic devices for communicating in a networked computing system utilizing reference signal time density configuration according to various embodiments of the present disclosure are shown. Figure 3 The embodiment of UE 116 shown in FIG. 1 is for illustration only, and Figure 1 UEs 111-115 may have the same or similar configurations. However, UEs may have a variety of configurations, and Figure 3 The scope of the present disclosure is not limited to any particular implementation of the UE.

[0151] like Figure 3 As shown, UE 116 includes antenna 305, transceiver 310, and microphone 320. UE 116 also includes speaker 330, processor 340, input / output (I / O) interface (IF) 345, input 350, display 355, and memory 360. Memory 360 includes operating system (OS) 361 and one or more applications 362.

[0152] Transceiver 310 receives incoming RF signals from antenna 305 transmitted by a gNB of network 100. The incoming RF signals are down-converted by transceiver 310 to generate an intermediate frequency (IF) or baseband signal. The IF or baseband signal is processed by RX processing circuitry and / or processor 340 in transceiver 310 by filtering, decoding, and / or digitizing the baseband or IF signal to generate a processed baseband signal. The RX processing circuitry transmits the processed baseband signal to speaker 330 (such as for voice data) or is processed by processor 340 (such as for web browsing data).

[0153] The TX processing circuitry and / or processor 340 in transceiver 310 receives analog or digital voice data from microphone 320 or other outgoing baseband data (such as web data, email, or interactive video game data) from processor 340. The TX processing circuitry encodes, multiplexes, and / or digitizes the outgoing baseband data to generate a processed baseband or IF signal. Transceiver 310 up-converts the baseband or IF signal to an RF signal that is transmitted through antenna 305.

[0154] The processor 340 may include one or more processors or other processing devices and execute the OS 361 stored in the memory 360 to control the overall operation of the UE 116. For example, the processor 340 may control the RF transceiver 310 to receive DL channel signals and transmit UL channel signals according to well-known principles. In some embodiments, the processor 340 includes at least one microprocessor or microcontroller.

[0155] The processor 340 is also capable of executing other processes and programs resident in the memory 360. The processor 340 can move data into or out of the memory 360 as needed for the executed process. In some embodiments, the processor 340 is configured to execute the application 362 based on the OS 361 or in response to a signal received from the BS or operator. The processor 340 is also coupled to the I / O interface 345, which provides the UE 116 with the ability to connect to other devices such as laptops and handheld computers. The I / O interface 345 is the communication path between these accessories and the processor 340.

[0156] Processor 340 is also coupled to input 350 including, for example, a touch screen, a keypad, etc., and display 355. An operator of UE 116 may enter data into UE 116 using input 350. Display 355 may be a liquid crystal display, a light emitting diode display, or other display capable of presenting text and / or at least limited graphics, such as from a website.

[0157] Memory 360 is coupled to processor 340. A portion of memory 360 may include random access memory (RAM), and another portion of memory 360 may include flash memory or other read-only memory (ROM).

[0158] although Figure 3 An example of UE 116 is shown, but the Figure 3 Make various changes. For example, Figure 3 The various components in may be combined, further subdivided, or omitted, and additional components may be added as required. As a specific example, processor 340 may be divided into multiple processors, such as one or more central processing units (CPUs) and one or more graphics processing units (GPUs). In another example, transceiver 310 may include any number of transceivers and signal processing chains, and may be connected to any number of antennas. In addition, although Figure 3 The UE 116 is shown configured as a mobile phone or smart phone, but the UE may be configured to operate as other types of mobile or stationary devices.

[0159] Figure 4A high-level flow chart illustrating an example of BS operations for supporting RS time density adaptation according to various embodiments of the present disclosure is shown. Figure 4 The embodiment is for illustration only. Other embodiments of process 400 may be used without departing from the scope of the present disclosure.

[0160] Figure 4 is an example of a method 400 for supporting the operation of AI / ML technology for RS time density adaptation at a BS. In operation 401, the BS receives UE capability information from the UE, including support for ML methods for RS time density adaptation. In operation 402, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS time density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 403, the BS sends an RS time high density configuration message to the UE. In operation 404, the BS receives auxiliary information from the UE, wherein the auxiliary information may include a recommended RS time density, which will be described below. In operation 405, the BS sends an RS time low density configuration message to the UE.

[0161] Figure 5 A high-level flow chart illustrating an example of UE operation for supporting RS time density adaptation according to various embodiments of the present disclosure is shown. Figure 5 The embodiment is for illustration only. Other embodiments of process 500 may be used without departing from the scope of the present disclosure.

[0162] Figure 5 is an example of a method 500 for supporting the operation of AI / ML techniques for RS time density adaptation at a UE. In operation 501, the UE reports capability information to the BS, including support for ML methods for RS time density adaptation. In operation 502, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS time density adaptation, the ML model to be used, trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 503, the UE receives an RS time high density configuration message from the BS. In operation 504, the UE sends auxiliary information to the BS; the auxiliary information may include a recommended RS time density, which will be described below. In operation 505, the UE receives an RS time low density configuration message from the BS.

[0163] Figure 6 A high-level flow chart illustrating an example of BS operations for supporting fallback to RS time high density configuration according to various embodiments of the present disclosure is shown. Figure 6The embodiment is for illustration only. Other embodiments of process 600 may be used without departing from the scope of the present disclosure.

[0164] Figure 6 is an example of a method 600 for supporting a request to fall back from RS time low density mode to RS time high density mode at a BS. In operation 601, the BS receives UE capability information from the UE, including support for the ML method for RS time density adaptation. In operation 602, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS time density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 603, the BS sends an RS time high density configuration message to the UE. In operation 604, the BS receives auxiliary information from the UE; the auxiliary information may include a recommended RS time density, which will be described below. In operation 605, the BS sends an RS time low density configuration message to the UE. In operation 606, the BS receives a request to fall back to RS time high density mode from the UE. The request will be described in further detail below. In operation 607, the BS sends an RS time high density configuration message to the UE.

[0165] Figure 7 A high-level flow chart illustrating an example of UE operation for supporting fallback to RS time high density configuration according to various embodiments of the present disclosure is shown. Figure 7 The embodiment is for illustration only. Other embodiments of process 700 may be used without departing from the scope of the present disclosure.

[0166] Figure 7 is an example of a method 700 for supporting a request to fall back from RS time low density mode to RS time high density mode at a UE. In operation 701, the UE reports capability information to the BS, including support for the ML method for RS time density adaptation. In operation 702, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS time density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 703, the UE receives an RS time high density configuration message from the BS. In operation 704, the UE sends auxiliary information to the BS; the auxiliary information may include a recommended RS time density, which will be described below. In operation 705, the UE receives an RS time low density configuration message from the BS. In operation 706, the UE sends a request to the BS to fall back to RS time high density mode. The request will be described in further detail below. In operation 707, the UE receives an RS time high density configuration message from the BS.

[0167] Figure 8A high-level flow chart illustrating an example of BS operations for supporting RS time disabling according to various embodiments of the present disclosure is shown. Figure 8 The embodiment is for illustration only. Other embodiments of process 800 may be used without departing from the scope of the present disclosure.

[0168] Figure 8 is an example of a method 800 for supporting the operation of AI / ML techniques for RS time disabling at a BS. In operation 801, the BS receives UE capability information from the UE, including support for ML methods for RS time density adaptation. In operation 802, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS time density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 803, the BS sends an RS time density configuration message to the UE. In operation 804, the BS receives auxiliary information from the UE. The auxiliary information may include a recommendation for RS time disabling, which will be described below. In operation 805, the BS sends an RS time disabling configuration message to the UE.

[0169] Fig. 9 A high-level flow chart illustrating an example of UE operation for supporting RS time disabling according to various embodiments of the present disclosure is shown. Fig. 9 The embodiment is for illustration only. Other embodiments of process 900 may be used without departing from the scope of the present disclosure.

[0170] Fig. 9 is an example of a method 900 for supporting the operation of AI / ML techniques for RS time disabling at a UE. At operation 901, the UE reports capability information to the BS, including support for ML methods for RS time density adaptation. At operation 902, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS time density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. At operation 903, the UE receives an RS time density configuration message from the BS. At operation 904, the UE sends auxiliary information to the BS. The auxiliary information may include a recommendation for RS time disabling, which will be described in further detail below. At operation 905, the UE receives an RS time disabling configuration message from the BS.

[0171] Fig.10 An example of RS time density adaptation according to various embodiments of the present disclosure is shown. Fig.10 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0172] In 5G NR, a time slot consists of 14 OFDM symbols. For example, in the time domain, a "time resource" may correspond to an OFDM symbol, or to a time slot, a subframe, a radio frame, etc. Example 1000 of RS time density adaptation Fig.10 , where the rectangles represent time slots. Within a time slot, a certain number of symbols may contain DMRS, where the ratio of this number to 14 may represent the DMRS time density. In this example, a time slot 1001 with a high DMRS time density may be configured with a specific periodicity 1004. Other time slots 1002 within the duration of a PDSCH transmission 1003 may be configured with a low DMRS time density.

[0173] In one embodiment, the BS may configure the DMRS time density for the UE via PDCCH signaling. In another example, the BS may configure the DMRS time density for the UE via RRC configuration. Table 1 is an example of defining an information element (IE) DMRS-DownlinkConfigH to configure a DMRS with high time density. In this example, the DMRS pattern may be configured as periodic, semi-persistent, or aperiodic. DMRS-Sym represents a tuple of (OFDM symbol index, frequency density) values, and specifies a tuple for each OFDM symbol in a time slot. Similarly, IE DMRS-DownlinkConfigL may be defined to configure a DMRS with low time density. These IEs may facilitate switching between a DMRS pattern with high time density and a DMRS pattern with low time density in consecutive time slots. For example, if a DMRS pattern with high time density is configured as periodic, the UE may expect to receive the pattern in certain time slots. In other time slots within the duration of the PDSCH transmission, the UE may expect to receive a DMRS pattern with low time density, assuming that the pattern has been configured as semi-persistent.

