Method for performance monitoring of a processing process

Abstract

A method includes determining at least one performance metric for a processing procedure performed by a wireless terminal device, and reporting, by the wireless terminal device, the at least one performance metric to a radio access network node. In other embodiments, the method comprises reporting, by a wireless terminal device, actual channel measurement information and / or predicted channel information for at least one dedicated occasion to a wireless access network node, the radio access network node may thereby compute at least one performance metric for the processing procedure based at least in part on the actual channel measurement information and / or the predicted channel information.

Description

Technical Field

[0001] This disclosure generally relates to performance monitoring of the processing procedure. Background Technology

[0002] Artificial Intelligence / Machine Learning (AI / ML) has been researched and applied across various fields. AI / ML can be used to extract features that cannot be derived mathematically. Typically, AI / ML models are trained on a dataset, implicitly learning the mapping relationships between data. The AI / ML model can then generate a set of outputs based on a set of inputs. However, because the performance of AI / ML models is closely related to the training dataset, generalization is a problem, as AI / ML models may not perform well in scenarios / configurations different from the training scenario / configuration.

[0003] Some research also aims to improve the efficiency of wireless communication systems. In these cases, User Equipment (UE) can deploy AI / ML models to obtain more accurate Channel State Information (CSI) and feedback from the network. However, due to the generalization problem of AI / ML models, they may not perform well in scenarios / configurations different from the training scenario / configuration. Therefore, the performance of AI / ML models may decrease under different scenarios / configurations. Summary of the Invention

[0004] This disclosure generally relates to performance monitoring of processing procedures. For processing procedures that may involve AI / ML models, having a performance monitoring process is crucial. If an AI / ML model fails to meet performance targets, it may be deactivated / updated / switched to another AI / ML model / rollback to a traditional approach. Therefore, performance monitoring of AI / ML models is essential as it can directly impact system performance. Currently, there is no comprehensive monitoring scheme / process for AI / ML models in communication systems. This disclosure addresses this deficiency and relates to methods for monitoring AI / ML models and / or processing procedures related to AI / ML models.

[0005] In various embodiments, "AL / ML model" can refer to a generic term used to describe a process, function, or feature that the UE is capable of performing. According to various embodiments, this function, process, or feature may include, involve, or be enabled by an AI / ML model. Furthermore, "UE-side model" can refer to a generic term used to describe a UE-side model in a single-sided UE model or a portion of a UE portion of a dual-sided model.

[0006] In some exemplary embodiments, a method performed by a wireless terminal device includes: determining at least one performance metric of a processing procedure, wherein the processing procedure is performed by the wireless terminal device; and reporting the at least one performance metric to a radio access network node. The method may include: measuring actual channel measurement information; generating predicted channel information using the processing procedure; and calculating the at least one performance metric based at least partially on the predicted channel information and the actual channel measurement information. Similarly, a method performed by a radio access network node may include receiving at least one performance metric of a processing procedure from a wireless terminal device, wherein the processing procedure is performed by the wireless terminal device.

[0007] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless terminal device determining at least one performance metric of a processing procedure for at least one timing, and reporting at least one performance metric of the processing procedure for at least one timing to a radio access network node. The timing may include at least one time unit, time slot, or symbol. The method may include: the wireless terminal device 104 receiving an indication of at least one dedicated timing for indicative of the performance of the processing procedure; determining at least one performance metric of the processing procedure for at least one dedicated timing based at least in part on actual channel measurement information and predicted channel information for at least one dedicated timing; and reporting at least one performance metric of the processing procedure for at least one dedicated timing.

[0008] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include at least one performance metric for a processing procedure in which a radio access network node receives at least one timing. In various examples, a timing includes at least one time unit, time slot, or symbol. The method may also include an indication of at least one dedicated timing for indicative of the performance of the processing procedure, and at least one performance metric for receiving at least one dedicated timing.

[0009] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include: a wireless terminal device receiving indications of a plurality of dedicated timings for indicative of the performance of a processing procedure; determining at least one performance metric of the processing procedure for a subset of the plurality of dedicated timings, based at least in part on actual channel measurement information and predicted channel information of a subset of the plurality of dedicated timings; and reporting at least one performance metric of the subset of the plurality of dedicated timings. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a radio access network node indicating indications of a plurality of dedicated timings for indicative of the performance of a processing procedure; and receiving at least one performance metric of a subset of the plurality of dedicated timings.

[0010] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include: a wireless terminal device receiving an indication of a plurality of dedicated timeframes for indicating the performance of a processing procedure; determining at least one performance metric of the processing procedure for the plurality of dedicated timeframes as an average performance metric of the plurality of dedicated timeframes, based at least in part on actual channel measurement information and predicted channel information of the plurality of dedicated timeframes; and reporting the average performance metric of the plurality of dedicated timeframes. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a wireless access network node indicating an indication of a plurality of dedicated timeframes for indicating the performance of a processing procedure; and receiving the average performance metric of the plurality of dedicated timeframes.

[0011] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include: a wireless terminal device receiving an indication of a plurality of dedicated timings for indicative of the performance of a processing procedure; determining at least one performance metric of the processing procedure for the plurality of dedicated timings as at least one average performance metric of at least one group of the plurality of dedicated timings, based at least in part on actual channel measurement information and predicted channel information of the plurality of dedicated timings; and reporting at least one average performance metric of at least one group of the plurality of dedicated timings. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a radio access network node indicating an indication of a plurality of dedicated timings for indicative of the performance of a processing procedure; and receiving at least one average performance metric of at least one group of the plurality of dedicated timings.

[0012] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include: a wireless terminal device receiving indications of a plurality of dedicated timings for indicative of the performance of a processing procedure; determining at least one performance metric of the processing procedure of a subset of the plurality of dedicated timings as at least one average performance metric of the subset of dedicated timings, based at least in part on actual channel measurement information and predicted channel information of a subset of the plurality of dedicated timings; and reporting at least one average performance metric of the subset of dedicated timings. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a radio access network node indicating indications of a plurality of dedicated timings for indicative of the performance of a processing procedure; and receiving at least one average performance metric of a subset of the plurality of dedicated timings.

[0013] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include: a wireless terminal device receiving indications of a plurality of dedicated timings for indicating the performance of a processing procedure; determining at least one performance metric of the processing procedure of a subset of the plurality of dedicated timings as at least one average performance metric of at least one group of at least one subset of the plurality of dedicated timings, based at least in part on actual channel measurement information and predicted channel information of a subset of the plurality of dedicated timings; and reporting at least one average performance metric of at least one group of at least one subset of the plurality of dedicated timings. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a wireless access network node indicating indications of a plurality of dedicated timings for indicating the performance of a processing procedure; and receiving at least one average performance metric of at least one group of at least one subset of the plurality of dedicated timings.

[0014] In some exemplary embodiments, this method can be combined with any of the other exemplary embodiments disclosed herein, and may include a wireless terminal device determining the validity period of predicted channel information, and / or determining the validity period of predicted channel information at least in part based on a performance metric, and / or determining the performance metric at the granularity of the performance metric calculation. In some exemplary embodiments, this method can be combined with any of the other exemplary embodiments disclosed herein, and may include a wireless access network node receiving the validity period of predicted channel information determined by the processing of the wireless terminal device, and / or indicating the granularity of the performance metric calculation to be performed by the wireless terminal device.

[0015] In some exemplary embodiments, this method can be combined with any of the other exemplary embodiments disclosed herein. The method may include a wireless terminal device reporting channel-related information and / or generating predicted channel information using a processing procedure, wherein reporting channel-related information includes reporting predicted channel information and actual channel measurement information. In some exemplary embodiments, this method can be combined with any of the other exemplary embodiments disclosed herein. The method may include a wireless access network node receiving channel-related information and / or receiving predicted channel information and actual channel measurement information, wherein the predicted channel information is generated by the wireless terminal device using a processing procedure.

[0016] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a wireless terminal device generating predicted channel information using a processing procedure, wherein reporting channel-related information includes reporting the difference between actual channel measurement information and predicted channel information; and reporting the predicted channel information. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a wireless access network node receiving predicted channel information and the difference between actual channel measurement information and predicted channel information, wherein the predicted channel information is generated by the wireless terminal device using a processing procedure.

[0017] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, a method may include a wireless terminal device measuring actual channel measurement information at at least one dedicated time to indicate the performance of a processing procedure, and reporting the actual channel measurement information for at least one dedicated time to a radio access network node. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, a method may include a radio access network node receiving actual channel measurement information from a wireless terminal device for at least one dedicated time, and calculating at least one performance metric of the processing procedure based at least in part on the actual channel measurement information.

[0018] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, a method may include a wireless terminal device generating predicted channel information for at least one dedicated time period to indicate the performance of the processing process using a processing procedure, wherein the processing procedure is performed by the wireless terminal device; and reporting the predicted channel information for at least one dedicated time period to a radio access network node. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, a method may include a radio access network node receiving predicted channel information for at least one dedicated time period to indicate the performance of the processing process; and calculating at least one performance metric of the processing process based at least in part on the predicted channel information, wherein the processing procedure is performed by the wireless terminal device.

[0019] In some exemplary embodiments, this method can be combined with any of the other exemplary embodiments disclosed herein. The method may include a wireless terminal device reporting channel-related information by reporting partial channel-related information, and / or generating predicted channel information using a processing procedure, wherein reporting the channel-related information includes reporting the predicted channel information and at least one performance metric. In some exemplary embodiments, this method can be combined with any of the other exemplary embodiments disclosed herein. The method may include a radio access network node receiving partial channel-related information, and / or receiving the predicted channel information and at least one performance metric, wherein the predicted channel information is generated by the wireless terminal device using a processing procedure.

