Device capability and performance monitoring for models

Abstract

Techniques are described for increasing the accuracy of channel state information (CSI) using an artificial intelligence / machine learning (AI / ML) model. Due to limited model generalization and dynamic wireless environment, several methods are proposed to monitor the performance of the model. A method of wireless communication includes transmitting, by a communication device, a capability message indicating that the communication device is capable of using a model used by the communication device to perform one or more wireless communication operations.

Description

Technical Field

[0001] The present disclosure is generally directed to digital wireless communications. Background Art

[0002] Mobile telecommunication technologies are driving the world towards an increasingly connected and networked society. Compared to existing wireless networks, next-generation systems and wireless communication technologies need to support a wider range of use case characteristics and provide more complex and sophisticated access requirements and flexibility.

[0003] Long Term Evolution (LTE) is a standard for wireless communications for mobile devices and data terminals developed by the 3rd Generation Partnership Project (3GPP). LTE Advanced (LTE-A) is a wireless communications standard that enhances the LTE standard. The fifth generation wireless system (i.e. 5G) advances the LTE and LTE-A wireless standards and is committed to supporting higher data rates, a large number of connections, ultra-low latency, high reliability and other emerging business requirements. Summary of the invention

[0004] Techniques for reporting UE capabilities related to AI / ML models are disclosed. Several solutions for monitoring model performance are discussed in some embodiments.

[0005] A wireless communication method includes transmitting, by a communication device, a capability message indicating that the communication device is capable of using a model used by the communication device to perform one or more wireless communication operations.

[0006] In some embodiments, the capability message includes a time requirement of the model. In some embodiments, the time requirement is associated with a model identifier.

[0007] In some embodiments, the model may be received by the communication device from a network.

[0008] In some embodiments, the capability message includes one of: a calculation time, a calculation time offset, or a model activation time. In some embodiments, the calculation time is the time required for the communication device to operate, wherein the calculation time offset is the additional time required for the communication device to operate, and wherein the model activation time is the time required for the communication device to activate the operation.

[0009] In some embodiments, the calculation time may be a constant that is independent of the selected model. In some embodiments, the value of the calculation time offset depends on the model. In some embodiments, the capability message indicates that a measurement report is created after a time interval T based on reception of a reference signal by the communication device, wherein T depends on the calculation time offset and the calculation time.

[0010] In some embodiments, the method further comprises transmitting, by the communication device, an identifier message indicating that the model is activated.

[0011] In some embodiments, the identifier message may be sent via the PUCCH. In some embodiments.

[0012] In some embodiments, the method further comprises receiving, by the communications device, a response indicating that the identifier message was successfully received.

[0013] In some embodiments, the method further comprises deactivating, by the communication device, the model when the response is not received by the communication device.

[0014] In some embodiments, the capabilities message includes capabilities of the communication device, wherein the capabilities include capabilities for model-dependent and non-model-dependent concurrency.

[0015] Another method of wireless communication includes reporting, by a communication device, a performance indication of a model.

[0016] In some embodiments, the performance indication comprises a relationship between the assistance information and actual measurement information made by the communications device.

[0017] In some embodiments, the relationship may be a similarity or a distance, wherein the similarity comprises cosine similarity, generalized cosine similarity or squared generalized cosine similarity, and wherein the distance comprises Euclidean distance or normalized mean square error.

[0018] In some embodiments, the communications device is configured with at least one dedicated occasion for performance indication of the model.In some embodiments, the performance indication comprises predicted measurements and actual measurements of the dedicated occasion.

[0019] In some embodiments, the actual measurements include RSRPs of a plurality of reference signals, wherein the predicted measurements include predicted RSRPs of the reference signals based on the model.

[0020] In some embodiments, the actual measurement includes a codebook-based precoding matrix indication (PMI).

[0021] In some embodiments, the performance indication also includes a relationship between an output of the model and an input of the model.