[0174] Table 1

[0175]

[0176] Table 1. Example of IE DMRS-DownlinkConfigH for configuring DMRS with high time density

[0177] In another embodiment, the BS may configure the CSI-RS time density for the UE via RRC configuration. Table 2 is an example of modifying the IECSI-RS-ResourceMapping to configure the time density of the CSI-RS. In this example, densityH and densityL represent predefined time density values ​​(in addition to the values ​​0.5, 1, and 3 already defined in this IE). In one example, densityH and densityL may represent predefined constants. In another example, densityH and densityL may represent predefined sets, each of which includes all possible values ​​of densityH and densityL, respectively. In another example, each predefined set may be defined by a 3-tuple (start_val, diff_val, end_val), where start_val and end_val represent the lower and upper limits on each set, and diff_val represents the difference between consecutive elements in each set. This IE may facilitate switching between CSI-RS modes with high and low time density in consecutive time slots because the information is contained in the IE NZP-CSI-RS-Resource. The NZP-CSI-RS-Resource IE may configure a CSI-RS pattern as periodic, semi-persistent, or aperiodic. For example, if a CSI-RS pattern with high time density is configured as periodic, the UE may expect to receive the pattern in certain time slots. In other time slots, the UE may expect to receive a CSI-RS pattern with low time density, assuming that the pattern has been configured as semi-persistent. For example, in the time domain, a "time resource" may correspond to an OFDM symbol, a time slot, a subframe, a radio frame, etc.

[0178] Table 2

[0179]

[0180] Table 2. Example of IE CSI-RS-ResourceMapping modification for configuring CSI-RS temporal density

[0181] In another example, if densityH and densityL represent predefined sets, each element in each set can also be configured as periodic, semi-persistent, or aperiodic. 1 ,H 2 ,H 3 ,H 4} has been configured as semi-persistent, each element in densityH can also be configured to have a periodicity of 4 time slots. Then, the UE can expect to receive a signal with density H in time slot 1. 1 The CSI-RS pattern with density H is received in time slot 2. 2 The CSI-RS pattern with density H is received in time slot 3. 3 The CSI-RS pattern with density H is received in time slot 4. 4 The CSI-RS pattern with density H is received in time slot 5. 1 In another example, if densityH={H 1 ,H 2 ,H 3 ,H 4}, the CSI-RS mode has been configured as semi-persistent, then the element H in densityH 2 It can also be configured as semi-persistent. Then, the UE can expect to receive a signal with density H in the subsequent time slot. 2 CSI-RS pattern.

[0182] As another example, the BS may configure the UE to (at least temporarily) disable DMRS in subsequent time slots. Table 3 is an example of modifying the IE DMRS-DownlinkConfig to disable DMRS in subsequent time slots. In another example, a new DCI format may be defined to support DMRS time disabling. For this DCI format, the CRC may be scrambled by the C-RNTI of the UE. The DCI format may consist of a DMRS time disabling indication for the UE.

[0183] Table 3

[0184]

[0185] Table 3. Example of IE DMRS-DownlinkConfig modification to support DMRS time disabling

[0186] In one embodiment, a new MAC CE may be defined for UE assistance information reporting. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and may consist of the following fields:

[0187] - Block Error Rate: This field indicates the observed block error rate of the UE, for example, the block error rate has been calculated over the last 1000 received transport blocks.

[0188] -Throughput: This field indicates the observed throughput of the UE, e.g. the throughput (in Mbit / s) that has been calculated over the last 1000 received transport blocks.

[0189] -UE Speed: This field indicates the UE's measurement of its speed (in meters per second).

[0190] -UE acceleration: This field indicates the UE's measurement of its acceleration in meters / (seconds*seconds).

[0191] - Estimated Coherence Time: This field indicates the UE's estimate of the DL channel coherence time (in milliseconds).

[0192] - Estimated Coherence Bandwidth: This field indicates the UE's estimate of the DL channel coherence bandwidth (in kilohertz).

[0193] -IR: This field indicates the presence of an octet containing the Recommended DMRS Time Density field. If the IR field is set to 1, then the octet containing the Recommended DMRS Time Density field is present. If the IR field is set to 0, then the octet containing the Recommended DMRS Time Density field is not present.

[0194] - Recommended DMRS Time Density: This field indicates the recommended DMRS time density of the UE, for example, an index of a DMRS time density value table.

[0195] ■ If the recommended DMRS time density is equal to 0, this message can be used as a recommendation for DMRS time disabling.

[0196] Fig.11 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown. Fig.11 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0197] Fig.11 An example of a new MAC CE 1100 for UE assistance information reporting is shown, where the block error rate 1101, UE throughput field 1102, UE speed field 1103, and UE acceleration field 1104 each have a length of 8 bits. An Indicator of Recommendation (IR) bit 1105 is included in the last byte together with a recommended DMRS time density field 1106 (which has a length of 7 bits).

[0198] In one embodiment, a new MAC CE may be defined for RS time high density backoff request. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and may consist of the following fields:

[0199] DMRS Time High Density Fallback: This field indicates that the UE is requesting the BS to fall back to the DMRS time high density mode, such as the index of this mode in the DMRS time density value table.

[0200] Fig.12 An example of a new MAC CE for RS time high density backoff request according to various embodiments of the present disclosure is shown. Fig.12 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0201] Fig.12 An example of a new MAC CE for RS Time High Density Backoff Request 1200 is shown, where the DMRS Time High Density Backoff field 1201 has a length of 8 bits.

[0202] Fig.13 An example of an AI / ML model architecture supporting RS time-density adaptation according to various embodiments of the present disclosure is shown. Fig.13 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0203] Fig.13 An example of an AI / ML model architecture 1300 that can support RS time density adaptation is shown. In this example, a DMRS high density pattern 1301 in one time slot (e.g., time T) can be combined with a DMRS low density pattern 1302, 1303 in subsequent time slots (e.g., time T+1, time T+2) and used (based on a channel estimate 1304) to infer channel estimates 1305, 1306 in those subsequent time slots. A CRNN can be used in this example, where the hidden states 1307, 1308, 1309 in each time slot can be a function of the previously received DMRS pattern. In each time slot, a complete two-dimensional channel over all subcarriers and OFDM symbols is provided as input to the CRNN, where all REs that do not contain DMRS are padded with zeros. Examples of CRNNs include convolutional LSTMs [4] and convolutional GRUs [5].

[0204] Examples of inputs to AI / ML models that can support RS time-density adaptation include:

[0205] - A complete two-dimensional receive channel over all subcarriers and OFDM symbols of a time slot.

[0206] All REs not containing RS are filled with zeros.

[0207] - The ground truth of the full 2D channel over all subcarriers and OFDM symbols of a slot.

[0208] This can be the training labels.

[0209] -Auxiliary information:

[0210] ◆UE speed.

[0211] ◆UE track.

[0212] ◆UE location.

[0213] ◆Relevant time

[0214] ◆Coherence bandwidth

[0215] Delayed expansion

[0216] ◆Doppler spread

[0217] ◆Block error rate.

[0218] ◆Throughput.

[0219] Examples of outputs from AI / ML models that can support RS time-density adaptation include:

[0220] - Complete 2D estimated channel over all subcarriers and OFDM symbols of a slot.

[0221] - The value of the (inferred) RS temporal density.

[0222] As described above, assuming that the receiver can estimate the channel in at least one other SB and / or BWP, RS signaling can be disabled (at least temporarily) in certain SBs and / or BWPs if the channel is static in frequency. The present disclosure describes a framework for supporting AI / ML techniques for RS frequency density adaptation based on statistics of an underlying randomly varying wireless channel. The corresponding signaling details are discussed in the present disclosure.

[0223] The following embodiments address the problem that RS frequency density adaptation currently depends on network implementation. The present disclosure provides a method that a network can use to configure the frequency density of RS patterns using an AI / ML-based solution. The present disclosure also provides a framework for self-adapting the frequency density of RS patterns based on UE inference and information. Details about supporting AI / ML techniques for RS frequency density adaptation are disclosed, including information elements to be exchanged between a transmitter and a receiver.

[0224] Fig.14A high-level flow chart illustrating an example of BS operations for supporting RS frequency density adaptation according to various embodiments of the present disclosure is shown. Fig.14 The embodiment is for illustration only. Other embodiments of process 1400 may be used without departing from the scope of the present disclosure.

[0225] Fig.14 is an example of a method 1400 for supporting the operation of AI / ML technology for RS frequency density adaptation at a BS. In operation 1401, the BS receives UE capability information from the UE, including support for ML methods for RS frequency density adaptation. In operation 1402, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 1403, the BS sends an RS frequency high density configuration message to the UE. In one example, the message may correspond to an SBA. In operation 1404, the BS receives auxiliary information from the UE, wherein the auxiliary information may include a recommended RS frequency density for a second SBB. In operation 1405, the BS sends an RS frequency low density configuration message to the UE. In one example, the message may correspond to an SBB.

[0226] Fig.15 A high-level flow chart illustrating an example of UE operation for supporting RS frequency density adaptation according to various embodiments of the present disclosure is shown. Fig.15 The embodiment is for illustration only. Other embodiments of process 1500 may be used without departing from the scope of the present disclosure.

[0227] Fig.15 is an example of a method 1500 for supporting the operation of AI / ML techniques for RS frequency density adaptation at a UE. In operation 1501, the UE reports capability information to the BS, including supporting ML methods for RS frequency density adaptation. In operation 1502, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 1503, the UE receives an RS frequency high density configuration message from the BS. In one example, the message may correspond to one SBA. In operation 1504, the UE sends auxiliary information to the BS; the auxiliary information may include a recommended RS frequency density for a second SBB. In operation 1505, the UE receives an RS frequency low density configuration message from the BS. In one example, the message may correspond to an SBB.

[0228] Fig.16A high-level flow chart illustrating an example of BS operations for supporting fallback to RS frequency high density configuration according to various embodiments of the present disclosure is shown. Fig.16 The embodiment is for illustration only. Other embodiments of process 1600 may be used without departing from the scope of the present disclosure.

[0229] Fig.16 is an example of a method 1600 for supporting a request to fall back from an RS frequency low density mode to an RS frequency high density mode at a BS. In operation 1601, the BS receives UE capability information from a UE, including support for an ML method for RS frequency density adaptation. In operation 1602, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 1603, the BS sends an RS frequency high density configuration message to the UE. In one example, the message may correspond to an SBA. In operation 1604, the BS receives auxiliary information from the UE; the auxiliary information may include a recommended RS frequency density of a second SBB. In operation 1605, the BS sends an RS frequency low density configuration message to the UE. In one example, the message may correspond to an SBB. In operation 1606, the BS receives a request to fall back to an RS frequency high density mode from the UE. In operation 1607, the BS sends an RS frequency high density configuration message to the UE. In one example, the message may correspond to SB B.