[0020] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless terminal device receiving a configuration of a processing procedure determined by a radio access network node at least in part based on at least one performance metric. This configuration may include the number of prediction opportunities, the length of a prediction window, or the length of the time span of the wireless terminal device's processing procedure, and / or the period of the prediction opportunities for the wireless terminal device's processing procedure. The method may include a configuration for receiving the processing procedure via at least one of the following: a Medium Access Control Control Element (MAC-CE) or Downlink Control Information (DCI) signaling.

[0021] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a radio access network node configuring a processing procedure of a wireless terminal device based at least in part on at least one performance metric; and instructing the configuration of the processing procedure. The configuration may include configuring the number of prediction opportunities, the length of a prediction window, or the length of the time span of the processing procedure of the wireless terminal device, and / or the period of the prediction opportunities for the processing procedure of the wireless terminal device. The method may also include instructing the configuration of the processing procedure via at least one of: a Media Access Control Element (MAC-CE) or Downlink Control Information (DCI) signaling.

[0022] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless terminal device performing performance monitoring of a processing procedure, wherein the processing procedure is performed by the wireless terminal device; and reporting an indication regarding the performance monitoring. The indication may include an indicator regarding the result of the performance monitoring. The method may include: performing performance monitoring of the processing procedure at least in part based on a performance threshold; and / or reporting an indication regarding the performance monitoring, the indication including a one-bit indication of whether the processing procedure is valid; and / or receiving a performance threshold; and / or reporting a performance threshold; and / or determining a performance threshold.

[0023] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless access network node receiving an indication of performance monitoring of a processing procedure performed by a wireless terminal device, wherein the processing procedure is performed by the wireless terminal device, and wherein the performance monitoring is performed by the wireless terminal device. The indication may include an indicator of the result of the performance monitoring. The method may further include receiving an indication of performance monitoring, including an indication of whether the processing procedure is valid, and / or receiving an indication of performance monitoring, including an indication of whether the processing procedure meets a performance threshold, and / or an indication of the performance threshold used by the wireless terminal device in the performance monitoring of the processing procedure. The indication of whether the processing procedure meets a performance threshold may include a one-bit indication of whether the processing procedure is valid, at least in part based on the performance threshold. The method may further include receiving a performance threshold.

[0024] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless terminal device reporting an indication regarding performance monitoring, including an indication of the current performance level of the processing. In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless access network node receiving an indication regarding performance monitoring, including an indication of the current performance level of the processing. The indication of the current performance level of the processing may include multiple bits for indicating the current performance level.

[0025] In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a wireless terminal device reporting a recommendation indication that includes a recommendation on whether a radio access network node should perform a decision process regarding a processing procedure, and / or reporting a recommendation indication that includes multiple bits indicating a specific recommended decision regarding the processing procedure. In some exemplary embodiments, which can be combined with any of the other exemplary embodiments disclosed herein, the method may include a radio access network node receiving a recommendation indication that includes a recommendation on whether a radio access network node should perform a decision process regarding a processing procedure, and / or receiving a recommendation indication that includes multiple bits indicating a specific recommended decision regarding the processing procedure.

[0026] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a configuration of a wireless terminal device receiving a processing procedure, wherein the configuration is at least in part based on instructions regarding performance monitoring, and / or on performance monitoring of the processing procedure to determine the action procedure of the wireless terminal device regarding the processing procedure, and / or on the action procedure of the processing procedure determined by the wireless terminal device.

[0027] In some exemplary embodiments, and in combination with any of the other exemplary embodiments disclosed herein, the method may include a wireless access network node determining an action procedure for a wireless terminal device's processing based on an instruction regarding performance monitoring, and / or configuring the wireless terminal device's processing based on an instruction regarding performance monitoring, and instructing the configuration of the processing procedure, and / or receiving an action procedure for the wireless terminal device's processing determined by the wireless terminal device based on performance monitoring of the processing procedure.

[0028] In some other embodiments, an apparatus for wireless communication, such as a network device, is disclosed. The network device may include one or more processors and one or more memories, wherein the one or more processors are configured to read computer code from the one or more memories to implement any of the methods described above. The apparatus for wireless communication may be a wireless access network node (e.g., a base station) or a wireless terminal device (e.g., a UE).

[0029] In some other embodiments, a computer program product is disclosed. The computer program product may include a non-transitory computer-readable medium storing computer code that, when executed by one or more processors, causes the one or more processors to implement any of the methods described above.

[0030] The above embodiments and other aspects and alternatives to the embodiments are described in more detail below with reference to the accompanying drawings, detailed description and claims. Attached Figure Description

[0031] Figure 1 A radio access network with exemplary uplink, downlink, and control channel configurations is shown.

[0032] Figure 2 It shows Figure 1 Various example processing components for wireless terminal devices and wireless access network nodes.

[0033] Figure 3 Illustrative examples of methods for determining performance metrics according to various embodiments are shown.

[0034] Figure 4 Another illustrative example of a method for determining performance metrics according to various embodiments is shown.

[0035] Figure 5 , Figure 6 , Figure 7 , Figure 8 , Figure 9 , Figure 10 and Figure 11 Illustrative examples of methods for reporting performance metrics according to various embodiments are shown respectively. Detailed Implementation

[0036] The techniques and examples of the implementation methods and / or embodiments described in this disclosure can be used to facilitate the allocation, configuration, and signaling of over-the-air radio resources in a radio access network, as well as the operational configuration of UEs and / or base stations within the radio access network, including monitoring and / or configuration processes that may involve AI / ML models or be enabled by AI / ML modules. The term "exemplary" is used to mean "an example of..." and, unless otherwise stated, does not imply an ideal or preferred example, implementation method, or embodiment. Section headings in this disclosure are used to facilitate understanding of the disclosed implementation methods and are not intended to limit the techniques disclosed in a section to the corresponding section. The disclosed implementation methods may be further embodied in various different forms, and therefore, the scope of this disclosure or the claimed subject matter is intended to be construed as not being limited to any embodiment set forth below. Various implementations may be practiced as methods, apparatus, components, systems, or non-transitory computer-readable media. Accordingly, embodiments in this disclosure may take the form of, for example, hardware, software, firmware, or any combination thereof.

[0037] Wireless Network Overview Wireless communication networks may include: a wireless access network for providing network access to wireless terminal devices, and a core network for routing data between access networks or between a wireless network and other types of data networks. In the wireless access network, wireless resources are provided for allocation and for transmitting data and control information. Figure 1 An exemplary wireless access network 100 is illustrated, comprising a wireless access network node (WANN) or wireless base station 102 (hereinafter referred to herein as a wireless base station, base station, wireless access node, wireless access network node, or WANN) and a wireless terminal equipment or user equipment (UE) 104 (hereinafter referred to herein as user equipment, UE, terminal equipment, or wireless terminal equipment), which communicate with each other via over-the-air (OTA) wireless communication resources 106. The wireless access network 100 may be implemented as, for example, a 2G, 3G, 4G / LTE, 5G, or 6G cellular wireless access network. Accordingly, the base station 102 may be implemented as a 2G base station, a 3G nodeB, an LTE eNB, or a 5G / 6G New Radio (NR) gNB. The user equipment 104 may be implemented as a mobile or fixed communication device equipped with a mobility identification module for accessing the base station 102. User equipment 104 may include, but is not limited to, mobile phones, laptops, tablets, personal digital assistants, wearable devices, distributed remote sensor devices, and desktop computers. Alternatively, wireless access network 100 may be implemented as other types of wireless access networks, such as Wi-Fi (Wireless Fidelity), Bluetooth, ZigBee, and WiMax (World Interoperability for Microwave Access) networks.

[0038] Figure 2 Further shown Figure 1The following are example processing components of WANN 102 and UE 104. UE 104 may include, for example, transceiver circuitry 206 coupled to one or more antennas 208 to enable wireless communication with WANN 102 (or other UEs). Transceiver circuitry 206 may also be coupled to processor 210, which may also be coupled to memory 212 or other storage devices. Memory 212 may be transient or non-transient and may store computer instructions or code therein that, when read and executed by processor 210, cause processor 210 to implement the various functions, methods, and processes of UE 104 as described herein. Memory 212 may also be utilized and allocated for buffering UL (Uplink) and DL (Downlink) transmissions in each frequency band / carrier. Memory 212 may include multiple memory modules (such as program memory, baseband memory, and / or RF memory, to name a few) allocated to different functions. Similarly, WANN 102 may include transceiver circuitry 214 coupled to one or more antennas 216, which may include antenna towers 218 of various forms, to enable wireless communication with UE 104. Transceiver circuitry 214 may also be coupled to one or more processors 220, which may also be coupled to memory 222 or other storage devices. Memory 222 may be transient or non-transient and may store instructions or code therein that, when read and executed by one or more processors 220, cause one or more processors 220 to implement the various functions, methods, and processes of WANN 102 as described herein.

[0039] Wireless communication resource scheduling / signaling Back Figure 1The wireless communication resources 106 for the air interface may include a combination of frequency, time, and / or spatial communication resources, which are organized into various resource units or elements in frequency, time, and / or space. The wireless communication resources 106 in the frequency domain may include portions of licensed radio frequency bands, portions of unlicensed radio frequency bands, or a mixture of both. The wireless communication resources 106, which can be used to carry wireless communication signals between base station 102 and user equipment 104, may be further divided into a physical downlink channel 110 for transmitting wireless signals from base station 102 to user equipment 104 and a physical uplink channel 120 for transmitting wireless signals from user equipment 104 to base station 102. The physical downlink channel 110 may also include a physical downlink control channel (PDCCH) 112 and a physical downlink shared channel (PDSCH) 114. Similarly, the physical uplink channel 120 may also include a physical uplink control channel (PUCCH) 122 and a physical uplink shared channel (PUSCH) 124. For simplicity, other types of downlink and uplink channels are not mentioned. Figure 1 As shown, but within the scope of this disclosure. Control channels PDCCH 112 and PUCCH 122 can be used to carry control information in the form of control messages 116 and 126 (referred to herein as downlink control information (DCI) messages or uplink control information (UCI) messages). Shared (shared between data and control information) channels PDSCH 114 and PUSCH 124 can be allocated and used to transmit downlink data transmission 118 and uplink data transmission 128 between base station 102 and user equipment 104.