[0022] In some embodiments, the performance indication comprises a relationship between MN remaining measurements of N measurements and a portion of the model output, wherein the N measurements are based on a reference signal, wherein the M measurements are based on the N measurements and are used as input to the model, wherein M<N.

[0023] In some embodiments, the relationship may be a similarity or a distance, wherein the similarity comprises cosine similarity, generalized cosine similarity or squared generalized cosine similarity, and wherein the distance comprises Euclidean distance or normalized mean square error.

[0024] In another exemplary aspect, the above method is embodied in the form of processor executable code and stored in a non-transitory computer readable storage medium. The code included in the computer readable storage medium causes the processor to implement the method described in this patent document when executed by the processor.

[0025] In yet another exemplary embodiment, a device configured or operable to perform the above method is disclosed.

[0026] The above-described aspects and other aspects and embodiments thereof are described in more detail in the drawings, the description and the claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 An example of artificial intelligence / machine learning (AI / ML) is shown.

[0028] Figure 2 Examples of calculation times and calculation time offsets are shown.

[0029] Figure 3 An example of the relationship between the auxiliary channel information and the actual channel measurement is shown.

[0030] Figure 4 An example of indicating dedicated channel measurement opportunities for monitoring model performance is shown.

[0031] Figure 5 An example of model performance monitoring using the relationship between a portion of an AI model output and an AI model input is shown.

[0032] Figure 6 An exemplary block diagram of a hardware platform is shown, which may be part of a network device or a communication device.

[0033] Figure 7 An example of wireless communication including a base station (BS) and a user equipment (UE) based on some embodiments of the disclosed technology is shown. DETAILED DESCRIPTION

[0034] The section headings used in this document are intended only to improve readability and do not limit the scope of the embodiments and techniques disclosed in each section to that section only. In addition, for ease of understanding, some embodiments are described with reference to the Third Generation Partnership Project (3GPP) New Radio (NR) standard ("5G"), while the described techniques may be implemented in different wireless systems that implement protocols other than the 5G protocol. In addition, the AI / ML model is an exemplary scenario, and the technical solutions described in this application may be generalized or applied to any model that determines the relationship between input and output.

[0035] Preliminary comments

[0036] Artificial intelligence / machine learning (AI / ML) has been studied and applied in various fields.

[0037] There have been some studies on artificial intelligence and machine learning to improve the efficiency of wireless communication systems, especially at the physical layer. For example, AI / ML models can be used to increase the accuracy of channel state information (CSI). In addition, AI / ML models can predict beam information. In order to support channel measurement using AI / ML models, UEs in the network should have corresponding capabilities.

[0038] This application discloses some processes proposed for reporting UE capabilities involving AI / ML models. Due to limited model generalization and dynamic environments, models may degrade their performance over time. Therefore, in this application, several solutions for monitoring model performance are discussed.

[0039] I. Introduction

[0040] AI / ML has been studied and used in various fields to extract features that cannot be derived by other mathematical methods. Generally speaking, AI / ML models are data-driven algorithms that apply AI / ML techniques to generate a set of outputs based on a set of inputs, including Figure 1 At least three parts are shown:

[0041] -AI model input: data fed into the AI / ML model,

[0042] -AI model output: the output of the AI / ML model, and

[0043] -AI Model: An algorithm used to derive the relationship between AI model inputs and AI model outputs.

[0044] To facilitate the discussion, the following terminology is introduced with some general descriptions.

[0045] AI model training is the process of training an AI / ML model by learning input / output relationships in a data-driven manner and obtaining a trained AI / ML model for reasoning.

[0046] AI model inference is the process of using a trained AI / ML model to produce a set of outputs based on a set of inputs.

[0047] In measurement reports, such as CSI reports or beam reports). The UE can use the AI ​​model to derive the measurement report. In order to support channel measurement by using the AI / ML model, the UE should have the corresponding capabilities.