[0230] Fig.17 A high-level flow chart illustrating an example of UE operation for supporting fallback to RS frequency high density configuration according to various embodiments of the present disclosure is shown. Fig.17 The embodiment is for illustration only. Other embodiments of process 1700 may be used without departing from the scope of the present disclosure.

[0231] Fig.17is an example of a method 1700 for supporting a request to fall back from an RS frequency low density mode to an RS frequency high density mode at a UE. In operation 1701, the UE reports capability information to the BS, including support for an ML method for RS frequency density adaptation. In operation 1702, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 1703, the UE receives an RS frequency high density configuration message from the BS. In one example, the message may correspond to an SBA. In operation 1704, the UE sends auxiliary information to the BS; the auxiliary information may include a recommended RS frequency density for a second SNB. In operation 1705, the UE receives an RS frequency low density configuration message from the BS. In one example, the message may correspond to an SBB. In operation 1706, the UE sends a request to fall back to an RS frequency high density mode to the BS. In operation 1707, the UE receives an RS frequency high density configuration message from the BS. In one example, the message may correspond to SB B.

[0232] Fig.18 A high-level flow chart illustrating an example of BS operation for supporting RS frequency disabling according to various embodiments of the present disclosure is shown. Fig.18 The embodiment is for illustration only. Other embodiments of process 1800 may be used without departing from the scope of the present disclosure.

[0233] Fig.18 is an example of method 1800 for supporting the operation of AI / ML technology for RS frequency disabling at a BS. In operation 1801, the BS receives UE capability information from the UE, including support for ML methods for RS frequency density adaptation. In operation 1802, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 1803, the BS sends an RS frequency density configuration message to the UE. In one example, the message may correspond to an SBA. In operation 1804, the BS receives auxiliary information from the UE; the auxiliary information may include a recommendation for RS frequency disabling. In operation 1805, the BS sends an RS frequency disabling configuration message to the UE. In one example, the message may correspond to an SBA.

[0234] Fig.19 A high-level flow chart illustrating an example of UE operation for supporting RS frequency disabling according to various embodiments of the present disclosure is shown. Fig.19The embodiment is for illustration only. Other embodiments of process 1900 may be used without departing from the scope of the present disclosure.

[0235] Fig.19 is an example of method 1900 for supporting the operation of AI / ML techniques for RS frequency disabling at a UE. In operation 1901, the UE reports capability information to the BS, including support for ML methods for RS frequency density adaptation. In operation 1902, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 1903, the UE receives an RS frequency density configuration message from the BS. In one example, the message may correspond to an SBA. In operation 1904, the UE sends auxiliary information to the BS; the auxiliary information may include a recommendation for RS frequency disabling. In operation 1905, the UE receives an RS frequency disabling configuration message from the BS. In one example, the message may correspond to an SBA.

[0236] Fig. 20 A high-level flow chart illustrating an example of BS operation for supporting RS frequency-based BWP switching according to various embodiments of the present disclosure is shown. Fig. 20 The embodiment is for illustration only. Other embodiments of process 2000 may be used without departing from the scope of the present disclosure.

[0237] Fig. 20The present invention is an example of a method 2000 for supporting an AI / ML technique for RS frequency-based BWP switching at a BS. In operation 2001, the BS receives UE capability information from a UE, including support for an ML method for RS frequency density adaptation. In operation 2002, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling an ML method for RS frequency density adaptation, an ML model to be used, trained model parameters, and / or whether to use model parameter updates reported by the UE. In operation 2003, the BS sends an RS frequency high density configuration message for active BWP A to the UE. In an example, the BS may configure the UE to receive high-density CSI-RS in BWPA. In operation 2004, the BS receives a CSI report of active BWP A from the UE. In operation 2005, the BS sends an RS frequency low density configuration message of an inactive BWP B to the UE. In an example, the BS may configure the UE to receive low-density CSI-RS in BWP B. In operation 2006, the BS receives a CSI report of inactive BWP B from the UE. At operation 2007, the BS receives assistance information from the UE; the assistance information may include a recommendation for BWP switching. At operation 2008, the BS sends a BWP switching command message to the UE. In one example, the command may correspond to a BWP switching from BWP A to BWP B.

[0238] Fig.21 A high-level flow chart illustrating an example of UE operation for supporting RS frequency-based BWP switching according to various embodiments of the present disclosure. Fig.21 The embodiment is for illustration only. Other embodiments of process 2100 may be used without departing from the scope of the present disclosure.

[0239] Fig.21is an example of a method 2100 for supporting AI / ML techniques for RS frequency-based BWP switching at a UE. In operation 2101, the UE reports capability information to the BS, including support for ML methods for RS frequency density adaptation. In operation 2102, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS frequency density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 2103, the UE receives an RS frequency high density configuration message for active BWP A from the BS. In operation 2104, the UE sends a CSI report for active BWP A to the BS. In operation 2105, the UE receives an RS frequency low density configuration message for inactive BWP B from the BS. In operation 2106, the UE sends a CSI report for inactive BWP B to the BS. In operation 2107, the UE sends auxiliary information to the BS; the auxiliary information may include a recommendation for BWP switching. In operation 2108, the UE receives a BWP switching command message from the BS. In one example, the command may correspond to a BWP switch from BWP A to BWP B.

[0240] Fig. 22 An example of RS frequency density adaptation according to various embodiments of the present disclosure is shown. Fig. 22 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0241] In 5G NR, a time slot consists of 14 OFDM symbols. For example, in the frequency domain, a "frequency resource" may correspond to a subcarrier, a resource block, a subband, etc. Example 2200 of RS frequency density adaptation Fig. 22 , where rectangles represent time slots and each time slot contains multiple SBs. Within a time slot and SB, a certain number of subcarriers may contain DMRS; the ratio of this number to the total number of subcarriers in the SB may represent the DMRS frequency density. In this example, for a given SB B, a time slot 2202 with a high DMRS frequency density may be configured with a specific period 2204; other time slots 2203 within the duration 2205 of the PDSCH transmission may be configured with a low DMRS frequency density. For another SB A, each time slot 2201 has a high DMRS frequency density.

[0242] In one embodiment, the BS may configure the DMRS time density for the UE via PDCCH signaling. In another example, the BS may configure the DMRS time density for the UE via RRC configuration. Table 4 is an example of defining an information element (IE) DMRS-DownlinkConfigH to configure a DMRS with high time density. In this example, for each SB, highDensity determines whether to utilize high frequency density. If high frequency density is used for a given SB, the corresponding DMRS mode may be configured as periodic, semi-persistent, or aperiodic. Different high frequency densities may be defined for each SB via timeFreqAllocation, where DMRS-Sym represents a tuple of (OFDM symbol index, frequency density) values; one tuple is specified for each OFDM symbol in a time slot. Similarly, IE DMRS-DownlinkConfigL may be defined to configure a DMRS with low frequency density. These IEs may facilitate switching between a DMRS mode with high time density and a DMRS mode with low time density in consecutive time slots. For example, for a given SB, if a DMRS pattern with high frequency density is configured as periodic, the UE can expect to receive the pattern in certain time slots. In other time slots of a given SB within the duration of a PDSCH transmission, the UE can expect to receive a DMRS pattern with low frequency density, assuming that the pattern has been configured as semi-persistent.

[0243] Table 4

[0244]

[0245]

[0246] Table 4. Example of IE DMRS-DownlinkConfigH for configuring DMRS with high frequency density

[0247] In another embodiment, the BS may configure the CSI-RS frequency density for the UE via RRC configuration. Table 5 is an example of modifying the IECSI-RS-ResourceMapping to configure the frequency density of the CSI-RS. In this example, densityH and densityL represent predefined frequency density values ​​(in addition to the values ​​0.5, 1, and 3 already defined in this IE); a frequency density value is specified for each SB. This IE can facilitate switching between a CSI-RS pattern with high frequency density and a CSI-RS pattern with low frequency density for a given SB in consecutive time slots because this information is contained in the IE NZP-CSI-RS-Resource. The NZP-CSI-RS-Resource IE can configure the CSI-RS pattern as periodic, semi-persistent, or aperiodic. For example, for a given SB, if a CSI-RS pattern with high frequency density is configured as periodic, the UE can expect to receive the pattern in certain time slots. In other time slots, for a given SB, the UE can expect to receive a CSI-RS pattern with low frequency density, assuming that the pattern has been configured as semi-persistent. For example, in the frequency domain, a "frequency resource" may correspond to a subcarrier, a resource block, a subband, etc.

[0248] Table 5

[0249]

[0250] Table 5. Example of IE CSI-RS-ResourceMapping modification for configuring CSI-RS frequency density

[0251] In another embodiment, the BS may configure DMRS frequency density hopping for the UE via RRC configuration. Table 6 is an example of defining IEDMS-DownlinkConfigH and DMRS-DownlinkConfigL to configure DMRS frequency density hopping. For DMRS-DownlinkConfigH, frequencyHopping (if present) determines whether a specific high-density DMRS pattern hops within a time slot or between time slots; frequencyHoppingOffset (if present) determines the frequency hopping pattern of the high-density DMRS pattern across available SBs. For DMRS-DownlinkConfigL, useDensityHHoppingOffset and useDensityHHoppingOffset (if present) are set to the same values ​​as frequencyHoppingOffset and frequencyHoppingOffset in DMRS-DownlinkConfigH, respectively. If the frequency hopping mode of the high-density DMRS mode is enabled, the UE can use the frequency hopping mode to determine the DMRS density within a specific SB in a specific time slot.

[0252] Table 6

[0253]

[0254]

[0255] Table 6. Example of IE DMRS-DownlinkConfig modification for configuring DMRS frequency density hopping

[0256] In another embodiment, the BS may configure CSI-RS frequency density hopping for the UE via RRC configuration. Table 7 is an example of defining IE CSI-RS-ResourceMappingH and CSI-RS-ResourceMappingL to configure CSI-RS frequency density hopping. For CSI-RS-ResourceMappingH, frequencyHopping (if present) determines whether a specific high-density CSI-RS pattern hops within a time slot or between time slots; frequencyHoppingOffset (if present) determines the frequency hopping pattern of the high-density CSI-RS pattern across available SBs. For CSI-RS-ResourceMappingL, useDensityHHopping and useDensityHHoppingOffset (if present) are set to the same values ​​as frequencyHopping and frequencyHoppingOffset in CSI-RS-ResourceMappingH, respectively. If the frequency hopping mode of the high-density CSI-RS pattern is enabled, the UE can use the frequency hopping mode to determine the CSI-RS density within a specific SB in a specific time slot.