[0040] The allocation and configuration of wireless communication resources associated with data channels (such as PDSCH and PUSCH) can be provided by one or more resource scheduling DCIs carried in the PDCCH. The PDCCH can be shared by multiple UEs in the access network. In various methods, a specific UE can be configured to perform a blind decoding process on a pre-configured UE-specific search space (USS) to detect and identify the payload of a UE-specific resource scheduling DCI carried in the PDCCH. Blind decoding can be performed at pre-configured monitoring times of the PDCCH associated with the USS. Such monitoring times can be referred to as a set of PDCCH candidates. Each PDCCH candidate can be associated with a set of Control Channel Elements (CCEs). The UE can specifically use its Radio Network Temporary Identifier (RNTI) to decode the PDCCH candidates. The RNTI can be used to demask the CRC (Cyclic Redundancy Check) of the PDCCH candidates. If no CRC error is detected, the UE determines that the PDCCH candidate carries its own control information. Then, the UE can process the DCI and extract resource allocation information related to the PDSCH and / or PUSCH to receive and / or transmit data.

[0041] Description of a new mechanism for performance monitoring of the processing process According to this disclosure, a method for providing performance monitoring of a processing procedure is disclosed. As described above, the processing procedure may be executed wholly or partially by an AI / ML model, may involve an AI / ML model, and / or may be enabled by an AI / ML model. Therefore, in various embodiments, the processing procedure may include processing utilizing AI or ML.

[0042] Due to the aforementioned AI / ML model generalization issues, AI / ML models may not perform well in deployment scenarios / configurations different from the training scenario / configuration. Considering the limited capabilities of AI / ML models affected by generalization, a public performance monitoring process is needed for the AI / ML models. In some cases, the performance monitoring process for the UE-side AI / ML model can be performed on the network side (e.g., base station 102) and / or the UE 104 side.

[0043] In some embodiments, performance monitoring of processes involving UE-side AI / ML models or processes enabled by UE-side AI / ML models is performed on the network side (e.g., at base station 102). In some embodiments, UE 104 reports one or more performance metrics to the network side.

[0044] In some embodiments, one or more performance metrics can be evaluated based on intermediate key performance indicators (KPIs). In some embodiments, intermediate KPIs can be calculated based on similarity (e.g., cosine similarity (CS) / generalized cosine similarity (GCS) / square generalized cosine similarity (SGCS)). In some embodiments, intermediate KPIs can be calculated based on difference / distance (e.g., Euclidean distance / mean square error (MSE) or normalized mean square error (NMSE)).

[0045] In one example, an AI / ML model is applied to predict Channel State Information (CSI). Figure 3 As shown, the predicted channel information and the actual channel measurement (e.g., via a reference signal) are respectively determined by H predict (H 预测 The inputs to the AI / ML model are denoted by H and H. The inputs are based on actual channel measurements from some prior time (e.g., observation time). previous (H 先前 Furthermore, the AI / ML model output is the predicted channel information H. predict In various embodiments, one or more performance metrics can be calculated based on predicted channel information and the actual channel.

[0046] In another example, an AI / ML model was applied to CSI prediction. For example... Figure 4 As shown, (one or more) predicted feature vectors / precoder information and (one or more) actual feature vectors / precoder information derived from channel measurements and some processing (e.g., Singular Vector Decomposition (SVD) of the channel measurements) are respectively derived by V predict (V 预测 ) and V represent the inputs to an AI / ML model. The input can be based on some prior time (e.g., the observation time). previous (V 先前 The actual channel measurement is used, while the AI / ML model output is the predicted channel information V. predictIn various embodiments, one or more performance metrics can be computed based on one or more predicted feature vectors / precoder information and one or more actual feature vectors / precoder information.

[0047] In some embodiments, one or more performance metrics can be evaluated based on a final KPI. In some embodiments, the final KPI can be calculated based on system metrics (e.g., throughput / average throughput / cell edge throughput / 5% tail throughput). In some embodiments, the final KPI can be calculated based on accuracy metrics (e.g., bit error rate (BER) / block error rate (BLER) / hypothetical BLER). In some embodiments, the final KPI can be calculated based on reliability metrics (e.g., signal-to-noise ratio (SNR) / signal-to-interference-plus-noise ratio (SINR) / modulation and coding scheme (MCS)).

[0048] According to various embodiments, a method for performance monitoring of a processing procedure performed by a wireless terminal device 104 may include: determining at least one performance metric of the processing procedure, wherein the processing procedure is performed by the wireless terminal device 104; and reporting the at least one performance metric to a wireless access network node 102 by the wireless terminal device 104. The method may include: the wireless terminal device 104 measuring actual channel measurement information; generating predicted channel information using the processing procedure; and calculating at least one performance metric based at least partially on the predicted channel information and the actual channel measurement information.

[0049] Similarly, a method for performance monitoring of a processing procedure performed by a wireless access network node 102 may include receiving at least one performance metric of the processing procedure from a wireless terminal device 104, wherein the processing procedure is performed by the wireless terminal device.

[0050] In some embodiments, at least one or more performance metrics at any given time are reported to the network (e.g., to base station 102).

[0051] In some embodiments, such as Figure 5 As shown, the network (e.g., base station 102) can indicate a dedicated timing for indicating model performance, and the UE 104 can report performance metrics for the dedicated timing.

[0052] In one example, an AI / ML model can be used for channel prediction. That is, the AI / ML model input at some previous time (e.g., the observation time) can be based on actual channel measurements. Then, the AI / ML model output at some future time (e.g., the prediction time) can be the predicted channel measurements. UE 104 may not need to acquire the actual channel measurements for the prediction time. However, if the time is designated as a dedicated time (e.g., a monitoring time) to indicate model performance, UE 104 can acquire both the actual and predicted channel measurements. UE 104 can then calculate performance metrics for the actual and predicted channel measurements and report these performance metrics to the network (e.g., base station 102) so that the network (e.g., base station 102) can perform performance monitoring and make subsequent decisions regarding the current AI / ML function / model. Note that (one or more) monitoring times can be dedicated monitoring times and / or times within the prediction time.

[0053] In some embodiments, the network (e.g., base station 102) may indicate more than one dedicated timing to indicate model performance, and the UE 104 may report one or more performance metrics of the dedicated timing.

[0054] In one example, the AI / ML model is used for channel prediction. That is, the AI / ML model input is based on actual channel measurements at some previous time (e.g., the observation time). Then, the AI / ML model output is the predicted channel measurement at some future time (e.g., the prediction time). UE 104 does not need to obtain the actual channel measurements at the prediction time. However, if some time is designated as a special time (e.g., the monitoring time) to indicate model performance, then UE 104 may need to obtain both the actual channel measurements and the predicted channel measurements. Then, as... Figure 6 As shown, UE 104 can calculate performance metrics for both actual and predicted channel measurements at each time point and report these performance metrics to the network (e.g., base station 102) so that the network can monitor performance and make subsequent decisions regarding the current AI / ML functions / models. For example, using squared generalized cosine similarity (SGCS) as a performance metric, the SGCS values ​​for the first three monitoring time points can be 0.9, 0.9, and 0.85, respectively, and these SGCS values ​​are reported by UE 104 to the network (e.g., base station 102).

[0055] According to various embodiments, the method performed by the wireless terminal device 104 may include determining at least one performance metric of a processing procedure for at least one timing, and reporting at least one performance metric of the processing procedure for at least one timing to the radio access network node 102. In various examples, a timing includes at least one time unit, time slot, or symbol. The method may include the wireless terminal device 104 receiving an indication from the radio access network node 102 of at least one dedicated timing for indicative of the performance of the processing procedure. The method may also include: determining at least one performance metric of a processing procedure for at least one dedicated timing by the wireless terminal device 104 based at least in part on actual channel measurement information and predicted channel information for at least one dedicated timing; and reporting at least one performance metric of the processing procedure for at least one dedicated timing to the radio access network node 102 by the wireless terminal device 104.

[0056] Similarly, the method performed by the radio access network node 102 may include receiving at least one performance metric of a processing procedure at least one timing from the wireless terminal device 104. In various examples, a timing includes at least one time unit, time slot, or symbol. The method may also include the radio access network node 102 indicating to the wireless terminal device 104 an indication of at least one dedicated timing for indicative of the performance of the processing procedure, and receiving at least one performance metric from the wireless terminal device 104 at least one dedicated timing.

[0057] In another example, such as Figure 7 As shown, UE 104 can calculate performance metrics for actual and predicted channel measurements at certain times and report these metrics to the network (e.g., base station 102) for performance monitoring and subsequent decision-making regarding the current AI / ML functions / models. For example, using SGCS as a performance metric, the SGCS values ​​for the first four monitoring times could be 0.5, 0.5, 0.45, and 0.43. UE 104 may not need to report all results for each monitoring time to reduce overhead, as the performance is very poor based on the first two results, and the current AI / ML model cannot function well, while reporting the last two results would be redundant.

[0058] According to various embodiments, a method performed by a wireless terminal device 104 may include receiving indications from a radio access network node 102 of a plurality of dedicated timings for indicative of the performance of a processing procedure. The method may further include: determining, by the wireless terminal device 104, at least in part based on actual channel measurement information and predicted channel information of a subset of the plurality of dedicated timings; and reporting, by the wireless terminal device 104, the at least one performance metric of the subset of the plurality of dedicated timings to the radio access network node 102. Similarly, a method performed by a radio access network node 102 may include: indicating to the wireless terminal device 104 the indications of a plurality of dedicated timings for indicative of the performance of a processing procedure; and receiving, by the wireless terminal device 104, at least one performance metric of a subset of the plurality of dedicated timings.