[0048] In this application, some processes for reporting UE capabilities involving AI / ML models are proposed. In addition, due to limited model generalization and dynamic environments, models may degrade their performance over time. In other words, the data used for AI model training and the data used for AI model inference may be very different, resulting in the AI ​​model being unable to obtain the expected AI model output. In this application, several solutions for monitoring model performance are discussed.

[0049] II. Example Embodiments

[0050] A. Capability signaling or auxiliary signaling

[0051] In some embodiments, the UE may report information related to its model (e.g., AI / ML model) to the network through UE capability signaling or auxiliary information signaling.

[0052] In some embodiments, model-related information may be reported after the network sends a request message to the UE, wherein the request message may include a model identifier.

[0053] In some embodiments, model-related information may be reported without first a request message from the network.

[0054] In some embodiments, the model-related information includes time requirements for the model.

[0055] In some embodiments, the model-related information may include a model identifier for associating a corresponding time requirement.

[0056] Time requirements may include one of the following:

[0057] 1) Calculation time

[0058] In an example, the computation time may include time required for the UE to perform / operate / execute the model and / or time required for the UE to prepare a measurement report associated with the model.

[0059] In some embodiments, the computation time is the same for all models.

[0060] In some embodiments, the calculation time is different for different reported quantities in the measurement report. In other words, the calculation time depends on what is included in the measurement report. For example, the calculation time for a beam report may be different from the calculation time for a CSI report.

[0061] In some embodiments, the calculation time does not need to be reported by the UE. The calculation time reuses the value defined for non-model based measurement reporting.

[0062] 2) Calculate time offset

[0063] In an example, calculating the time offset may include the additional time required for the UE to perform / operate / execute with respect to the model.

[0064] In some embodiments, each model has its own computation time offset.

[0065] In some embodiments, the calculated time offset does not need to be reported by the UE. In this scenario, the value of the calculated time offset is set to zero by default. Alternatively, the UE can report a zero value for calculating the time offset.

[0066] In some embodiments, Figure 2 As shown, the UE may only report measurement reports based on reference signals received T time units earlier than the time when the measurement report is transmitted, where T time units are the calculation time plus the calculation time offset.

[0067] 3) Model activation / application time

[0068] In some embodiments, when the UE receives a media access control (MAC) signaling including an indication command to use a model, the UE may need some time to activate the model. For example, when the UE transmits a PUCCH with HARQ-ACK information at time n corresponding to the PDSCH carrying the indication command, the UE assumes that the corresponding model to be prepared / activated for inference should be applied at time n+model activation time.

[0069] In some embodiments, when the UE receives a downlink control indication (DCI) including an indication command to use a model, the UE may need some time to activate the model. For example, when the UE receives a PDCCH carrying an indication command at time n, the UE assumes that the corresponding model to be prepared / activated for inference should be applied at time n+model activation time.

[0070] In some embodiments, the UE may send signaling to inform the network that the model is ready / activated for inference. In some embodiments, if the UE may send signaling to inform the network that the model is ready / activated for inference and if the UE is instructed to operate using this model, the UE does not need additional time (e.g., activation time) to activate the model.

[0071] In some embodiments, signaling may be transmitted over PUCCH, where each model may be associated with a dedicated PUCCH resource.

[0072] In some embodiments, signaling may be transmitted over the PRACH, where each model may be associated with a dedicated PRACH.

[0073] In some embodiments, the UE may receive a response from the network to confirm that the signaling has been successfully received by the network. In some embodiments, the response is transmitted via MAC signaling. In some embodiments, if the UE detects a PDCCH scrambled by a dedicated RNTI, the UE confirms that the signaling has been successfully received by the network.

[0074] In some embodiments, if the UE does not receive a response, the UE may automatically deactivate the model.

[0075] In some embodiments, the UE may automatically deactivate the model when the waiting time from when the UE sends the signaling exceeds a time threshold.

[0076] In some embodiments, the UE may report concurrent / mixed UE capabilities for both model-related and non-model-related.