[0257] Table 7

[0258]

[0259] Table 7. Example of IE CSI-RS-ResourceMapping modification for configuring CSI-RS frequency density hopping

[0260] As another example, the BS may configure the UE to (at least temporarily) disable DMRS in certain SBs for subsequent time slots. Table 8 is an example of modifying the IE DMRS-DownlinkConfig to disable DMRS in certain SBs for subsequent time slots. In another example, a new DCI format may be defined to support DMRS frequency disabling. For this DCI format, the CRC may be scrambled by the C-RNTI of the UE. The DCI format may consist of a set of tuples (DMRS frequency disabling indication, SB index) for the UE, wherein one tuple is included for each SB.

[0261] Table 8

[0262]

[0263] Table 8. Example of IE DMRS-DownlinkConfig modification to support DMRS frequency disabling

[0264] In one embodiment, a new MAC CE may be defined for UE assistance information reporting. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and may consist of the following fields:

[0265] - Block Error Rate: This field indicates the observed block error rate of the UE, for example, the block error rate has been calculated over the last 1000 received transport blocks.

[0266] -Throughput: This field indicates the observed throughput of the UE, e.g., the throughput calculated over the last 1000 received transport blocks (in Mbit / s).

[0267] -UE Speed: This field indicates the UE's measurement of its speed (in meters per second).

[0268] -UE acceleration: This field indicates the UE's measurement of its acceleration in meters / (seconds*seconds).

[0269] - IR1: This field indicates the presence of an octet containing the Recommended RS Frequency Density field. If the IR1 field is set to 1, then the octet containing the Recommended RS Frequency Density field is present. If the IR1 field is set to 0, then the octet containing the Recommended RS Frequency Density field is not present.

[0270] - Recommended RS Frequency Density: This field indicates the recommended RS frequency density of the UE for each SB, such as an index of an RS frequency density value table.

[0271] ◆ If for a given SB, the recommended RS frequency density is equal to 0, then this message can serve as a recommendation for RS frequency disabling for that SB.

[0272] - IR2: This field indicates the presence of an octet containing the Recommended BWP field. If the IR2 field is set to 1, then the octet containing the Recommended BWP field is present. If the IR2 field is set to 0, then the octet containing the Recommended BWP field is not present.

[0273] - Recommended BWP: This field indicates the recommended active BWP for the UE, eg, an index to a BWP table.

[0274] Fig.23 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown. Fig.23 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0275] Fig.23 An example of a new MAC CE 2300 for UE assistance information reporting is shown, where the block error rate field 2301, the UE throughput field 2302, the UE speed field 2303, and the UE acceleration field 2304 each have a length of 8 bits. The recommended RS frequency density field 2306 has a length of 15 bits for two SBs. The recommended BWP field 2308 has a length of 7 bits.

[0276] In one embodiment, a new MAC CE may be defined for RS frequency high density fallback request. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and consist of the following fields (one entry per SB):

[0277] -DMRS Frequency High Density Fallback: This field indicates that the UE is requesting the BS to fall back to the DMRS frequency high density mode for a given SB, such as the index of this mode in the DMRS frequency density value table.

[0278] Fig.24 An example of a new MAC CE for RS frequency high density backoff request according to various embodiments of the present disclosure is shown. Fig.24 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0279] Fig.24 An example of a new MAC CE 2400 for RS frequency high density backoff request is shown, where the DMRS frequency high density backoff field has a length of 24 bits for three SBs.

[0280] Fig.25 An example of an AI / ML model architecture supporting RS frequency density adaptation according to various embodiments of the present disclosure is shown. Fig.25 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0281] Fig.25An example of an AI / ML model architecture 2500 that can support RS frequency density adaptation is shown. In this example, a DMRS low density pattern 2501 in one SB A is provided as an input to a CNN 2502. A DMRS high density pattern 2503 in another SB B (which serves as auxiliary information for the model [6]) is provided as an input to another CNN 2504. The outputs of the two CNNs 2502, 2504 are then convolved (block 2505), and the output of the convolution is provided as input to another CNN 2506, which infers a channel estimate 2507 in SB A. For both SB A and SB B, all REs that do not contain DMRS are padded with zeros.

[0282] Examples of inputs to AI / ML models that can support RS frequency density adaptation include:

[0283] - A complete two-dimensional receive channel over all subcarriers and OFDM symbols of one time slot and one SBA.

[0284] All REs not containing RS are filled with zeros.

[0285] - The ground truth of the complete 2D channel over all subcarriers and OFDM symbols of one time slot and one SBA.

[0286] This can be the training labels.

[0287] -Auxiliary information:

[0288] A complete two-dimensional receive channel on all subcarriers and OFDM symbols of one time slot and another SBB.

[0289] All REs not containing RS are filled with zeros

[0290] ◆UE speed.

[0291] ◆UE track.

[0292] ◆UE location.

[0293] ◆Relevant time

[0294] ◆Coherence bandwidth

[0295] Delayed expansion

[0296] ◆Doppler spread

[0297] ◆Block error rate.

[0298] ◆Throughput.

[0299] Examples of outputs from AI / ML models that can support RS frequency density adaptation include:

[0300] - Complete two-dimensional estimated channel over all subcarriers and OFDM symbols of one time slot and one SBA.

[0301] -The value of the (inferred) RS frequency density of SB A.

[0302] It may be advantageous to vary the spatial density of the RS pattern based on the statistics of the underlying randomly varying wireless channel. For example, if the channel selectivity in space is reduced, reducing the spatial density of the RS pattern may have a negligible effect on the CSI estimation error - while reducing the signaling overhead. As another example, assuming that the receiver can estimate the channel for at least one other antenna port, if the channel is static in space, RS signaling can be disabled for certain antenna ports (at least temporarily).

[0303] 5G NR supports flexibility in selecting RS patterns, where the selection of RS patterns is based on the statistics of the underlying randomly varying radio channel. For example, the parameter dmrs-AdditionalPosition can be used to increase the number of DMRS in a given time slot in high mobility scenarios. As another example, the parameters periodicityAndOffset-p and periodicityAndOffset-sp can be used to change the periodicity (and time slot offset) of the SRS. The details of the algorithm used to select the RS pattern are generally left to the network.

[0304] This disclosure describes a framework for supporting AI / ML techniques for RS spatial density adaptation based on the statistics of the underlying randomly varying wireless channel. The corresponding signaling details are discussed in this disclosure.

[0305] The present disclosure solves the problem that RS spatial density adaptation currently depends on network implementation. The present disclosure provides a method that the network can use to configure the spatial density of RS patterns using AI / ML-based solutions. The present disclosure also provides a framework for adapting the spatial density of RS patterns based on UE inference and information.

[0306] Details are disclosed on supporting AI / ML techniques for RS spatial density adaptation, including information elements to be exchanged between transmitters and receivers.

[0307] Fig.26 A high-level flow chart illustrating an example of BS operations for supporting RS spatial density adaptation according to various embodiments of the present disclosure is shown. Fig.26 The embodiment is for illustration only. Other embodiments of process 2600 may be used without departing from the scope of the present disclosure.

[0308] Fig.26is an example of a method 2600 for supporting the operation of AI / ML technology for RS spatial density adaptation at a BS. In operation 2601, the BS receives UE capability information from the UE, including support for the ML method for RS spatial density adaptation. In operation 2602, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS spatial density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 2603, the BS sends an RS spatial high density configuration message to the UE. In one example, the message may correspond to one antenna port A. At operation 2604, the BS sends auxiliary information to the UE; the auxiliary information may include beamforming information, which will be described below. In operation 2605, the BS receives auxiliary information from the UE; the auxiliary information may include a recommended RS spatial density for the second antenna port B. In operation 2606, the BS sends an RS spatial low density configuration message to the UE. In one example, the message may correspond to antenna port B.

[0309] Fig. 27 A high-level flow chart illustrating an example of UE operation for supporting RS spatial density adaptation according to various embodiments of the present disclosure is shown. Fig. 27 The embodiment is for illustration only. Other embodiments of process 2700 may be used without departing from the scope of the present disclosure.

[0310] Fig. 27 is an example of a method 2700 for supporting the operation of AI / ML technology for RS spatial density adaptation at a UE. In operation 2701, the UE reports capability information to the BS, including support for the ML method for RS spatial density adaptation. In operation 2702, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS spatial density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameters reported by the UE will be updated. In operation 2703, the UE receives an RS spatial high density configuration message from the BS. In one example, the message may correspond to one antenna port A. In operation 2704, the UE receives auxiliary information from the BS; the auxiliary information may include beamforming information, which will be described below. In operation 2705, the UE sends auxiliary information to the BS; the auxiliary information may include a recommended RS spatial density for the second antenna port B. In operation 2706, the UE receives an RS spatial low density configuration message from the BS. In one example, the message may correspond to antenna port B.

[0311] Fig.28A high-level flow chart illustrating an example of BS operations for supporting fallback to RS spatial high density configuration according to various embodiments of the present disclosure is shown. Fig.28 The embodiment is for illustration only. Other embodiments of process 2800 may be used without departing from the scope of the present disclosure.

[0312] Fig.28 is an example of a method 2800 for supporting a request to fall back from an RS spatial low density mode to an RS spatial high density mode at a BS. In operation 2801, the BS receives UE capability information from a UE, including support for an ML method for RS spatial density adaptation. In operation 2802, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS spatial density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 2803, the BS sends an RS spatial high density configuration message to the UE. In one example, the message may correspond to one antenna port A. In operation 2804, the BS sends auxiliary information to the UE; the auxiliary information may include beamforming information, which will be described below. In operation 2805, the BS receives auxiliary information from the UE; the auxiliary information may include a recommended RS spatial density for a second antenna port B. In operation 2806, the BS sends an RS spatial low density configuration message to the UE. In one example, the message may correspond to antenna port B. At operation 2807, the BS receives a request to fall back to the RS spatial high density mode from the UE. At operation 2808, the BS sends an RS spatial high density configuration message to the UE. In one example, the message may correspond to antenna port B.

[0313] Fig.29 A high-level flow chart illustrating an example of UE operation for supporting fallback to RS spatial high density configuration according to various embodiments of the present disclosure is shown. Fig.29 The embodiment is for illustration only. Other embodiments of process 2900 may be used without departing from the scope of the present disclosure.