[0059] In another example, such as Figure 8 As shown, UE 104 can calculate performance metrics for the actual channel measurement and predicted channel measurement for each of multiple time points, and report the average performance metric for each of the multiple time points to the network (e.g., base station 102) so that the network can perform performance monitoring and make subsequent decisions on the current AI / ML function / model. For example, SGCS can be used as the performance metric, and the SGCS values ​​for the first four monitoring time points can be 0.9, 0.9, and 0.85. Then, UE 104 can report the average of the three monitoring time points to the network, which is 0.88.

[0060] According to various embodiments, a method performed by a wireless terminal device 104 may include receiving indications from a radio access network node 102 of a plurality of dedicated timeframes for indicating the performance of a processing procedure. The method may further include: the wireless terminal device 104 determining at least one performance metric of the processing procedure for the plurality of dedicated timeframes as an average performance metric of the plurality of dedicated timeframes, based at least in part on actual channel measurement information and predicted channel information of the plurality of dedicated timeframes; and the wireless terminal device 104 reporting the average performance metric of the plurality of dedicated timeframes to the radio access network node 102. Similarly, a method performed by a radio access network node 102 may include: indicating to the wireless terminal device 104 the indications of a plurality of dedicated timeframes for indicating the performance of a processing procedure; and receiving the average performance metric of the plurality of dedicated timeframes from the wireless terminal device 104.

[0061] In another example, such as Figure 9As shown, UE 104 can calculate performance metrics for the actual and predicted channel measurements of a time group and report the average performance metric for each time group to the network (e.g., base station 102) so that the network can monitor performance and make subsequent decisions regarding the current AI / ML functions / models. For example, SGCS can be used as the performance metric, and the SGCS values ​​for the first six monitoring times can be 0.9, 0.9, 0.85, 0.8, 0.75, and 0.7. The six monitoring times can be divided into, for example, two groups, and UE 104 can report the average value for each time group to the network; in this example, the average values ​​are 0.88 and 0.75, respectively.

[0062] According to various embodiments, a method performed by a wireless terminal device 104 may include receiving an indication from a radio access network node 102 of a plurality of dedicated time slots for indicating the performance of a processing procedure. The method may further include: determining, at least one performance metric of the processing procedure for the plurality of dedicated time slots as at least one average performance metric of at least one group of the plurality of dedicated time slots, based at least in part on actual channel measurement information and predicted channel information of the plurality of dedicated time slots; and reporting, by the wireless terminal device 104, at least one average performance metric of the at least one group of the plurality of dedicated time slots to the radio access network node 102. Similarly, a method performed by a radio access network node 102 may include: indicating, to the wireless terminal device 104, the indication of a plurality of dedicated time slots for indicating the performance of a processing procedure; and receiving, at least one average performance metric of at least one group of the plurality of dedicated time slots from the wireless terminal device 104.

[0063] In another example, such as Figure 10 As shown, UE 104 can calculate the performance metrics of the actual channel measurement and the predicted channel measurement for each / part of the time period, and report the average performance metrics on the corresponding part of the time period to the network (e.g., base station 102) so that the network can monitor performance and make subsequent decisions on the current AI / ML function / model.

[0064] According to various embodiments, a method performed by a wireless terminal device 104 may include receiving indications from a radio access network node 102 of a plurality of dedicated timings for indicative of the performance of a processing procedure. The method may further include: determining, at least one performance metric of the processing procedure of a subset of the plurality of dedicated timings as at least one average performance metric of the subset of dedicated timings, based at least in part on actual channel measurement information and predicted channel information of a subset of the plurality of dedicated timings; and reporting, by the wireless terminal device 104, the at least one average performance metric of the subset of dedicated timings to the radio access network node 102. Similarly, a method performed by a radio access network node 102 may include: indicating, by the radio access network node 102, to the wireless terminal device 104 of the indications of a plurality of dedicated timings for indicative of the performance of a processing procedure; and receiving, by the radio access network node 102, at least one average performance metric of the subset of dedicated timings from the wireless terminal device 104.

[0065] In another example, such as Figure 11 As shown, UE 104 can calculate the performance metrics of actual channel measurements and predicted channel measurements for multiple groups of timings, and report a set of average performance metrics for the corresponding partial timing groups to the network (e.g., base station 102) so that the network can monitor performance and make subsequent decisions on the current AI / ML functions / models.

[0066] According to various embodiments, a method performed by a wireless terminal device 104 may include receiving an indication from a radio access network node 102 of a plurality of dedicated timings for indicative of the performance of a processing procedure. The method may further include: determining, at least one performance metric of the processing procedure of a subset of the plurality of dedicated timings as at least one average performance metric of at least one group of at least one subset of the plurality of dedicated timings, based at least in part on actual channel measurement information and predicted channel information of a subset of the plurality of dedicated timings; and reporting, by the wireless terminal device 104, at least one average performance metric of at least one group of at least one subset of the plurality of dedicated timings to the radio access network node 102. Similarly, a method performed by a radio access network node 102 may include: indicating, to the wireless terminal device 104, the indication of a plurality of dedicated timings for indicative of the performance of a processing procedure; and receiving, at least one average performance metric of at least one group of at least one subset of the plurality of dedicated timings from the wireless terminal device 104.

[0067] In some embodiments, one or more performance metrics may include effective time. In one example, an AI / ML model is applied to CSI prediction. If UE 104 can predict CSI in future timings, it can be understood that UE 104 is capable of estimating the effective time of the predicted CSI (e.g., based on offline testing / online inference). Effective time can be evaluated by performance metrics that meet some thresholds. For example, if the SCGS value exceeds a threshold of 0.8, the AI / ML model is effective in monitoring timings. Then, effective time can be the statistical length of the effective prediction timing / slot. Therefore, UE 104 can report the effective time in the CSI report to the network (e.g., base station 102) to identify the performance of the AI / ML model and further configuration. To ensure that the network (e.g., base station 102) can trust the metrics of effective time reported by UE 104, some coordination may be needed between UE 104 and the network (e.g., base station 102) to confirm the validity of the thresholds.

[0068] In some embodiments, performance metrics(s) can be evaluated at different granularities. In one approach, the network (e.g., base station 102) can indicate the granularity of the performance metric calculation to the UE 104. In another approach, the UE 104 can select the granularity of the performance metric calculation and report to the network (e.g., base station 102) which granularity the performance metric is based on. In some embodiments, the granularity includes temporal granularity (e.g., OFDM symbol, single timeslot / multiple timeslots, etc.). In some embodiments, the granularity includes spatial granularity (e.g., antenna port / port group, etc.). In some embodiments, the granularity includes frequency granularity (e.g., resource element (RE), resource element group (REG), physical resource block (PRB), subband, etc.).

[0069] According to various embodiments, a method performed by a wireless terminal device 104 may include determining the validity period of predicted channel information. The method may include the wireless terminal device 104 determining the validity period of the predicted channel information based at least in part on a performance metric. The method may include the wireless terminal device 104 determining a performance metric at the granularity of a performance metric calculation. Similarly, a method performed by a radio access network node 102 may include receiving the validity period of predicted channel information determined by the processing of the wireless terminal device from the wireless terminal device 104. The method may also include the radio access network node 102 instructing the wireless terminal device 104 on the granularity of the performance metric calculation to be performed by the wireless terminal device.

[0070] In some embodiments, UE 104 may report channel-related information to the network side (e.g., base station 102). In some embodiments, UE 104 may report actual channel measurements and predicted channel measurements at specific times in the measurement report. In one example, an AI / ML model may be applied to CSI prediction. In some embodiments, actual channel measurements may include a conventional codebook-based precoding matrix indicator (PMI). In some embodiments, actual channel measurements may include high-resolution channel information. In one example, a high-resolution quantization method (e.g., scalar quantization / vector quantization), and in another example, a high-resolution codebook-based PMI (e.g., a codebook with a new combination of parameters).

[0071] According to various embodiments, a method performed by a wireless terminal device 104 may include reporting channel-related information to a radio access network node 102. The method may include the wireless terminal device 104 generating predicted channel information using a processing procedure, wherein reporting the channel-related information includes reporting the predicted channel information and actual channel measurement information to the radio access network node 102. Similarly, a method performed by a radio access network node 102 may include receiving channel-related information from the wireless terminal device 104. The method may also include the radio access network node 102 receiving the predicted channel information and actual channel measurement information from the wireless terminal device 104, wherein the predicted channel information is generated by the wireless terminal device using a processing procedure.

[0072] In some embodiments, the difference between the actual channel measurement and the predicted channel measurement may be reported in the measurement report along with the predicted channel measurement at a specific time (or along with the actual channel measurement). In some embodiments, the difference information may be evaluated using an error function, such as Euclidean distance, MSE, NMSE, etc. In some embodiments, the difference information may be evaluated using a similarity function, such as GCS, SGCS, etc.

[0073] According to various embodiments, the method performed by the wireless terminal device 104 may include: generating predicted channel information using a processing procedure; wherein reporting channel-related information includes: reporting the difference between actual channel measurement information and predicted channel information to the radio access network node 102; and reporting the predicted channel information to the radio access network node 102. Similarly, the method performed by the radio access network node 102 may include receiving predicted channel information and the difference between actual channel measurement information and predicted channel information from the wireless terminal device 104, wherein the predicted channel information is generated by the wireless terminal device using a processing procedure.