[0077] In some embodiments, the UE may report the number of measurement reports using the model and the number of measurement reports not using the model that the UE may be configured with.

[0078] For example, the UE may report that when the UE is configured with M CSI reports using the model in a bandwidth part / component carrier (Bandwidth Part / ComponentCarrier, BWP / CC), the UE does not expect to be configured with more than N other CSI reports not using the model. Here, the value of N may either not need to be reported or may be reported by a value of zero, both of which indicate that when the CSI report using the model is configured, the UE cannot be configured with other CSI reports not using the model.

[0079] In another example, the UE may report that when the UE is configured with M beam reports using the model in the BWP / CC (bandwidth part / component carrier), the UE does not expect to be configured with more than N other beam reports not using the model. Here, the value of N may not need to be reported or may be reported by a value of zero, both of which indicate that when the beam reports using the model are configured, the UE cannot be configured with other beam reports not using the model.

[0080] In some embodiments, the UE may report: The UE may process measurement reports using the model and measurement reports not using the model simultaneously.

[0081] For example, the UE may simultaneously process M beam reports using the model and N beam reports not using the model in a CC (or across all CCs). The value of N does not need to be reported or may be reported with a value of zero.

[0082] For example, the UE may simultaneously process M beam reports using the model and N CSI reports not using the model in a CC (or across all CCs). The value of N does not need to be reported or may be reported with a value of zero.

[0083] B. Model Monitoring Process

[0084] On the UE side, the performance of the deployed model may not always be good. Due to limited model generalization and dynamic environment, the performance of the model may degrade. Therefore, both the network and the UE should continuously monitor the performance of the model.

[0085] Here, we propose three solutions for model performance monitoring.

[0086] Method 1: The network provides auxiliary channel information to assist model performance monitoring.

[0087] In some embodiments, the UE may report the relationship between the supplementary channel information and the actual channel measurement.

[0088] exist Figure 3 In the example shown, the auxiliary channel information and the actual channel measurement are respectively represented by H assist and H represents the output of the first AI model. And its AI model input is based on the auxiliary channel information. The second AI model output is And its AI model input is based on actual channel measurements.

[0089] One metric used to evaluate model performance is similarity, such as cosine similarity, generalized cosine similarity, or squared generalized cosine similarity. Assume H assist The similarity between S and H is S input ,and and The similarity between them is S output In the example, if S input The value is equal to (or approximately equal to) S output If the value of , the AI ​​model can still work well with some guaranteed performance.

[0090] Another metric used to evaluate model performance is distance, such as Euclidean distance or normalized mean square error. Assume H assist The distance between and H is D input ,and and The distance between them is D output In the example, if D input The value is equal to (or approximately equal to) D output If the value of , the AI ​​model can still work well with some guaranteed performance.

[0091] In some embodiments, the UE needs to report a performance indication regarding the results of performance monitoring to the network.

[0092] In some embodiments, the performance indication includes an indicator in the measurement report. For example, values ​​0 and 1 indicate that the AI ​​model is invalid and valid, respectively. More specifically, a value of 0 means that S output / S input or D output / D input Below the threshold.

[0093] In some embodiments, the performance indication includes the similarity (S input and S output ) or distance (D input and D output ). In some embodiments, the performance indicator includes S output / S input or D output / D input The value of .

[0094] In some embodiments, the network may provide a set of supplementary channel information. In this case, each supplementary channel information may be associated with a performance indicator.

[0095] Method 2: Dedicated channel measurement opportunities are used to monitor model performance.

[0096] In some embodiments, the network may indicate at least one dedicated channel measurement opportunity for monitoring model performance.