[0314] Fig.29is an example of a method 2900 for supporting a request to fall back from an RS spatial low density mode to an RS spatial high density mode at a UE. In operation 2901, the UE reports capability information to the BS, including support for an ML method for RS spatial density adaptation. In operation 2902, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS spatial density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 2903, the UE receives an RS spatial high density configuration message from the BS. In one example, the message may correspond to one antenna port A. In operation 2904, the UE receives auxiliary information from the BS; the auxiliary information may include beamforming information, which will be described below. In operation 2905, the UE sends auxiliary information to the BS; the auxiliary information may include a recommended RS spatial density for a second antenna port B. In operation 2906, the UE receives an RS spatial low density configuration message from the BS. In one example, the message may correspond to antenna port B. At operation 2907, the UE sends a request to the BS to fall back to the RS spatial high density mode. At operation 2908, the UE receives an RS spatial high density configuration message from the BS. In one example, the message may correspond to antenna port B.

[0315] Fig.30 A high-level flow chart illustrating an example of BS operations for supporting RS spatial disabling according to various embodiments of the present disclosure is shown. Fig.30 The embodiment is for illustration only. Other embodiments of process 3000 may be used without departing from the scope of the present disclosure.

[0316] Fig.30 is an example of a method 3000 for supporting the operation of AI / ML techniques for RS spatial disabling at a BS. In operation 3001, the BS receives UE capability information from the UE, including support for ML methods for RS spatial density adaptation. In operation 3002, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for RS spatial density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 3003, the BS sends an RS spatial density configuration message to the UE. In one example, the message may correspond to one antenna port A. In operation 3004, the BS sends auxiliary information to the UE; the auxiliary information may include beamforming information. In operation 3005, the BS receives auxiliary information from the UE; the auxiliary information may include a recommendation for RS spatial disabling. At operation 3006, the BS sends an RS spatial disabling configuration message to the UE. In one example, the message may correspond to antenna port A.

[0317] Fig.31 A high-level flow chart illustrating an example of UE operation for supporting RS spatial disabling according to various embodiments of the present disclosure is shown. Fig.31 The embodiment is for illustration only. Other embodiments of process 3100 may be used without departing from the scope of the present disclosure.

[0318] Fig.31 is an example of a method 3100 for supporting the operation of AI / ML techniques for RS spatial disabling at a UE. In operation 3101, the UE reports capability information to the BS, including support for ML methods for RS spatial density adaptation. In operation 3102, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for RS spatial density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 3103, the UE receives an RS spatial density configuration message from the BS. In one example, the message may correspond to one antenna port A. In operation 3104, the UE receives auxiliary information from the BS; the auxiliary information may include beamforming information. In operation 3105, the UE sends auxiliary information to the BS; the auxiliary information may include a recommendation for RS spatial disabling. In operation 3106, the UE receives an RS spatial disabling configuration message from the BS. In one example, the message may correspond to antenna port A.

[0319] Fig.32 A high-level flow chart illustrating an example of BS operations for supporting RS space-based port switching according to various embodiments of the present disclosure is shown. Fig.32 The embodiment is for illustration only. Other embodiments of process 3200 may be used without departing from the scope of the present disclosure.

[0320] Fig.32is an example of a method 3200 for supporting an AI / ML technique for RS space-based port switching at a BS. In operation 3201, the BS receives UE capability information from a UE, including support for an ML method for RS space density adaptation. In operation 3202, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling an ML method for RS space density adaptation, an ML model to be used, trained model parameters, and / or whether to use model parameter updates reported by the UE. In operation 3203, the BS sends a DMRS space high density configuration message for antenna port A to the UE. In operation 3204, the BS sends a CSI-RS space low density configuration message for antenna port B to the UE. In one example, the BS uses antenna port A for PDSCH transmission and does not use antenna port B for the transmission. In operation 3205, the BS sends auxiliary information to the UE; the auxiliary information may include beamforming information. In operation 3206, the BS receives a CSI report for antenna port B from the UE. At operation 3207, the BS receives assistance information from the UE; the assistance information may include a recommendation for antenna port switching. At operation 3208, the BS sends a port switching indication to the UE. In one example, the indication may correspond to a port switching from antenna port A to antenna port B.

[0321] Fig.33 A high-level flow chart illustrating an example of UE operation for supporting RS space-based port switching according to various embodiments of the present disclosure is shown. Fig.33 The embodiment is for illustration only. Other embodiments of process 3300 may be used without departing from the scope of the present disclosure.

[0322] Fig.33is an example of a method 3300 for supporting an AI / ML technique for RS space-based port switching at a UE. In operation 3301, the UE reports capability information to the BS, including support for an ML method for RS space density adaptation. In operation 3302, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling an ML method for RS space density adaptation, an ML model to be used, trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 3303, the UE receives a DMRS space high density configuration message for antenna port A from the BS. In operation 3304, the UE receives a CSI-RS space low density configuration message for antenna port B from the BS. In one example, the BS uses antenna port A for PDSCH transmission and does not use antenna port B for the transmission. In operation 3305, the UE receives auxiliary information from the BS; the auxiliary information may include beamforming information. In operation 3306, the UE sends a CSI report for antenna port B to the BS. In operation 3307, the UE sends auxiliary information to the BS; the auxiliary information may include a recommendation for antenna port switching. At operation 3208, the UE receives a port switching indication from the BS. In one example, the indication may correspond to a port switching from antenna port A to antenna port B.

[0323] Fig.34 An example of RS spatial density adaptation according to various embodiments of the present disclosure is shown. Fig.34 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0324] In 5G NR, a time slot consists of 14 OFDM symbols. An example of RS spatial density adaptation is shown in Fig.34 , where rectangles represent time slots, and multiple antenna ports are used for transmission during each time slot. Within a time slot, for each antenna port, a certain number of REs may contain DMRS; the ratio of the number of REs within a time slot for that antenna port to the total number may represent the DMRS spatial density. In this example, for a given antenna port B, a time slot 3402 with a high DMRS spatial density may be configured with a specific periodicity 3404; other time slots 3403 within the duration 3405 of the PDSCH transmission may be configured with a low DMRS spatial density. For another antenna port A, each time slot 3401 has a high DMRS spatial density.

[0325] In one embodiment, the BS may configure the DMRS spatial density for the UE via PDCCH signaling. In another example, the BS may configure the DMRS spatial density for the UE via RRC configuration. Table 9 is an example of defining IE DMRS-DownlinkConfigH to configure DMRS with high spatial density. In this example, for each antenna port, highDensity determines whether to utilize high spatial density. If high spatial density is used for a given antenna port, the corresponding DMRS pattern can be configured as periodic, semi-persistent, or non-periodic. Different high spatial densities can be defined for each antenna port via timeFreqAllocation, where DMRS-Sym represents a tuple of (OFDM symbol index, frequency density) values; one tuple is specified for each OFDM symbol in a time slot. Similarly, IE DMRS-DownlinkConfigL can be defined to configure DMRS with low spatial density. These IEs can facilitate switching between DMRS modes with high spatial density and low spatial density for a given antenna port in consecutive time slots. For example, for a given antenna port, if a DMRS pattern with high spatial density is configured as periodic, the UE can expect to receive the pattern in certain time slots. In other time slots for the given antenna port within the duration of the PDSCH transmission, the UE can expect to receive a DMRS pattern with low spatial density, assuming that the pattern has been configured as semi-persistent.

[0326] Table 9

[0327]

[0328]

[0329] Table 9. Example of IE DMRS-DownlinkConfigH for configuring DMRS with high spatial density

[0330] In another embodiment, the BS may configure the CSI-RS spatial density for the UE via RRC configuration. Table 10 is an example of modifying the IECSI-RS-ResourceMapping to configure the spatial density of the CSI-RS. In this example, densityH and densityL represent predefined spatial density values ​​(in addition to the values ​​0.5, 1, and 3 already defined in this IE); a spatial density value is specified for each antenna port. This IE can facilitate switching between CSI-RS patterns with high spatial density and low spatial density for a given antenna port in consecutive time slots because this information is contained in the IE NZP-CSI-RS-Resource. The NZP-CSI-RS-Resource IE can configure the CSI-RS pattern as periodic, semi-persistent, or aperiodic. For example, for a given antenna port, if a CSI-RS pattern with high spatial density is configured as periodic, the UE may expect to receive the pattern in certain time slots. In other time slots, for a given antenna port, the UE may expect to receive a CSI-RS pattern with low spatial density, assuming that the pattern has been configured as semi-persistent.

[0331] Table 10

[0332]

[0333] Table 10. Example of IE CSI-RS-ResourceMapping modification for configuring CSI-RS spatial density

[0334] In another embodiment, the BS may configure DMRS spatial density hopping for the UE via RRC configuration. Table 11 is an example of defining IE DMRS-DownlinkConfigH and DMRS-DownlinkConfigL to configure DMRS spatial density hopping. For DMRS-DownlinkConfigH, portHopping (if present) determines whether a specific high-density DMRS pattern hops within a time slot or between time slots; portHoppingOffset (if present) determines the frequency hopping pattern of the high-density DMRS pattern across available antenna ports. For DMRS-DownlinkConfigL, useDensityHHopping and useDensityHHoppingOffset (if present) are set to the same values ​​as portHopping and portHoppingOffset in DMRS-DownlinkConfigH, respectively. If the frequency hopping mode of the high-density DMRS mode is enabled, the UE can use the frequency hopping mode to determine the DMRS density of a specific antenna port for a specific time slot.

[0335] Table 11

[0336]

[0337]

[0338] Table 11. Example of IE DMRS-DownlinkConfig modification for configuring DMRS spatial density hopping

[0339] In another embodiment, the BS may configure CSI-RS spatial density hopping for the UE via RRC configuration. Table 12 is an example of defining IE CSI-RS-ResourceMappingH and CSI-RS-ResourceMappingL to configure CSI-RS spatial density hopping. For CSI-RS-ResourceMappingH, portHopping (if present) determines whether a specific high-density CSI-RS pattern hops within a time slot or between time slots; portHoppingOffset (if present) determines the frequency hopping pattern of the high-density CSI-RS pattern across available antenna ports. For CSI-RS-ResourceMappingL, useDensityHHopping and useDensityHHoppingOffset (if present) are set to the same values ​​as portHopping and portHoppingOffset in CSI-RS-ResourceMappingH, respectively. If the frequency hopping mode of the high-density CSI-RS pattern is enabled, the UE can use the frequency hopping mode to determine the CSI-RS density of a specific antenna port for a specific time slot.