[0074] In some embodiments, the measurement report only reports the actual channel measurements at one or more dedicated moments that indicate the performance of the processing procedure. The network (e.g., base station 102) can then calculate one or more performance metrics based on the actual channel measurements at one or more dedicated moments and channel measurements at one or more previous moments to identify channel variation information for network monitoring of processing performance. In some embodiments, the channel variation information can be evaluated using an error function (e.g., Euclidean distance, MSE, NMSE, etc.). In some embodiments, the channel variation information can be evaluated using a similarity function (e.g., GCS, SGCS, etc.). In various examples, the network (e.g., base station 102) can perform any of the methods discussed above for calculating one or more performance metrics performed by UE 104.

[0075] According to various embodiments, the method performed by the wireless terminal device 104 may include measuring actual channel measurement information at at least one dedicated time to indicate the performance of the processing procedure, and reporting the actual channel measurement information for the at least one dedicated time to the radio access network node 102. Similarly, the method performed by the radio access network node 102 may include: receiving actual channel measurement information for at least one dedicated time from the wireless terminal device 104 to indicate the performance of the processing procedure; and calculating at least one performance metric of the processing procedure based at least in part on the actual channel measurement information.

[0076] In some embodiments, the measurement report only reports predicted channel measurements for one or more dedicated moments that indicate the performance of the processing procedure. The network (e.g., base station 102) can then calculate one or more performance metrics based on the predicted channel measurements for one or more dedicated moments and channel measurements for one or more previous moments to identify channel change information for network monitoring of the processing method's performance. In some embodiments, the channel change information can be evaluated using an error function (e.g., Euclidean distance, MSE, NMSE, etc.). In some embodiments, the channel change information can be evaluated using a similarity function (e.g., GCS, SGCS, etc.). In various examples, the network (e.g., base station 102) can perform any of the methods discussed above for calculating one or more performance metrics performed by UE 104.

[0077] According to various embodiments, a method performed by a wireless terminal device 104 may include: generating predicted channel information for at least one dedicated timing, indicating the performance of the processing procedure, using a processing procedure performed by the wireless terminal device; and reporting the predicted channel information for at least one dedicated timing to a radio access network node 102. Similarly, a method performed by a radio access network node 102 may include: receiving predicted channel information for at least one dedicated timing, indicating the performance of the processing procedure, from the wireless terminal device 104; and calculating at least one performance metric of the processing procedure based at least in part on the predicted channel information, wherein the processing procedure is performed by the wireless terminal device.

[0078] In some embodiments, UE 104 may report actual channel measurements and compressed channel information at specific times in the measurement report. In one example, an AI / ML model may be applied to CSI compression. In some embodiments, actual channel measurements include conventional codebook-based precoding matrix indicators (PMIs). In some embodiments, actual channel measurements include high-resolution channel information. In one example, a high-resolution quantization method (e.g., scalar quantization / vector quantization) or a high-resolution codebook-based PMI (e.g., a codebook with a new combination of parameters) is utilized.

[0079] In some embodiments, UE 104 may report some channel-related information to the network side (e.g., to base station 102).

[0080] In one example, an AI / ML model can be applied to CSI prediction. The UE can calculate performance metrics, such as SGCS values, based on actual and predicted channel measurements across four monitoring points, for example, values ​​of 0.9, 0.85, 0.65, and 0.6 respectively. Based on the performance metrics, UE 104 can observe poor prediction performance for the two most recent points, and UE 104 may choose not to report channel information for these two points, as inaccurate CSI can harm network system performance. Furthermore, UE 104 can save some CSI feedback overhead by reporting only partial channel-related information. Note that UE 106 can explicitly / implicitly indicate to the network the number of channel-related information reporting points.

[0081] In one example, the AI / ML model is again applied to CSI prediction. After some initial rounds of reporting, the network (e.g., base station 102) calculates an average performance metric, such as an SGCS value, based on the reported actual and predicted channel measurements from four monitoring times. This metric could be 0.9, 0.85, 0.65, and 0.6, respectively. Based on the performance metric, the network (e.g., base station 102) can observe poor overall prediction performance for the two most recent times, and can instruct UE 104 not to report channel information about the two most recent times in future monitoring sessions, since inaccurate CSI is useless to the network and there is no need to waste uplink resources reporting channel information. Furthermore, the network (e.g., base station 102) can modify resource allocation for UE 104 based on the monitoring results.

[0082] According to various embodiments, the method performed by the wireless terminal device 104 may include reporting channel-related information by reporting some channel-related information to the radio access network node 102. Similarly, the method performed by the radio access network node 102 may include receiving some channel-related information from the wireless terminal device 104.

[0083] In some embodiments, UE 104 may report channel-related information and performance metrics to the network side (e.g., to base station 102). In one example, an AI / ML model may be applied to CSI prediction. UE 104 may calculate performance metrics, such as SGCS values, based on actual and predicted channel measurements at four monitoring points, for example, values ​​of 0.9, 0.85, 0.65, and 0.6 respectively. To reduce feedback overhead, UE 104 may report predicted channel measurements at one or more monitoring points and performance metrics at that point / these points to the network to identify the performance of the AI / ML model. This way, UE 104 does not need to report actual channel measurements in its measurement report, which would otherwise incur significant feedback overhead.

[0084] According to various embodiments, the method performed by the wireless terminal device 104 may include: generating predicted channel information using a processing procedure, wherein reporting channel-related information includes: reporting the predicted channel information and at least one performance metric to the radio access network node 102. Similarly, the method performed by the radio access network node 102 may include: receiving predicted channel information and at least one performance metric from the wireless terminal device 104, wherein the predicted channel information is generated by the wireless terminal device 104 using a processing procedure.

[0085] In some embodiments, the network (e.g., base station 102) can dynamically configure and / or modify the configuration based on the results of performance monitoring. In one approach, the network (e.g., base station 102) can dynamically configure / modify the prediction timing / prediction window length via signaling. For example, UE 104 can report performance metrics (e.g., SGCS values) for multiple monitoring moments to the network. The network can infer performance changes based on the values ​​of these metrics. In one example, the SGCS values ​​reported by UE 104 in four monitoring moments might be, for example, 0.9, 0.85, 0.7, and 0.6, respectively. The network (e.g., base station 102) can determine that the current AI / ML model is not predicting well in further prediction moments. Therefore, the network can reduce the prediction timing / prediction window length of the ongoing AI / ML model used by UE 104 via signaling to ensure prediction performance.

[0086] In another approach, the network (e.g., a base station) can dynamically configure and / or modify the period of prediction timing via signaling. For example, UE 104 can report performance metrics (e.g., SGCS values) for multiple monitoring moments to the network. The network can infer performance changes from the values ​​of these metrics. In one example, the SGCS values ​​reported by the UE for four monitoring moments with a period of 5ms might be, for example, 0.9, 0.85, 0.7, and 0.6. The network (e.g., the base station) can determine that the current AI / ML model is not predicting well in further prediction moments. Therefore, the network can reduce the period of the prediction moment for the ongoing AI / ML model used by UE 104 via signaling (e.g., 2.5ms) to ensure prediction performance.

[0087] In some embodiments, dynamic configuration can be configured via Media Access Control Element (MAC-CE) signaling based on the results of performance monitoring and can be indicated to UE 104.

[0088] In some embodiments, dynamic configuration can be configured via downlink control information (DCI) signaling based on the results of performance monitoring and can be indicated to UE 104.

[0089] According to various embodiments, the method performed by the wireless terminal device 104 may include receiving a configuration of a processing procedure determined by the radio access network node 102, at least in part, based on at least one performance metric. In a particular embodiment, the configuration includes the number of prediction opportunities, the length of a prediction window, or the length of the time span of the processing procedure of the wireless terminal device. In other embodiments, the configuration includes the period of the prediction opportunities of the processing procedure of the wireless terminal device. The method may include the wireless terminal device 104 receiving the configuration of the processing procedure from the radio access network node 102 via at least one of: a Media Access Control Element (MAC-CE) or Downlink Control Information (DCI) signaling.

[0090] Similarly, the method performed by the radio access network node 102 may include: configuring the processing procedure of the wireless terminal device at least in part based on at least one performance metric; and instructing the wireless terminal device 104 of the configuration of the processing procedure. In various embodiments, the configuration includes configuring the number of prediction opportunities, the length of the prediction window, or the time span of the processing procedure of the wireless terminal device by the radio access network node 102. In various embodiments, the configuration includes configuring the period of the prediction opportunities of the processing procedure of the wireless terminal device by the radio access network node 102. The method may also include the radio access network node 102 instructing the wireless terminal device 104 of the configuration of the processing procedure via at least one of the following: Media Access Control Element (MAC-CE), or Downlink Control Information (DCI) signaling.

[0091] In some embodiments, the network (e.g., base station 102) makes one or more decisions regarding the processing procedure based on performance monitoring results. In some embodiments, the network may need to make decisions regarding the processing procedure based on monitored performance to ensure system performance / efficiency. For example, if the processing procedure is enabled by an AI / ML model, performance may significantly degrade when the model's generalization ability is poor. Therefore, appropriate decisions made by the network can mitigate this situation. When the network (e.g., base station 102) identifies that the performance of the current AI / ML model remains poor over a period of time, model switching / updating may be necessary to achieve better performance. Otherwise, falling back to a traditional mechanism can at least maintain acceptable performance. In some embodiments, if the processing procedure is enabled by an AI / ML model / feature, the decisions at least include model / feature selection / switching / updating / activation / deactivation / fallback to a traditional approach.

[0092] In some embodiments, performance monitoring of processes involving UE-side AI / ML models or processes enabled by UE-side AI / ML models is performed on the UE side. In some embodiments, UE 104 reports an indication to the network side. In some embodiments, this indication includes indicators regarding the results of performance monitoring. The results of performance monitoring may be influenced by performance metrics at one or more monitoring times.