[0097] In one example of using an AI model for channel prediction, the AI ​​model input is based on actual channel measurements at some previous occasions (e.g., observation occasions). The AI ​​model output is a predicted channel measurement value that can be used for some future occasions (e.g., prediction occasions). Therefore, the UE does not have to obtain the actual channel measurement at the prediction occasion. However, if the occasion is indicated as a dedicated channel measurement occasion (e.g., monitoring occasion) for monitoring model performance, such as Figure 4 As shown, the UE may need to obtain both the actual channel measurement and the predicted channel measurement. The UE / network may then compare the actual channel measurement with the predicted channel measurement to check whether the channel prediction is accurate enough.

[0098] In some embodiments, the UE must report both actual channel measurements and predicted channel measurements for monitoring occasions in the measurement report.

[0099] In one example, an AI model is used for beam prediction.

[0100] In some embodiments, the actual channel measurement includes RSRPs of a plurality of reference signals, and the predicted channel measurement includes RSRPs of the plurality of reference signals.

[0101] In some embodiments, the network may indicate the number of RSRPs included in the actual channel measurement and the predicted channel measurement.

[0102] In some embodiments, the number of RSRPs included in the actual channel measurement is the same as the number of RSRPs included in the predicted channel measurement.

[0103] In some embodiments, each RSRP is associated with a reference signal index.

[0104] In another example, an AI model may be employed to perform CSI prediction.

[0105] In some embodiments, the actual channel measurement comprises a conventional codebook based precoding matrix indication (PMI).

[0106] Method 3: Dedicated module of AI module for model performance monitoring

[0107] In some embodiments, a portion of the AI ​​model output may be used for model performance monitoring.

[0108] In some embodiments, the portion of the AI ​​model output includes at least an indicator indicating whether the AI ​​model is valid. The indicator may be reported in a measurement report.

[0109] In some embodiments, the relationship between a portion of the AI ​​model output and the AI ​​model input can be used for model performance monitoring.

[0110] In such Figure 5 In the example shown, the AI ​​model input is the channel information H, and the AI ​​model output of the first submodule is the compressed channel information And the AI ​​model input of the second submodule is the reconstructed channel information

[0111] In some embodiments, the measurement report may include an indication of the relationship.

[0112] In some embodiments, the relationship may be a similarity / distance between the channel information and the reconstructed channel information.

[0113] In some embodiments, the measurement report may include both an indication of the relationship and the compressed channel information.

[0114] In some embodiments, the UE obtains N measurement results based on the received reference signal. The AI ​​model input may include M measurement results out of the N measurement results, where M is less than N. In this scenario, the remaining NM measurement results may be used for model performance monitoring.

[0115] In some embodiments, the N measurement results are N RSRP values, each of which corresponds to a beam between the network and the UE. The AI ​​model input includes only M RSRP values ​​out of the N RSRP values, where M < N. Therefore, the remaining NM RSRP values ​​can be used for model performance monitoring. The remaining NM RSRP values ​​correspond to a set of beams between the network and the UE. In addition, a portion of the AI ​​model output includes predicted NM RSRP values ​​corresponding to a set of beams between the network and the UE.

[0116] In some embodiments, the UE may report a validity indicator based on the remaining NM RSRP values ​​and the predicted NM RSRP values.

[0117] In some embodiments, the effectiveness indicator may be a similarity between the remaining NM RSRP values ​​and the predicted NM RSRP values.

[0118] In some embodiments, the validity indicator may be a difference (or distance) between the remaining NM RSRP values ​​and the predicted NM RSRP values.

[0119] In some embodiments, the UE may report both the remaining NM RSRP values ​​and the predicted NM RSRP values ​​in the measurement report.

[0120] In some embodiments, one of the predicted NM RSRP values ​​is reported relative to one of the remaining NM RSRP values ​​corresponding to the same beam between the network and the UE.

[0121] In some embodiments, the N measurement results are N channel measurement results corresponding to an N-port reference signal. The AI ​​model input includes only M (M < N) channel measurement results of the N channel measurement results. Therefore, the remaining NM channel measurement results can be used for model performance monitoring. The remaining NM channel measurement results correspond to a set of ports of the N-port reference signal. In addition, a portion of the AI ​​model output includes predicted NM channel measurement results corresponding to a set of ports of the N-port reference signal.