[0340] Table 12

[0341]

[0342] Table 12. Example of IE CSI-RS-ResourceMapping modification for configuring CSI-RS spatial density hopping

[0343] As another example, the BS may configure the UE to (at least temporarily) disable DMRS for certain antenna ports for subsequent time slots. Table 13 is an example of modifying the IE DMRS-DownlinkConfig to disable DMRS for certain antenna ports for subsequent time slots. In another example, a new DCI format may be defined to support DMRS spatial disabling. For this DCI format, the CRC may be scrambled by the C-RNTI of the UE. The DCI format may consist of a set of tuples (DMRS spatial disabling indication, antenna port index) for the UE, including one tuple for each antenna port.

[0344] Table 13

[0345]

[0346] Table 13. Example of IE DMRS-DownlinkConfig modification to support DMRS space disabling

[0347] In one embodiment, a new MAC CE may be defined for BS assistance information reporting. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and may consist of the following fields:

[0348] -Beamforming Information: This field indicates information that the UE can use to infer the channel for a specific antenna port, including:

[0349] ◆PMI used by the BS for transmission to the UE.

[0350] ◆BS-side mapping from (logical) antenna ports to physical antenna elements.

[0351] ◆ BS-side spacing between physical antenna elements (e.g., in centimeters).

[0352] Fig.35 An example of a new MAC CE for BS assistance information reporting according to various embodiments of the present disclosure is shown. Fig.35 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0353] Fig.35 An example of a new MAC CE 3500 for BS assistance information reporting is shown, where the beamforming information field 3501 has a length of 24 bits.

[0354] UE assistance information reporting provides several advantages over relying on existing signaling. For example, the BS can use SRS to estimate the UL (and DL, depending on reciprocity) channel from the UE. However, the minimum period of SRS is 2 milliseconds (ms); in contrast, the interval between consecutive DMRS can be configured to be less than 1 ms. Therefore, the UE can use the received DMRS to perform more fine-grained measurements of the DL channel than the BS uses the received SRS to measure the UL channel.

[0355] As another example, the UE may report local information that may not be available to the BS. The UE may use a camera therein to detect an oncoming vehicle that will cross its line of sight with the BS within T seconds. The UE may then report this information to the BS and make a preemptive recommendation for a transmission mode switch (e.g., switch to a relatively robust mode such as transmit diversity) within T seconds.

[0356] In one embodiment, a new MAC CE may be defined for UE assistance information reporting. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and may consist of the following fields:

[0357] - Block Error Rate: This field indicates the observed block error rate of the UE, for example, the block error rate has been calculated over the last 1000 received transport blocks.

[0358] -Throughput: This field indicates the observed throughput of the UE, e.g., the throughput calculated over the last 1000 received transport blocks (in Mbit / s).

[0359] -UE Speed: This field indicates the UE's measurement of its speed (in meters per second).

[0360] -UE acceleration: This field indicates the UE's measurement of its acceleration in meters / (seconds*seconds).

[0361] - IR1: This field indicates the presence of an octet containing the Recommended RS Spatial Density field. If the IR1 field is set to 1, then there is an octet containing the Recommended RS Spatial Density field. If the IR1 field is set to 0, then there is no octet containing the Recommended RS Spatial Density field.

[0362] - Recommended RS spatial density: This field indicates the recommended RS spatial density of the UE for each antenna port, for example, an index to an RS spatial density value table.

[0363] ◆ If for a given antenna port the recommended RS spatial density is equal to 0, then this message may serve as a recommendation for RS spatial disabling for that antenna port.

[0364] ◆This field may also include the reason for the recommended RS spatial density, for example, an index into a table of codes. For example, one code may correspond to the reason "oncoming vehicle" while another code may correspond to the reason "report traffic accident at mile marker 30".

[0365] -IR2: This field indicates the presence of an octet containing the Recommended Port field. If the IR2 field is set to 1, then the octet containing the Recommended Port field is present. If the IR2 field is set to 0, then the octet containing the Recommended Port field is not present.

[0366] - Recommended Port: This field indicates the recommended BS antenna port for the UE, for example, an index (or multiple indexes) of an antenna port table.

[0367] ◆This field may also include the reason for the recommended port, such as an index into a table of codes. For example, one code may correspond to the reason "oncoming vehicle" while another code may correspond to the reason "reporting traffic accident at mile marker 30".

[0368] The BS may configure the UE with an RS spatial density threshold T, where the UE may compare the recommended RS spatial density for each port of the UE with T. The UE may be configured to apply the recommended RS spatial density for each port of the UE on the UL transmission of the UE if the recommended RS spatial density for each port of the UE exceeds T, without waiting for a configuration message from the BS.

[0369] Fig.36 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown. Fig.36 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0370] Fig.36 An example of a new MAC CE 3600 for UE assistance information reporting is shown, where the block error rate field 3601, the UE throughput field 3602, the UE speed field 3603, and the UE acceleration field 3604 each have a length of 8 bits. The recommended RS spatial density field 3606 (depending on the IR1 field 3605) has a length of 15 bits for two antenna ports. The recommended port field 3608 (depending on the IR2 field 3607) has a length of 7 bits.

[0371] In one embodiment, a new MAC CE may be defined for RS spatial high density fallback request. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and consist of the following fields (one entry per antenna port):

[0372] -DMRS spatial high density fallback: This field indicates that the UE is requesting the BS to fall back to the DMRS spatial high density mode for a given antenna port, for example, the index of this mode in the DMRS spatial density value table.

[0373] Fig.37 An example of a new MAC CE for RS spatial high-density backoff request according to various embodiments of the present disclosure is shown. Fig.37 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0374] Fig.37 An example of a new MAC CE 3700 for RS spatial high density backoff request is shown, where the DMRS spatial high density backoff field 3701 has a length of 24 bits for three antenna ports.

[0375] Fig.38 An example of an AI / ML model architecture supporting RS spatial density adaptation according to various embodiments of the present disclosure is shown. Fig.38 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0376] Fig.38 An example of an AI / ML model architecture 3800 that can support RS spatial density adaptation is shown. In this example, a DMRS low density pattern 3801 for one antenna port A is provided as an input to a CNN 3802. A DMRS high density pattern 3803 for another antenna port B (used as auxiliary information for the model [6]) is provided as an input to another CNN 3804. The outputs of the two CNNs 3802, 3804 are then convolved (block 3805), and the output of the convolution is provided as input to another CNN 3806, which infers a channel estimate 3807 for antenna port A. For antenna ports A and B, all REs that do not contain DMRS are filled with zeros.

[0377] Examples of inputs to AI / ML models that can support RS spatial density adaptation include:

[0378] - A complete two-dimensional receive channel over all subcarriers and OFDM symbols of one time slot and one antenna port A.

[0379] ◆All REs not containing RS are filled with zeros.

[0380] - The ground truth of the complete 2D channel over all subcarriers and OFDM symbols for one time slot and one antenna port A.

[0381] ◆This can be a training label.

[0382] ◆Auxiliary information.

[0383] ◆ A complete two-dimensional receive channel over all subcarriers and OFDM symbols for one time slot and another antenna port B.

[0384] ■All REs not containing RS are filled with zeros.

[0385] ◆UE speed.

[0386] ◆UE track.

[0387] ◆UE location.

[0388] ◆Relevant time

[0389] ◆Coherence bandwidth

[0390] Delayed expansion

[0391] ◆Doppler spread

[0392] Block Error Rate

[0393] ◆Throughput.

[0394] ◆PMI used by the BS for transmission to the UE.

[0395] ◆BS-side mapping from (logical) antenna ports to physical antenna elements.

[0396] ◆BS-side spacing between physical antenna elements.

[0397] Examples of outputs from AI / ML models that can support RS spatial density adaptation include:

[0398] - Complete two-dimensional estimated channel over all subcarriers and OFDM symbols for one time slot and one antenna port A.

[0399] -The value of the (inferred) RS spatial density for antenna port A.

[0400] When utilizing multi-TRP transmissions, it may be advantageous to vary the density of RS patterns based on the statistics of the underlying randomly varying wireless channel. For example, if the UE moves along a straight line between two TRPs in an indoor environment, the channel between the UE and each TRP should have similar statistics. As another example, if the UE moves along a straight line between two TRPs in a rural environment, the channel between the UE and each TRP should have similar statistics. Reducing the density of RS patterns from one TRP may have a negligible impact on the CSI estimation error—while reducing signaling overhead—because the UE can utilize the RS pattern from another TRP to estimate its channel with the first TRP.

[0401] 5G NR supports flexibility in selecting RS patterns, where the selection of RS patterns is based on the statistics of the underlying randomly varying radio channel. For example, the parameter dmrs-AdditionalPosition can be used to increase the number of DMRS in a given time slot in high mobility scenarios. As another example, the parameters periodicityAndOffset-p and periodicityAndOffset-sp can be used to change the periodicity (and time slot offset) of the SRS. The details of the algorithm used to select the RS pattern are generally left to the network.

[0402] This disclosure describes a framework for supporting AI / ML techniques for multi-TRP RS density adaptation based on statistics of underlying randomly varying wireless channels. The corresponding signaling details are discussed in this disclosure.

[0403] The present disclosure solves the problem that multi-TRP RS density adaptation currently depends on network implementation. The present disclosure provides a method that the network can use to configure the density of multi-TRP RS using an AI / ML-based solution. The present disclosure also provides a framework for adapting the density of multi-TRP RS based on UE inference and information.

[0404] Details are disclosed regarding AI / ML techniques supporting multi-TRP RS density adaptation, including information elements to be exchanged between transmitters and receivers.

[0405] Fig.39 A high-level flowchart illustrating an example of BS operations for supporting multi-TRP RS density adaptation according to various embodiments of the present disclosure is shown. Fig.39 The embodiment is for illustration only. Other embodiments of process 3900 may be used without departing from the scope of the present disclosure.

[0406] Fig.39is an example of a method 3900 for supporting the operation of AI / ML techniques for multi-TRP RS density adaptation at a BS. In operation 3901, the BS receives UE capability information from the UE, including support for the ML method for multi-TRP RS density adaptation. In operation 3902, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for multi-TRP RS density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 3903, the BS sends a multi-TRP CSI-RS density configuration message to the UE. In one example, the message may correspond to two TRP A and TRP B. In operation 3904, the BS receives auxiliary information from the UE; the auxiliary information may include a measurement of multi-TRP channel similarity, which will be described below. In operation 3905, the BS sends a multi-TRP DMRS density configuration message to the UE. In one example, the message may correspond to TRP B.