[0093] In one example, UE 104 may report a 1-bit indicator in its measurement report to the network (e.g., base station 102) to indicate whether the AI / ML model is valid. A value of 0 indicates that the current AI / ML model is invalid; a value of 1 indicates that the AI / ML model is valid, and vice versa. In another example, UE 104 may report a 1-bit indicator in its measurement report to the network to indicate whether the AI / ML model performance meets a specific threshold. A value of 0 indicates that the current AI / ML model performance metric does not meet the threshold; a value of 1 indicates that the AI / ML model performance metric meets the threshold. In some embodiments, the threshold may be configured by the network (e.g., base station 102) and indicated to UE 104. UE 104 and network 102 need to reach a consensus on the threshold.

[0094] In another example, UE 104 may, for instance, report a 1-bit indicator and an additional value for a threshold to the network (e.g., base station 102) in a measurement report to indicate whether the AI / ML model performance meets a specific threshold. A value of 0 indicates that the current AI / ML model performance metric fails to meet the threshold; and a value of 1 indicates that the AI / ML model performance metric meets the threshold. In some embodiments, the threshold is set by UE 104 and is transparent to the network. Therefore, the value of the threshold should also be reported to the network to help ensure that the network is aware of and trusts the results of performance monitoring. Similarly, consensus on the threshold is required between the UE and the network.

[0095] According to various embodiments, a method performed by a wireless terminal device 104 may include performance monitoring of a processing procedure performed by the wireless terminal device; and reporting an indication regarding the performance monitoring to a radio access network node 102. In a particular example, the indication includes an indicator regarding the result of the performance monitoring. The method may include the wireless terminal device 104 performing performance monitoring of the processing procedure at least in part based on a performance threshold. The method may include the wireless terminal device 104 reporting an indication regarding the performance monitoring to the radio access network node 102, the indication including a one-bit indication of whether the processing procedure is valid. The method may include the wireless terminal device 104 receiving a performance threshold from the radio access network node 102. The method may include the wireless terminal device 104 reporting the performance threshold to the radio access network node 102. The method may include the wireless terminal device 104 determining the performance threshold.

[0096] Similarly, the method performed by the radio access network node 102 may include receiving an indication from the wireless terminal device 104 regarding performance monitoring of a processing procedure, wherein the processing procedure is performed by the wireless terminal device, and wherein the performance monitoring is performed by the wireless terminal device. In a particular embodiment, the indication includes an indicator regarding the result of the performance monitoring. The method may also include the radio access network node 102 receiving an indication from the wireless terminal device 104 regarding performance monitoring, the indication including an indication of whether the processing procedure is valid. The method may also include the radio access network node 102 receiving an indication from the wireless terminal device 104 regarding performance monitoring, the indication including an indication of whether the processing procedure meets a performance threshold. The method may also include the radio access network node 102 indicating to the wireless terminal device 104 a performance threshold used by the wireless terminal device in performance monitoring of the processing procedure. In a particular embodiment, the indication of whether the processing procedure meets a performance threshold includes a one-bit indication of whether the processing procedure is valid, at least in part based on the performance threshold. The method may also include the radio access network node 102 receiving a performance threshold from the wireless terminal device 104.

[0097] In some embodiments, the indication includes an indicator about the current performance level. In one example, UE 104 may report a K-bit indicator of the performance level to the network (e.g., base station 102) in a measurement report to indicate the current performance level of the AI / ML function / model. The performance level may be associated with a performance metric, which may include at least intermediate KPIs and a final KPI. For example, the mapping between performance levels and performance metrics can be shown in the following table, which is known to both UE 104 and the network. For example, Table 1 below shows an example mapping between performance levels and SGCS:

[0098] Table 1 Table 1 above is just an example, and different performance metrics / performance metrics at different granularities / mapping relationships between indicators and performance metrics also apply.

[0099] According to various embodiments, the method performed by the wireless terminal device 104 may include reporting an indication regarding performance monitoring to the radio access network node 102, the indication including an indication of the current performance level of the processing. Similarly, the method performed by the radio access network node 102 may include receiving an indication regarding performance monitoring from the wireless terminal device 104, the indication including an indication of the current performance level of the processing. In a particular embodiment, the indication of the current performance level of the processing includes a plurality of bits for indicating the current performance level.

[0100] In some embodiments, the indication includes an indicator regarding decision recommendations to the network. In one example, UE 104 may report a 1-bit indicator to the network (e.g., base station 102) in a measurement report to suggest whether further decisions should be made regarding the AI / ML function / model. A value of 0 may indicate that the current AI / ML function / model is recommended to continue using, and no further decisions need to be made regarding the current AI / ML function / model. A value of 1 may indicate that UE 104 recommends that the network make some decisions regarding the current AI / ML function / model. For example, the 1-bit indicator may also implicitly indicate the effectiveness of the current AI / ML model's performance; for instance, a value of 0 may implicitly indicate that the current AI / ML function / model is performing well, while a value of 1 may implicitly indicate that the current AI / ML function / model may experience performance degradation. In some embodiments, function / model decisions include at least selecting / switching / updating / activating / deactivating / falling back to a legacy scheme.

[0101] In one example, UE 104 may report an indicator in a field of X bits in the measurement report to the network to suggest specific follow-up decisions that should be made regarding the AI / ML function / model. For example, the indicator table in Table 2 below may include a variety of decision options known to both UE 104 and the network. For instance, if the indicator is 000, it indicates that the UE is suggesting to the network that it make a decision to fall back to a legacy approach.

[0102]

[0103] Table 2 The table above is just an example, and different mappings between indicators and functions / models may also apply.

[0104] According to various embodiments, a method performed by a wireless terminal device 104 may include reporting a recommendation indication to a radio access network node 102, the recommendation indication including a recommendation on whether the radio access network node should perform a decision process regarding the processing procedure. The method may include the wireless terminal device 104 reporting the recommendation indication to the radio access network node 102, the recommendation indication including multiple bits indicating a specific recommended decision regarding the processing procedure. Similarly, a method performed by a radio access network node 102 may include receiving a recommendation indication from the wireless terminal device 104, the recommendation indication including a recommendation on whether the radio access network node should perform a decision process regarding the processing procedure. The method may also include the radio access network node 102 receiving a recommendation indication from the wireless terminal device 104, the recommendation indication including multiple bits indicating a specific recommended decision regarding the processing procedure.

[0105] In some embodiments, the network (e.g., base station 102) makes one or more decisions regarding the processing procedure based on performance monitoring results. In some embodiments, the network needs to make decisions regarding the processing procedure based on monitored performance to ensure system performance / efficiency. For example, if the processing procedure is enabled by an AI / ML model, performance may significantly degrade when the model's generalization ability is poor. Therefore, appropriate decisions made by the network can mitigate this situation. When the network identifies that the current AI / ML model's performance on a specific function remains poor over a period of time, model / function switching / updating may be necessary to achieve better performance. Otherwise, falling back to a traditional mechanism can at least maintain acceptable performance. In some embodiments, if the processing procedure is enabled by an AI / ML model / function, the decisions may at least include model / function selection / switching / updating / activation / deactivation / fallback to a traditional approach.

[0106] According to various embodiments, the method performed by the wireless terminal device 104 may include receiving a configuration of a processing procedure from the wireless access network node 102, wherein the configuration is at least partially based on instructions regarding performance monitoring. Similarly, the method performed by the wireless access network node 102 may include determining the operational procedures of the processing procedure of the wireless terminal device based on instructions regarding performance monitoring. The method may also include the wireless access network node 102 configuring the processing procedure of the wireless terminal device based on instructions regarding performance monitoring; and instructing the wireless terminal device 104 on the configuration of the processing procedure.

[0107] In some embodiments, UE 104 may make one or more decisions regarding the processing procedure based on performance monitoring results. In these embodiments, UE 104 may make one or more decisions regarding the processing procedure based on monitored performance to ensure system performance / efficiency. For example, if the processing procedure is enabled by an AI / ML model, performance may significantly degrade when the model's generalization ability is poor. Therefore, appropriate decisions made by UE 104 can quickly mitigate this situation. Within the same AI / ML function, UE 104 may have multiple models that can implement the same function (e.g., the same model configuration / applicable scenarios, etc.). When UE 104 identifies that the performance of the current AI / ML model remains poor over a period of time, decisions regarding one or more models within the same function (e.g., model switching) can be transparent to the network to achieve rapid performance compensation. Otherwise, when the model / function is not feasible, UE 104 does not wish to make one or more decisions regarding the model / function.

[0108] In some embodiments, UE 104 may make one or more decisions regarding the processing based on monitored performance and report these decisions to the network to ensure system performance / efficiency. For example, if the processing is enabled by an AI / ML model, performance may significantly degrade when the model's generalization ability is poor. Therefore, appropriate decisions made by UE 104 can quickly mitigate this situation. Within the same AI / ML function, the UE may have multiple models that perform the same function (e.g., the same model configuration / applicable scenarios, etc.). When UE 104 identifies that the performance of the current AI / ML model remains poor over a period of time, it may make one or more decisions (e.g., model switching) regarding the model within the same function to achieve rapid performance compensation. UE 104 may then report these decisions to the network so that the network is aware of the performance changes in the AI / ML model. Otherwise, UE 104 does not wish to make one or more decisions regarding the model / function when the model / function is not feasible.

[0109] In some embodiments, if the process is enabled by an AI / ML model / function, the decision includes at least model / function selection / switching / updating / activation / deactivation / rollback to a traditional approach.