[0122] In some embodiments, the UE may report a validity indicator based on the remaining NM channel measurement results and the predicted NM channel measurement results.

[0123] In some embodiments, the validity indicator may be a similarity between the remaining NM channel measurements and the predicted NM channel measurements.

[0124] In some embodiments, the validity indicator may be the difference (or distance) between the remaining NM channel measurements and the predicted NM channel measurements.

[0125] Figure 6 An exemplary block diagram of a hardware platform 600 is shown, which may be part of a network device (e.g., a base station) or a communication device (e.g., a user equipment (UE)). The hardware platform 600 includes at least one processor 610 and a memory 605 having instructions stored thereon. The instructions, when executed by the processor 610, configure the hardware platform 600 to execute the Figures 1 to 5 and Figure 7 As well as the operations described in the various embodiments described in this patent document. Transmitter 615 transmits or sends information or data to another device. For example, a network device transmitter can send a message to a user device. Receiver 620 receives information or data transmitted or sent by another device. For example, a user device can receive a message from a network device.

[0126] The embodiments discussed above are applicable to wireless communications. Figure 7An example of a wireless communication system (e.g., a 5G or NR cellular network) including a base station 720 and one or more user equipments (UEs) 711, 712, and 713 is shown. In some embodiments, the UE accesses a BS (e.g., a network) using a communication link to the network (sometimes referred to as an uplink direction, as depicted by dashed arrows 731, 732, 733), which then enables subsequent communication from the BS to the UE (e.g., shown in the direction from the network to the UE, sometimes referred to as a downlink direction, as shown by arrows 741, 742, 743). In some embodiments, the BS sends information to the UE (sometimes referred to as a downlink direction, as shown by arrows 741, 742, 743), which then enables subsequent communication from the UE to the BS (e.g., shown in the direction from the UE to the BS, sometimes referred to as an uplink direction, as shown by dashed arrows 731, 732, 733). The UE may be, for example, a smartphone, a tablet, a mobile computer, a machine to machine (M2M) device, an Internet of Things (IoT) device, and the like.

[0127] A wireless communication method includes transmitting, by a communication device, a capability message indicating that the communication device is capable of using a model used by the communication device to perform one or more wireless communication operations.

[0128] In some embodiments, the capability message includes a time requirement of the model. In some embodiments, the time requirement is associated with a model identifier.

[0129] In some embodiments, the model may be received by the communication device from a network.

[0130] In some embodiments, the capability message includes one of: a calculation time, a calculation time offset, or a model activation time. In some embodiments, the calculation time is the time required for the communication device to operate, wherein the calculation time offset is the additional time required for the communication device to operate, and wherein the model activation time is the time required for the communication device to activate the operation.

[0131] In some embodiments, the calculation time may be a constant that is independent of the selected model. In some embodiments, the value of the calculation time offset depends on the model. In some embodiments, the capability message indicates that the measurement report is created after a time interval T based on reception of a reference signal by the communication device, where T depends on the calculation time offset and the calculation time.

[0132] In some embodiments, the method further comprises transmitting, by the communication device, an identifier message indicating that the model is activated.

[0133] In some embodiments, the identifier message may be sent via the PUCCH. In some embodiments.

[0134] In some embodiments, the method further comprises receiving, by the communications device, a response indicating that the identifier message was successfully received.

[0135] In some embodiments, the method further comprises deactivating, by the communication device, the model when the response is not received by the communication device.

[0136] In some embodiments, the capability message includes capabilities of the communication device, wherein the capabilities include capabilities for model-dependent and non-model-dependent concurrency.

[0137] Another method of wireless communication includes reporting, by a communication device, a performance indication of a model.

[0138] In some embodiments, the performance indication comprises a relationship between the assistance information and actual measurement information made by the communications device.