[0407] Fig.40 A high-level flow chart illustrating an example of UE operation for supporting multi-TRP RS density adaptation according to various embodiments of the present disclosure is shown. Fig.40 The embodiment is for illustration only. Other embodiments of process 4000 may be used without departing from the scope of the present disclosure.

[0408] Fig.40 is an example of a method 4000 for supporting the operation of AI / ML techniques for multi-TRP RS density adaptation at a UE. In operation 4001, the UE reports capability information to the BS, including support for the ML method for multi-TRP RS density adaptation. In operation 4002, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for multi-TRP RS density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 4003, the UE receives a multi-TRP CSI-RS density configuration message from the BS. In one example, the message may correspond to two TRP A and TRP B. In operation 4004, the UE sends auxiliary information to the BS; the auxiliary information may include a measurement of multi-TRP channel similarity, which will be described below. In operation 4005, the UE receives a multi-TRP DMRS density configuration message from the BS. In one example, the message may correspond to TRP B.

[0409] Fig.41 A high-level flow chart illustrating an example of BS operation for supporting multi-TRP RS disabling according to various embodiments of the present disclosure is shown. Fig.41The embodiment is for illustration only. Other embodiments of process 4100 may be used without departing from the scope of the present disclosure.

[0410] Fig.41 is an example of a method 4100 for supporting the operation of AI / ML techniques for multi-TRP RS disabling at a BS. In operation 4101, the BS receives UE capability information from the UE, including support for ML methods for multi-TRP RS density adaptation. In operation 4102, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for multi-TRP RS density adaptation, the ML model to be used, the trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 4103, the BS sends a multi-TRP RS density configuration message to the UE. In one example, the message may correspond to two TRP A and TRP B. In operation 4104, the BS receives auxiliary information from the UE; the auxiliary information may include a recommendation for multi-TRP RS disabling, which will be described below. In one example, the message may correspond to TRPB. In operation 4105, the BS sends a multi-TRP RS disabling configuration message to the UE. In one example, the message may correspond to TRPB.

[0411] Fig.42 A high-level flow chart illustrating an example of UE operation for supporting multi-TRP RS disabling according to various embodiments of the present disclosure is shown. Fig.42 The embodiment is for illustration only. Other embodiments of process 4200 may be used without departing from the scope of the present disclosure.

[0412] Fig.42 is an example of a method 4200 for supporting AI / ML techniques for multi-TRP RS disabling at a UE. In operation 4201, the UE reports capability information to the BS, including support for the ML method for multi-TRP RS density adaptation. In operation 4204, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for multi-TRP RS density adaptation, the ML model to be used, the trained model parameters, and / or whether the model parameter update reported by the UE will be used. In operation 4203, the UE receives a multi-TRP RS density configuration message from the BS. In one example, the message may correspond to two TRP A and TRP B. In operation 4204, the UE sends auxiliary information to the BS; the auxiliary information may include a recommendation for multi-TRP RS disabling, which will be described below. In one example, the message may correspond to TRP B. In operation 4205, the UE receives a multi-TRP RS disabling configuration message from the BS. In one example, the message may correspond to TRP B.

[0413] The UE may also support AI / ML techniques for inferring its channel with TRP B given the CSI received from another TRP A, assuming that the channels have similar statistics. In one example, the AI / ML model may be trained offline with a dataset where each (training example, training label) pair maps to (CSI of TRP A, CSI of TRP B). The dataset may be collected within a range of distances between the UE and TRP A and TRP B. The UE may then run the AI / ML model in inference mode.

[0414] Fig.43 A high-level flow chart illustrating an example of BS operation for supporting multi-TRP CSI information according to various embodiments of the present disclosure is shown. Fig.43 The embodiment is for illustration only. Other embodiments of process 4300 may be used without departing from the scope of the present disclosure.

[0415] Fig.43 is an example of a method 4300 for supporting operations of AI / ML techniques for multi-TRP CSI information at a BS. In operation 4301, the BS receives UE capability information from the UE, including support for ML methods for multi-TRP CSI information. In operation 4302, the BS sends configuration information to the UE, which may include ML-related configuration information, such as enabling / disabling the ML method for multi-TRP CSI information, the ML model to be used, trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 4303, the BS sends a CSI-RS configuration message to the UE. In one example, the message may correspond to TRP A. In operation 4304, the BS receives auxiliary information from the UE; in one example, the auxiliary information may include CSI information of another TRP B, which will be described below.

[0416] Fig.44 A high-level flow chart illustrating an example of UE operation for supporting multi-TRP CSI information according to various embodiments of the present disclosure is shown. Fig.44 The embodiment is for illustration only. Other embodiments of process 4400 may be used without departing from the scope of the present disclosure.

[0417] Fig.44is an example of a method 4400 for supporting operations of AI / ML techniques for multi-TRP CSI information at a UE. In operation 4401, the UE reports capability information to the BS, including support for ML methods for multi-TRP CSI information. In operation 4402, the UE receives configuration information from the BS, which may include ML-related configuration information, such as enabling / disabling the ML method for multi-TRP CSI information, the ML model to be used, trained model parameters, and / or whether to use the model parameter update reported by the UE. In operation 4403, the UE receives a CSI-RS configuration message from the BS. In one example, the message may correspond to TRP A. In operation 4404, the UE sends auxiliary information to the BS; in one example, the auxiliary information may include CSI information of another TRP B, which will be described below.

[0418] In one embodiment, the BS may configure the UE to (at least temporarily) disable DMRS in subsequent time slots of a given TRP. Table 14 is an example of modifying the IE DMRS-DownlinkConfig to disable DMRS in subsequent time slots of a given TRP, where the IE is included in the IE PDSCH-Config already configured for the TRP. In another example, a new DCI format may be defined to support DMRS time disabling for a given TRP. The DCI format may include a TCI state corresponding to a QCL relationship between 1) a DM-RS port for the PDSCH of the TRP and 2) a CSI-RS port for a given CSI-RS resource of the TRP. The DCI format may include a DMRS time disabling indication for a given TRP of the UE.

[0419] Table 14

[0420]

[0421] Table 14. Example of IE DMRS-DownlinkConfig modification to support DMRS time disabling for TRP

[0422] UE assistance information reporting provides several advantages over relying on existing signaling. For example, the BS can use SRS to estimate the UL (and DL, depending on reciprocity) channel from the UE. However, the minimum periodicity of SRS is 2ms; in contrast, the interval between consecutive DMRS can be configured to be less than 1ms. Therefore, the UE can use the received DMRS to perform more fine-grained measurements of the DL channel than the BS uses the received SRS to measure the UL channel.

[0423] As another example, the UE may report local information that may not be available to the BS. The UE may use its camera to detect oncoming vehicles that will cross its line of sight with the BS within T seconds. The UE may then report this information to the BS and make a preemptive recommendation for transmission mode switching (e.g., switching to a relatively robust mode such as transmit diversity) within T seconds.

[0424] In one embodiment, a new MAC CE may be defined for UE assistance information reporting. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [2]. The MAC CE may have a variable size and may consist of the following fields:

[0425] - Block Error Rate: This field indicates the observed block error rate of the UE, for example, the block error rate has been calculated over the last 1000 received transport blocks.

[0426] -Throughput: This field indicates the observed throughput of the UE, e.g. the throughput (in Mbit / s) that has been calculated over the last 1000 received transport blocks.

[0427] -UE Speed: This field indicates the UE's measurement of its speed (in meters per second).

[0428] -UE acceleration: This field indicates the UE's measurement of its acceleration in meters / (seconds*seconds).

[0429] -IR1: This field indicates the presence of an octet containing the TRP Channel Affinity field. If the IR1 field is set to 1, then the octet containing the TRP Channel Affinity field is present. If the IR1 field is set to 0, then the octet containing the TRP Channel Affinity field is not present.

[0430] -TRP Channel Similarity: This field indicates the UE's measure of the similarity between its channel and two TRPs; examples of similarity measures include NMSE, cosine similarity, etc.

[0431] ◆Before calculating the similarity metric, the UE may use an AI / ML method or a non-AI / ML method to estimate its channels with the two TRPs.

[0432] -IR2: This field indicates the presence of an octet containing the Recommended TRP DMRS Time Density field. If the IR2 field is set to 1, then the octet containing the Recommended TRP DMRS Time Density field is present. If the IR2 field is set to 0, then the octet containing the Recommended TRP DMRS Time Density field is not present.

[0433] -Recommended TRP DMRS time density: This field indicates the recommended TRP DMRS time density for the UE, for example, an index into a table of DMRS time density values ​​for a given TRP.

[0434] ◆This field may also include the reason for the recommended TRP DMRS time density, for example, an index into a code table. For example, one code may correspond to the reason "oncoming vehicle" while another code may correspond to the reason "reporting traffic accident at mile marker 30".

[0435] The BS may configure the UE with a TRP DMRS time density threshold T, where the UE may compare its recommended DMRS time density for each TRP with T. The UE may be configured to apply its recommended DMRS time density for each TRP on the PUSCH if its recommended DMRS time density for each TRP exceeds T, without waiting for a configuration message from the BS.

[0436] Fig.45 An example of a new MAC CE for UE assistance information reporting according to various embodiments of the present disclosure is shown. Fig.45 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0437] Fig.45 An example of a new MAC CE 4500 for UE assistance information reporting is shown, wherein the block error rate field 4501, UE throughput 4502, UE speed 4503 and UE acceleration field 4504 each have a length of 8 bits; the TRP channel similarity field 4506 and the recommended TRP DMRS time density field 4508 each have a length of 7 bits.

[0438] In another embodiment, a new MAC CE may be defined for DMRS disabling recommendations for a given TRP. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [2]. The MAC CE may have a variable size and consist of the following fields:

[0439] -CSI-RS Process ID: This field indicates the ID of the CSI-RS process corresponding to a given TRP; the CSI-RS ports of the associated CSI-RS resources have a QCL relationship with the DM-RS ports of the PDSCH of this TRP.

[0440] -TCI State: This field indicates the ID of the TCI state of the QCL relationship between the CSI-RS port of the CSI-RS resource of a given TRP and the DM-RS port of the PDSCH.