[0110] According to various embodiments, the method performed by the wireless terminal device 104 may include determining the action procedures for the processing of the wireless terminal device based on performance monitoring of the processing. The method may include the wireless terminal device 104 reporting the action procedures determined by the wireless terminal device to the radio access network node 102. Similarly, the method performed by the radio access network node 102 may include receiving the action procedures for the processing of the wireless terminal device 104 determined by the wireless terminal device 104 based on performance monitoring of the processing.

[0111] The above description and accompanying drawings provide specific example embodiments and implementations. However, the described subject matter can be embodied in a variety of different forms, and therefore, the covered or claimed subject matter is intended to be construed as not being limited to any of the example embodiments set forth herein. The claimed or covered subject matter is intended to have a reasonably broad scope. Among other things, the subject matter can be implemented as a method, apparatus, component, system, or non-transitory computer-readable medium for storing computer code. Accordingly, embodiments can take the form of, for example, hardware, software, firmware, storage medium, or any combination thereof. For example, the above-described method embodiments can be implemented by a component, apparatus, or system including a memory and a processor by executing computer code stored in the memory.

[0112] Throughout the specification and claims, terms may have nuanced meanings suggested or implied in the context beyond their explicitly stated meanings. Similarly, the phrase "in one embodiment / implementation / example / method" as used herein does not necessarily refer to the same embodiment, and the phrase "in another embodiment / implementation / example / method" as used herein does not necessarily refer to different embodiments. For example, the claimed subject matter is intended to include, in whole or in part, combinations of exemplary embodiments.

[0113] Generally, terms can be understood at least in part from their usage in the context. For example, terms such as “and,” “or,” or “and / or” as used herein can include a variety of meanings that can depend at least in part on the context in which they are used. Typically, “or,” when used to associate a list such as A, B, or C, is intended to mean A, B, and C (inclusive meaning) and A, B, or C (in exclusive meaning). Furthermore, depending at least in part on the context, the term “one or more,” as used herein, can be used to describe any feature, structure, or characteristic in a singular sense, or can be used to describe a combination of features, structures, or characteristics in a plural sense. Similarly, depending at least in part on the context, terms such as “a,” “an,” or “the” can be understood to convey either a singular or a plural usage. Additionally, the term “based on,” depending at least in part on the context, can be understood to not necessarily convey an exclusive set of factors, but rather to allow for the presence of additional factors that are not necessarily explicitly described.

[0114] References to features, advantages, or similar language throughout this specification do not imply that all features and advantages achievable using this solution should be, or be included, in any single implementation thereof. Rather, references to such features and advantages are to be understood as indicating that a particular feature, advantage, or characteristic described in connection with an embodiment is included in at least one embodiment of this solution. Therefore, throughout this specification, discussions of features and advantages and similar language may, but do not necessarily, refer to the same embodiment.

[0115] Furthermore, the features, advantages, or characteristics described in this solution can be combined in one or more embodiments in any suitable manner. Based on the description herein, those skilled in the art will recognize that this solution can be practiced without one or more of the specific features or advantages of a particular embodiment. In other instances, additional features and advantages may be recognized in certain embodiments that may not be present in all embodiments of this solution.

Claims

1. A method for performance monitoring of a processing procedure, comprising: At least one performance metric of the processing procedure is determined by the wireless terminal device, wherein the processing procedure is performed by the wireless terminal device; and The wireless terminal device reports the at least one performance metric to the wireless access network node.

2. The method according to claim 1, comprising: The wireless terminal device measures the actual channel measurement information; The wireless terminal device generates predictive channel information using the processing procedure. as well as The wireless terminal device calculates the at least one performance metric based at least in part on the predicted channel information and the actual channel measurement information.

3. The method according to any one of claims 1 to 2, comprising: At least one performance metric of the processing procedure in which at least one timing is determined by the wireless terminal device; as well as The wireless terminal device reports at least one performance metric at the at least one time to the wireless access network node.

4. The method according to claim 3, in, The timing includes at least one unit of time, time slot, or symbol.

5. The method according to any one of claims 3 to 4, comprising: The wireless terminal device receives at least one dedicated timing indication from the wireless access network node for indicating the performance of the processing procedure; The wireless terminal device determines at least one performance metric of the processing procedure for the at least one dedicated time slot, based at least in part on actual channel measurement information and predicted channel information of the at least one dedicated time slot. as well as The wireless terminal device reports the at least one performance metric for the at least one dedicated time to the wireless access network node.

6. The method according to any one of claims 3 to 4, comprising: The wireless terminal device receives indications from the wireless access network node at multiple dedicated points for indicating the performance of the processing procedure. The wireless terminal device determines at least one performance metric of the processing of the subset of the plurality of dedicated time slots, based at least in part on actual channel measurement information and predicted channel information of the subset of the plurality of dedicated time slots; as well as The wireless terminal device reports at least one performance metric of a subset of the plurality of dedicated time slots to the wireless access network node.

7. The method according to any one of claims 3 to 4, comprising: The wireless terminal device receives indications from the wireless access network node at multiple dedicated points for indicating the performance of the processing procedure. The wireless terminal device determines at least one performance metric of the processing of the multiple dedicated time slots as the average performance metric of the multiple dedicated time slots, based at least in part on the actual channel measurement information and predicted channel information of the multiple dedicated time slots. as well as The wireless terminal device reports the average performance metrics of the multiple dedicated time slots to the wireless access network node.

8. The method according to any one of claims 3 to 4, comprising: The wireless terminal device receives indications from the wireless access network node at multiple dedicated points for indicating the performance of the processing procedure. The wireless terminal device determines at least one performance metric of the processing of the plurality of dedicated time slots as at least one average performance metric of at least one group of the plurality of dedicated time slots, based at least in part on actual channel measurement information and predicted channel information of the plurality of dedicated time slots. as well as The wireless terminal device reports at least one average performance metric for at least one group of the plurality of dedicated timings to the wireless access network node.

9. The method according to any one of claims 3 to 4, comprising: The wireless terminal device receives indications from the wireless access network node at multiple dedicated points for indicating the performance of the processing procedure. The wireless terminal device determines at least one performance metric of the processing of the subset of the plurality of dedicated time slots as at least one average performance metric of the subset of the plurality of dedicated time slots, based at least in part on actual channel measurement information and predicted channel information of the subset of the plurality of dedicated time slots. as well as The wireless terminal device reports at least one average performance metric of a subset of the plurality of dedicated time slots to the wireless access network node.

10. The method according to any one of claims 3 to 4, comprising: The wireless terminal device receives indications from the wireless access network node at multiple dedicated points for indicating the performance of the processing procedure. The wireless terminal device determines at least one performance metric of the processing of the subset of the plurality of dedicated time slots as at least one average performance metric of at least one group of the subset of the plurality of dedicated time slots, based at least in part on actual channel measurement information and predicted channel information of the subset of the plurality of dedicated time slots. as well as The wireless terminal device reports at least one average performance metric of at least one group of at least one subset of the plurality of dedicated time slots to the wireless access network node.

11. The method according to any one of claims 2 to 10, comprising: The effective time of the predicted channel information is determined by the wireless terminal device.

12. The method of claim 11, comprising: The effective time of the predicted channel information is determined by the wireless terminal device, at least in part, based on the performance metric.

13. The method according to any one of claims 2 to 12, comprising: The performance metric is determined by the wireless terminal device at a granularity calculated using the performance metric.

14. The method according to any one of claims 1 to 13, comprising: The wireless terminal device reports channel-related information to the wireless access network node.

15. The method of claim 14, comprising: The wireless terminal device generates predictive channel information using the processing procedure. The channel-related information reported includes: The wireless terminal device reports the predicted channel information and the actual channel measurement information to the wireless access network node.

16. The method of claim 14, comprising: The wireless terminal device generates predictive channel information using the processing procedure. The channel-related information reported includes: The wireless terminal device reports the difference between the actual channel measurement information and the predicted channel information to the wireless access network node; and The predicted channel information is reported by the wireless terminal device to the wireless access network node.

17. The method according to claim 14, in, The channel-related information mentioned in the report includes: The wireless terminal device reports some channel-related information to the wireless access network node.

18. The method of claim 14, comprising: The wireless terminal device generates predictive channel information using the processing procedure. The channel-related information reported includes: The wireless terminal device reports the predicted channel information and the at least one performance metric to the wireless access network node.

19. The method according to any one of claims 1 to 18, comprising: The wireless terminal device receives a configuration of the processing procedure from the wireless access network node, which is determined by the wireless access network node at least in part based on the at least one performance metric.

20. The method according to claim 19, in, The configuration includes the number of prediction opportunities, the length of the prediction window, or the length of the time span of the processing procedure of the wireless terminal device.

21. The method according to any one of claims 19 to 20, in, The configuration includes the period for predicting the timing of the processing of the wireless terminal device.

22. The method according to any one of claims 19 to 21, comprising: The wireless terminal device receives the configuration of the processing from the wireless access network node via at least one of the following: Media Access Control Element (MAC-CE) or Downlink Control Information (DCI) signaling.

23. The method according to any one of claims 1 to 22, in, The processing includes processing using artificial intelligence (AI) or machine learning (ML) models.

24. A method for monitoring the performance of a processing procedure, comprising: The wireless access network node receives at least one performance metric of the processing procedure from the wireless terminal device, wherein the processing procedure is performed by the wireless terminal device.

25. The method of claim 24, comprising: At least one performance metric of the processing procedure in which the wireless access network node receives at least one instance of information from the wireless terminal device.

26. The method according to claim 25, in, The timing includes at least one unit of time, time slot, or symbol.

27. The method according to any one of claims 25 to 26, comprising: The wireless access network node instructs the wireless terminal device to indicate at least one dedicated timing for indicative of the performance of the processing procedure; as well as The wireless access network node receives at least one performance metric from the at least one dedicated timing of the wireless terminal device.