[0139] In some embodiments, the relationship may be a similarity or a distance, wherein the similarity comprises cosine similarity, generalized cosine similarity or squared generalized cosine similarity, and wherein the distance comprises Euclidean distance or normalized mean square error.

[0140] In some embodiments, the communications device is configured with at least one dedicated occasion for performance indication of the model.In some embodiments, the performance indication comprises predicted measurements and actual measurements of the dedicated occasion.

[0141] In some embodiments, the actual measurements include RSRPs of a plurality of reference signals, wherein the predicted measurements include predicted RSRPs of the reference signals based on the model.

[0142] In some embodiments, the actual measurement comprises a codebook based precoding matrix indication (PMI).

[0143] In some embodiments, the performance indication also includes a relationship between an output of the model and an input of the model.

[0144] In some embodiments, the performance indication comprises a relationship between MN remaining measurements of N measurements and a portion of the model output, wherein the N measurements are based on a reference signal, wherein the M measurements are based on the N measurements and are used as input to the model, wherein M<N.

[0145] In some embodiments, the relationship may be a similarity or a distance, wherein the similarity comprises cosine similarity, generalized cosine similarity or squared generalized cosine similarity, and wherein the distance comprises Euclidean distance or normalized mean square error.

[0146] In another exemplary aspect, the above method is embodied in the form of processor executable code and stored in a non-transitory computer readable storage medium. The code included in the computer readable storage medium causes the processor to implement the method described in this patent document when executed by the processor.

[0147] In yet another exemplary embodiment, a device configured or operable to perform the above method is disclosed.

[0148] The above-described aspects and other aspects and embodiments thereof are described in more detail in the drawings, the description and the claims.

[0149] The term "exemplary" is used in this document to mean "an example of," and does not imply an ideal or preferred embodiment unless otherwise stated.

[0150] Some of the embodiments described herein are described in the general context of methods or processes, which, in one embodiment, may be implemented by a computer program product, embodied in a computer-readable medium, including computer-executable instructions, such as program codes, executed by a computer in a networked environment. Computer-readable media may include removable and non-removable storage devices, including but not limited to read-only memory (ROM), random access memory (RAM), compact discs (CD), digital versatile discs (DVD), etc. Therefore, computer-readable media may include non-transitory storage media. In general, program modules may include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. Computer-executable instructions or processor-executable instructions, associated data structures, and program modules represent examples of program codes for executing the steps of the methods disclosed herein. A specific sequence of such executable instructions or associated data structures represents an example of corresponding behavior for implementing the functions described in such steps or processes.

[0151] Some of the disclosed embodiments may be implemented as devices or modules using hardware circuits, software or a combination thereof. For example, a hardware circuit implementation may include discrete analog and / or digital components, which are, for example, integrated as a part of a printed circuit board. Alternatively or additionally, the disclosed components or modules may be implemented as a dedicated integrated circuit (Application Specific Integrated Circuit, ASIC) and / or implemented as a field programmable gate array (Field Programmable Gate Array, FPGA) device. Some embodiments may additionally or alternatively include a digital signal processor (digital signal processor, DSP), which is a specialized microprocessor with an architecture optimized for the operational requirements of digital signal processing associated with the functions disclosed in the present application. Similarly, various components or subcomponents within each module may be implemented with software, hardware or firmware. The connection between the modules and / or components within the modules may be provided using any of the connection methods and media known in the art, including but not limited to communications on the Internet, wired or wireless networks using appropriate protocols.

[0152] Although this document contains many details, these should not be interpreted as limitations on the scope of the claimed invention or the invention that may be claimed, but should be interpreted as descriptions of features that are specific to a particular embodiment. Certain features described in this document in the context of a separate embodiment may also be implemented in combination in a single embodiment. Conversely, the various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any suitable sub-combination. In addition, although the above-mentioned features may be described as working in certain combinations and even initially claimed as such, in some cases, one or more features from the claimed combination may be deleted from the combination, and the claimed combination may be directed to a sub-combination or a variant of a sub-combination. Similarly, although operations are depicted in a particular order in the accompanying drawings, this should not be understood as requiring such operations to be performed in the particular order or sequential order shown, or requiring all illustrated operations to be performed to achieve the desired result.