[0441] -IR: This field indicates the presence of an octet containing a CORESET Pool ID field. If the IR field is set to 1, then the octet containing the CORESET Pool ID field is present. If the IR field is set to 0, then the octet containing the CORESET Pool ID field is not present.

[0442] -CORESET Pool ID: This field indicates the ID of the CORESET pool corresponding to a given TRP; the CORESET pool is used to support multi-DCI based NC-JT.

[0443] Fig.46 An example of a new MAC CE recommended for DMRS disabling for a given TRP according to various embodiments of the present disclosure is shown. Fig.46 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0444] Fig.46 An example of a new MAC CE 4600 for DMRS disabling recommendation for a given TRP is shown, where the CSI-RS process ID field 4601 and the TCI status field 4602 each have a length of 8 bits; the CORESET pool ID field 4604 has a length of 7 bits.

[0445] In another embodiment, a new MAC CE may be defined for the CSI information of a given TRP. The MAC CE may be identified by a MAC subheader with a logical channel ID, which may be specified in Table 6.2.1-2 in [3]. The MAC CE may have a variable size and may consist of the following fields:

[0446] -CQI: This field indicates the value of the inferred CQI for a given TRP, e.g., an index into a CQI table. In one example, this value may be the inferred wideband CQI.

[0447] -PMI: This field indicates the value of the inferred PMI for a given TRP, such as an index into a PMI table. In one example, this value may be the inferred wideband PMI.

[0448] Fig.47 An example of a new MAC CE for CSI information of a given TRP according to various embodiments of the present disclosure is shown. Fig.47 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0449] Fig.47An example of a new MAC CE 4700 for CSI information for a given TRP is shown, where the CQI field 4701 and the PMI field 4702 each have a length of 8 bits.

[0450] Fig.48 An example of an AI / ML model architecture supporting multi-TRP RS density adaptation according to various embodiments of the present disclosure is shown. Fig.48 The embodiments shown are for illustration only. Other embodiments may be used without departing from the scope of the present disclosure.

[0451] Fig.48 An example of an AI / ML model architecture 4800 that can support multi-TRP RS density adaptation is shown. In this example, a standard CNN can be enhanced as follows:

[0452] - A feature selection layer 4802[8] can be placed before the CNN 4803 to learn the relative weights of different CSI-RS tones 4801 for TRP A; the learned weights can be used to infer the throughput maximizing RS density for TRP A.

[0453] -The CSI-RS pattern 4804 of another TRP B can be used as auxiliary information [6] to infer the channel information of TRP A in CNN 4805 (e.g., full CSI, throughput maximizing RS density).

[0454] -The outputs of these two CNNs 4803, 4805 are then convolved (box 4806), and the output of the convolution is provided as input to another CNN 4807, which infers the channels 4808 for TRP A and TRP B.

[0455] In each time slot, a complete two-dimensional channel across all subcarriers and OFDM symbols is provided as input to the AI / ML model architecture 4800, where all REs not containing CSI-RS are padded with zeros.

[0456] Examples of inputs to AI / ML models that can support multi-TRP RS density adaptation include:

[0457] - A complete two-dimensional receive channel over all subcarriers and OFDM symbols of one time slot and one TRP A.

[0458] ◆All REs not containing RS are filled with zeros.

[0459] -The ground truth of the complete two-dimensional channel over all subcarriers and OFDM symbols of one time slot and one TRP A (or another TRP B, as in operation 4404).

[0460] ◆This can be a training label.

[0461] - Auxiliary information.

[0462] ◆Complete two-dimensional receive channel on all subcarriers and OFDM symbols of one time slot and another TRP B.

[0463] ■All REs not containing RS are filled with zeros.

[0464] ◆UE speed.

[0465] ◆UE track.

[0466] ◆UE location.

[0467] ◆Relevant time

[0468] ◆Coherence bandwidth

[0469] Delayed expansion

[0470] ◆Doppler spread

[0471] Block Error Rate

[0472] ◆Throughput.

[0473] ◆Spatial separation between TRP A and TRP B.

[0474] Examples of outputs from AI / ML models that can support multi-TRP RS density adaptation include:

[0475] - A complete two-dimensional estimated channel over all subcarriers and OFDM symbols of one time slot and one TRP A (or another TRP B, as in operation 4404).

[0476] - The (inferred) throughput maximizing value of RS density for TRP A.

[0477] For the purpose of illustration, the steps of the above algorithm are described continuously. However, some of these steps can be performed in parallel with each other. The operation diagram illustrates an example method that can be implemented according to the principles of the present disclosure, and various changes can be made to the method illustrated in the flow chart. For example, although shown as a series of steps, the various steps in each figure can overlap, occur in parallel, occur in different orders, or occur multiple times. In another example, a step can be omitted or replaced by another step.

[0478] Although the present disclosure has been described with exemplary embodiments, various changes and modifications may be suggested to one skilled in the art. The present disclosure is intended to encompass such changes and modifications as fall within the scope of the appended claims.

Claims

1. A method, include: sending information to a base station (BS) indicating a capability of the UE to support machine learning (ML) adaptation of a reference signal (RS) density in a domain; receiving first configuration information from the BS, the first configuration information indicating one or more of: enabling or disabling ML adaptation of RS density in the domain, an ML model for adaptation of RS density in the domain, updated model parameters of the ML model, and whether the model parameters received from the UE are to be used for ML adaptation of RS density in the domain; receiving second configuration information from the BS, where the second configuration information indicates a density of RSs in the first domain; receiving an RS having an RS density in a first domain; Send auxiliary information to BS; In response to transmitting the auxiliary information, receiving third configuration information indicating a density of RSs in the second domain from the BS, wherein the density of RSs in the first domain is greater than the density of RSs in the second domain; and An RS having a density of RSs in a second domain is received.

2. The method according to claim 1, in, The domain is one of a time domain, a frequency domain, and a spatial domain.

3. The method according to claim 1, include: Send a fallback request to the BS, where the fallback request indicates the RS density in the domain; receiving, from the BS, a fourth configuration indicating a density of RSs in a third domain; and receiving an RS having an RS density in a third domain, The RS density in the third domain is greater than the RS density in the second domain.

4. The method according to claim 1, include: receiving an information element indicating one of an RS frequency or time density value of consecutive resources and an RS frequency or time density value of some resources mapped to consecutive antenna ports configured for RS; and The RS is received in one of the continuous resources and the resources mapped to the continuous antenna ports by switching between the RS density in the first domain and the RS density in the second domain based on the information element.

5. The method according to claim 1, in: The auxiliary information includes one or more of a block error rate, a UE speed, a UE acceleration, and a recommended RS density in the domain.

6. The method according to claim 1, further comprising: include: A configuration of RS density in a disabled area is received from a BS.

7. A user equipment (UE), include: processor; as well as A transceiver operatively coupled to the processor, the transceiver being configured to: Sending information to a base station (BS) indicating the capability of the UE to support machine learning (ML) adaptation of reference signal (RS) density in a domain, receiving first configuration information from the BS, the first configuration information indicating one or more of: enabling or disabling ML adaptation of RS density in the domain, an ML model for adaptation of RS density in the domain, updated model parameters of the ML model, and whether the model parameters received from the UE are to be used for ML adaptation of RS density in the domain, receiving second configuration information from the BS, the second configuration information indicating a density of RSs in the first domain, receiving an RS having an RS density in a first domain, Send auxiliary information to BS; In response to sending the auxiliary information, receiving third configuration information indicating a density of RSs in the second domain from the BS, wherein the density of RSs in the first domain is greater than the density of RSs in the second domain, and An RS having a density of RSs in a second domain is received.

8. The UE according to claim 7, in, The domain is one of a time domain, a frequency domain, and a spatial domain.

9. The UE according to claim 7, in, The transceiver is also configured to: Send a fallback request to the BS, the fallback request indicates the RS density in the domain, receiving, from the BS, a fourth configuration indicating a density of RSs in a third domain, and receiving an RS having an RS density in a third domain, The RS density in the third domain is greater than the RS density in the second domain.

10. The UE according to claim 7, in, The transceiver is also configured to: receiving an information element indicating one of an RS frequency or time density value of a continuous resource and an RS frequency or time density value of some resources mapped to continuous antenna ports configured for the RS, and The RS is received in one of the continuous resources and the resources mapped to the continuous antenna ports by switching between the RS density in the first domain and the RS density in the second domain based on the information element.

11. The UE according to claim 7, in, The transceiver is also configured to: Receive the configuration of RS density in the disabled domain from the BS, The auxiliary information includes one or more of a block error rate, a UE speed, a UE acceleration, and a recommended RS density in the domain.

12. A base station (BS), include: processor; and A transceiver operatively coupled to the processor, the transceiver being configured to receiving information from a user equipment (UE) indicating a capability of the UE to support machine learning (ML) adaptation of reference signal (RS) density in a domain, sending first configuration information to the UE, the first configuration information indicating one or more of: enabling or disabling ML adaptation of RS density in the domain, an ML model for RS density adaptation in the domain, updated model parameters of the ML model, and whether the model parameters received from the UE are to be used in ML adaptation of RS density in the domain, Sending second configuration information to the UE, where the second configuration information indicates the RS density in the first domain, sending an RS having an RS density in a first domain, Receive assistance information from the UE, In response to receiving the auxiliary information, third configuration information indicating a density of RSs in the second domain is sent to the UE, wherein the density of RSs in the first domain is greater than the density of RSs in the second domain, and An RS having an RS density in a second domain is transmitted.

13. The BS according to claim 12, in, The transceiver is also configured to: Receive a fallback request from the UE, the fallback request indicating the RS density in the domain, sending a fourth configuration indicating a density of RSs in a third domain to the UE, and Sending RS with RS density in the third domain, where domain is one of time, frequency, or space, and The RS density in the third domain is greater than the RS density in the second domain.

14. The BS according to claim 12, in, The transceiver is also configured to: transmitting an information element indicating one of an RS frequency or time density value of a continuous resource and an RS frequency or time density value of some resources mapped to continuous antenna ports configured for RS, and The RS is transmitted in one of the continuous resources and the resources mapped to the continuous antenna ports by switching between the RS density in the first domain and the RS density in the second domain based on the information element.

15. The BS according to claim 12, in: The auxiliary information includes one or more of a block error rate, a UE speed, a UE acceleration, and a recommended RS density in the domain.