28. The method according to any one of claims 25 to 26, comprising: The wireless access network node instructs the wireless terminal device to provide instructions on multiple dedicated timings for indicative of the performance of the processing procedure. as well as The wireless access network node receives at least one performance metric from a subset of the plurality of dedicated times from the wireless terminal device.

29. The method according to any one of claims 25 to 26, comprising: The wireless access network node instructs the wireless terminal device to provide instructions on multiple dedicated timings for indicative of the performance of the processing procedure. as well as The average performance metric of the plurality of dedicated time slots received by the wireless access network node from the wireless terminal device.

30. The method according to any one of claims 25 to 26, comprising: The wireless access network node instructs the wireless terminal device to provide instructions on multiple dedicated timings for indicative of the performance of the processing procedure. as well as The wireless access network node receives at least one average performance metric from at least one group of the plurality of dedicated times from the wireless terminal device.

31. The method according to any one of claims 25 to 26, comprising: The wireless access network node instructs the wireless terminal device to provide instructions on multiple dedicated timings for indicative of the performance of the processing procedure. as well as The wireless access network node receives at least one average performance metric from a subset of the plurality of dedicated time slots from the wireless terminal device.

32. The method according to any one of claims 25 to 26, comprising: The wireless access network node instructs the wireless terminal device to provide instructions on multiple dedicated timings for indicative of the performance of the processing procedure. as well as The wireless access network node receives at least one average performance metric from at least one group of at least one subset of the plurality of dedicated time slots from the wireless terminal device.

33. The method according to any one of claims 24 to 32, comprising: The effective time for the wireless access network node to receive the predicted channel information determined by the processing procedure of the wireless terminal device from the wireless terminal device.

34. The method according to any one of claims 24 to 33, comprising: The wireless access network node instructs the wireless terminal device on the granularity of the performance metric calculation to be performed by the wireless terminal device.

35. The method according to any one of claims 24 to 34, comprising: The wireless access network node receives channel-related information from the wireless terminal device.

36. The method of claim 35, comprising: The wireless access network node receives predicted channel information and actual channel measurement information from the wireless terminal device, wherein the predicted channel information is generated by the wireless terminal device using the processing procedure.

37. The method of claim 35, comprising: The wireless access network node receives predicted channel information and the difference information between the actual channel measurement information and the predicted channel information from the wireless terminal device, wherein the predicted channel information is generated by the wireless terminal device using the processing procedure.

38. The method of claim 35, comprising: The wireless access network node receives some channel-related information from the wireless terminal device.

39. The method of claim 35, comprising: The wireless access network node receives predicted channel information and at least one performance metric from the wireless terminal device, wherein the predicted channel information is generated by the wireless terminal device using the processing procedure.

40. The method according to any one of claims 24 to 39, comprising: The processing procedure of the wireless terminal device is configured by the wireless access network node based at least in part on the at least one performance metric; as well as Instruct the wireless terminal device on the configuration of the processing procedure.

41. The method according to claim 40, in, The configuration includes: The wireless access network node configures the number of prediction opportunities, the length of the prediction window, or the time span of the processing procedure of the wireless terminal device.

42. The method according to any one of claims 40 to 41, in, The configuration includes: The wireless access network node configures the period for predicting the timing of the processing of the wireless terminal device.

43. The method according to any one of claims 40 to 42, comprising: The configuration of the processing is indicated to the wireless terminal device via at least one of the following: Media Access Control Element (MAC-CE) or Downlink Control Information (DCI) signaling.

44. A method for performance monitoring of a processing procedure, comprising: Performance monitoring of the processing procedure performed by the wireless terminal device, wherein the processing procedure is performed by the wireless terminal device; and The wireless terminal device reports instructions regarding the performance monitoring to the wireless access network node.

45. The method according to claim 44, in, The indications include indicators regarding the results of the performance monitoring.

46. ​​The method according to any one of claims 44 to 45, comprising: The performance monitoring of the process is performed by the wireless terminal device based at least in part on a performance threshold.

47. The method of claim 46, comprising: The wireless terminal device reports an indication regarding the performance monitoring to the wireless access network node, the indication including a one-bit indication of whether the processing is effective.

48. The method according to any one of claims 46 to 47, comprising: The performance threshold is received by the wireless terminal device from the wireless access network node.

49. The method according to any one of claims 46 to 47, comprising: The performance threshold is reported by the wireless terminal device to the wireless access network node.

50. The method according to any one of claims 46 to 47 and 49, comprising: The performance threshold is determined by the wireless terminal device.

51. The method according to any one of claims 44 to 50, comprising: The wireless terminal device reports an indication regarding the performance monitoring to the wireless access network node, the indication including an indication of the current performance level of the processing.

52. The method according to claim 51, in, The indication of the current performance level of the processing includes multiple bits used to indicate the current performance level.

53. The method according to any one of claims 44 to 52, comprising: The wireless terminal device reports a recommendation to the wireless access network node, the recommendation including a suggestion on whether the wireless access network node should perform a decision process regarding the processing procedure.

54. The method according to any one of claims 44 to 53, comprising: The wireless terminal device reports a recommendation indication to the wireless access network node, the recommendation indication including multiple bits indicating a specific recommended decision regarding the processing procedure.

55. The method according to any one of claims 44 to 54, comprising: The wireless terminal device receives a configuration of the processing procedure from the wireless access network node, wherein the configuration is at least in part based on instructions regarding the performance monitoring.

56. The method according to any one of claims 44 to 54, comprising: The wireless terminal device determines the operation process of the processing procedure based on the performance monitoring of the processing procedure.

57. The method of claim 56, comprising: The wireless terminal device reports the actions determined by the wireless terminal device regarding the processing procedure to the wireless access network node.

58. The method according to any one of claims 44 to 57, in, The processing includes processing using artificial intelligence (AI) or machine learning (ML) models.

59. A method for monitoring the performance of a processing procedure, comprising: The wireless access network node receives an instruction from the wireless terminal device regarding performance monitoring of the processing procedure, wherein the processing procedure is performed by the wireless terminal device, and wherein the performance monitoring is performed by the wireless terminal device.

60. The method according to claim 59, in, The indications include indicators regarding the results of the performance monitoring.

61. The method according to any one of claims 59 to 60, comprising: The wireless access network node receives an indication from the wireless terminal device regarding the performance monitoring, the indication including whether the processing is effective.

62. The method according to any one of claims 59 to 61, comprising: The wireless access network node receives an indication from the wireless terminal device regarding the performance monitoring, the indication including whether the processing meets a performance threshold.

63. The method of claim 62, comprising: The wireless access network node indicates to the wireless terminal device the performance threshold used by the wireless terminal device for performance monitoring during the processing.

64. The method according to claim 63, in, The indication of whether the processing meets the performance threshold includes a one-bit indication of whether the processing is effective, at least in part, based on the performance threshold.

65. The method according to any one of claims 62 to 64, comprising: The performance threshold is received by the wireless access network node from the wireless terminal device.

66. The method according to any one of claims 59 to 65, comprising: The wireless access network node receives the indication from the wireless terminal device regarding the performance monitoring, the indication including an indication of the current performance level of the processing.

67. The method according to claim 66, in, The indication of the current performance level of the processing includes multiple bits used to indicate the current performance level.

68. The method according to any one of claims 59 to 67, comprising: The wireless access network node receives a suggestion instruction from the wireless terminal device, the suggestion instruction including a suggestion on whether the wireless access network node should perform a decision process regarding the processing procedure.

69. The method according to any one of claims 59 to 68, comprising: The wireless access network node receives a suggestion instruction from the wireless terminal device, the suggestion instruction including multiple bits indicating a specific suggested decision regarding the processing procedure.

70. The method according to any one of claims 59 to 69, comprising: The wireless access network node determines the action procedure for the processing of the wireless terminal device based on the indications regarding the performance monitoring.

71. The method of claim 70, comprising: The processing procedure by which the wireless access network node configures the wireless terminal device based on the indication regarding the performance monitoring; as well as Instruct the wireless terminal device on the configuration of the processing procedure.

72. The method according to any one of claims 59 to 69, comprising: The wireless access network node receives the action process of the wireless terminal device's processing procedure from the wireless terminal device, which is determined by the wireless terminal device based on the performance monitoring of the processing procedure.

73. The method according to any one of claims 59 to 72, in, The processing includes processing using artificial intelligence (AI) or machine learning (ML) models.

74. A method for monitoring the performance of a processing procedure, comprising: The actual channel measurement information is measured by the wireless terminal device at at least one dedicated moment used to indicate the performance of the processing; as well as The wireless terminal device reports the actual channel measurement information for the at least one dedicated time slot to the wireless access network node.

75. A method for monitoring the performance of a processing procedure, comprising: The wireless access network node receives actual channel measurement information from the wireless terminal device for at least one dedicated moment, which is used to indicate the performance of the processing procedure. as well as The wireless access network node calculates at least one performance metric of the processing procedure based at least in part on the actual channel measurement information.

76. A method for monitoring the performance of a processing procedure, comprising: A wireless terminal device generates predictive channel information for at least one dedicated timing, indicating the performance of the processing procedure, using the processing procedure performed by the wireless terminal device; and The wireless terminal device reports the predicted channel information for the at least one dedicated time slot to the wireless access network node.

77. A method for monitoring the performance of a processing procedure, comprising: The wireless access network node receives predictive channel information from the wireless terminal device for at least one dedicated timing, which is used to indicate the performance of the processing. as well as The wireless access network node calculates at least one performance metric of the processing procedure based at least in part on the predicted channel information, wherein the processing procedure is performed by the wireless terminal device.

78. An apparatus for wireless communication, comprising a processor configured to perform the method according to any one of claims 1 to 77.

79. A non-transitory computer-readable medium having code stored on it, the code, when executed by a processor, causing the processor to perform the method according to any one of claims 1 to 77.

Images