[0153] Only a few implementations and examples are described and other implementations, enhancements and variations can be made based on what is described and illustrated in this disclosure.

Claims

1. A wireless communication method, comprising: A capabilities message is transmitted by a communication device, the capabilities message indicating that the communication device is capable of using a model used by the communication device to perform one or more wireless communication operations.

2. The method according to claim 1, wherein: The capability message includes the timing requirements of the model.

3. The method according to claim 2, wherein: The time requirement is associated with a model identifier.

4. The method according to claim 3, wherein: The model may be received by the communication device from a network.

5. The method according to claim 2, wherein: The capability message includes one of the following: calculation time, calculation time offset, or model activation time, The calculation time is the time required for the communication device to operate, the calculation time offset is the additional time required for the communication device to operate, and the model activation time is the time required for the communication device to activate the operation.

6. The method according to claim 5, wherein: The calculation time can be constant independent of the selected model.

7. The method according to claim 5, wherein: The value of the calculated time offset depends on the model.

8. The method according to claim 5, wherein: The capability message indicates that a measurement report is created after a time interval T based on reception of a reference signal by the communication device, wherein T depends on the calculation time offset and the calculation time.

9. The method of claim 1, further comprising transmitting, by the communication device, an identifier message indicating that the model is activated.

10. The method according to claim 9, wherein: The identifier message can be sent through the PUCCH.

11. The method of claim 9, further comprising receiving, by the communication device, a response indicating that the identifier message was successfully received.

12. The method of claim 11, further comprising deactivating, by the communication device, the model when the response is not received by the communication device.

13. The method according to claim 1, wherein: The capability message includes the capabilities of the communication device, wherein the capabilities include concurrent capabilities for model-related and non-model-related.

14. A wireless communication method, comprising: An indication of the performance of the model is reported by the communication device.

15. The method according to claim 14, wherein: The performance indication comprises a relationship between assistance information and actual measurement information performed by the communication device.

16. The method according to claim 15, wherein: The relationship can be a similarity or a distance, wherein the similarity comprises cosine similarity, generalized cosine similarity or squared generalized cosine similarity, wherein the distance comprises Euclidean distance or normalized mean square error.

17. The method according to claim 14, wherein: The communication device is configured with at least dedicated opportunities for performance indication of the model.

18. The method according to claim 14, wherein: The performance indication includes predicted measurements and actual measurements of the dedicated occasions.

19. The method according to claim 18, wherein: The actual measurements include RSRPs of a plurality of reference signals, wherein the predicted measurements include predicted RSRPs of the reference signals based on the model.

20. The method according to claim 18, wherein: The actual measurement includes a codebook-based Precoding Matrix Indicator (PMI).

21. The method according to claim 14, wherein: The performance indication also includes a relationship between an output of the model and an input of the model.

22. The method according to 14, wherein: The performance indication comprises a relationship between MN remaining measurement results out of N measurement results and a portion of the model output, wherein the N measurement results are based on the reference signal, wherein the M measurement results are based on the N measurement results and are used as input to the model, wherein M<N.

23. The method according to claim 21 or 22, wherein: The relationship can be a similarity or a distance, wherein the similarity can be at least one of the following: cosine similarity, generalized cosine similarity or squared generalized cosine similarity, wherein the distance can be at least one of the following: Euclidean distance or normalized mean square error.

24. An apparatus for wireless communication, comprising a processor configured to implement the method according to one or more of claims 1 to 24.

25. A non-transitory computer-readable program storage medium having stored thereon codes which, when executed by a processor, cause the processor to implement the method according to one or more of claims 1 to 24.

Images