Method for reporting performance related information associated to a computational model for predicting channel state information

Abstract

A communication device (730) and method therein for handling communication in a wireless communications network (700). The communication device (730) receives a first signaling message from a network node (711) for reporting a performance related information associated to a computational model for predicting a Channel State Information (CSI) in one or multiple future time instances. The communication device (730) sends a report on the performance related information according to the first signaling message.

Description

[0001] METHOD FOR REPORTING PERFORMANCE RELATED INFORMATION ASSOCIATED TO A COMPUTATIONAL MODEL FOR PREDICTING CHANNEL STATE INFORMATION

[0002] TECHNICAL FIELD

[0003] Embodiments herein relate to a network node, a communication device and methods performed therein for handling transmitting performance monitoring reports in a wireless communication system.

[0004] BACKGROUND

[0005] In a typical wireless communication network or system, wireless devices, also known as wireless communication devices, mobile stations, and / or user equipment (UE), communicate via a Radio Access Network (RAN) to one or more core networks (CN). The RAN covers a geographical area which is divided into service areas or cell areas, which may also be referred to as a beam or a beam group, with each service area or cell area being served by a radio network node such as a radio access node e.g., a Wi-Fi access point or a radio base station (RBS), which in some networks may also be denoted, for example, a “NodeB” or “eNodeB” or “eNB” or “gNB”. A service area or cell area is a geographical area where radio coverage is provided by the radio network node. A radio network node communicates over an air interface operating on radio frequencies with one or more wireless communication device within a range of the radio network node.

[0006] A Universal Mobile Telecommunications System (UMTS) is a third generation (3G) telecommunication network, which evolved from the second generation (2G) Global System for Mobile Communications (GSM). Specifications for the Evolved Packet System (EPS), also called a Fourth Generation (4G) network or Long Term Evolution (LTE) have been completed within the 3rd Generation Partnership Project (3GPP) and this work continues in the coming 3GPP releases, for example to specify a Fifth Generation (5G) New Radio (NR) network, Next Generation (NG) and upcoming releases.

[0007] With the 5G technologies such as NR, focus is on a set of features such as the use of very many transmit- and receive-antenna elements that makes it possible to utilize beamforming, such as transmit-side and receive-side beamforming. Transmit-side beamforming means that the transmitter can amplify the transmitted signals in a selected direction or directions, while suppressing the transmitted signals in other directions. Similarly, on the receive-side, a receiver can amplify signals from a selected direction or directions, while suppressing unwanted signals from other directions.

[0008] Artificial Intelligence (Al) and Machine Learning (ML) have been investigated, both in academia and industry, as promising tools to optimize the design of the air-interface in wireless communication networks. Example use cases include using autoencoders for Channel State Information (CSI) compression to reduce the feedback overhead and improve channel prediction accuracy; using deep neural networks for classifying Line-of-Sight (LOS) and Non-LOS (NLOS) conditions to enhance the positioning accuracy; and using reinforcement learning for beam selection at the network node side and / or the UE side to reduce the signaling overhead and beam alignment latency; using deep reinforcement learning to learn an optimal precoding policy for complex Multiple Input Multiple Output (MIMO) precoding problems.

[0009] In 3GPP NR standardization work, a study item on AI / ML for the NR air interface has been started. This study item explored the benefits of augmenting the air-interface with features enabling improved support of AI / ML based algorithms for enhanced performance and / or reduced complexity / overhead. Through studying and specifying a few selected use cases, e.g. CSI feedback, beam management, and positioning etc., this study item aims at laying the foundation for future air-interface use cases leveraging AI / ML techniques and aims to design the mechanisms to accommodate AI / ML into the 3GPP standard.

[0010] General aspects for NR Rel-18 AI / ML for NR air interface

[0011] Lifecycle Management (LCM) operations for AI / ML for NR air interface:

[0012] An important part in Al development and operation is the lifecycle management (LCM) of the AI / ML model e.g., model training, model deployment, model inference, model monitoring, model updating, and AI / ML functionality.

[0013] In NR Rel-18 AI / ML for NR air interface study item, the LCM procedure is studied for the case that an AI / ML model has a model ID with associated information and / or for the case that a given functionality is provided by some AI / ML operations.

[0014] Two types of LCM operations were studied in NR Rel-18, functionality -based LCM and model-ID based LCM:

[0015] Functionality refers to an AI / ML-enabled Feature / feature group (FG) enabled by configuration(s), where configuration(s) is(are) supported based on conditions indicated by UE capability. Correspondingly, functionality-based LCM operates based on, at least, one configuration of an AI / ML-enabled Feature / FG or specific configurations of an AI / ML-enabled Feature / FG. In functionality-based LCM, the network node indicates activation / deactivation / fallback / switching of AI / ML functionality via 3 GPP signaling, e.g., radio resource control (RRC), medium access control (MAC)-control element (CE), DCI. Models may not be identified at the Network, and UE may perform model-level LCM. Whether and how much awareness / interaction network (NW) node should have about modellevel LCM requires further study. For functionality identification, there may be either one or more than one functionalities defined within an AI / ML-enabled feature, whereby AI / ML-enabled Feature refers to a Feature where AI / ML may be used.

[0016] In model-ID-based LCM, models are identified at the Network, and Network / UE may activate / deactivate / select / switch individual AI / ML models via model ID. A model may be associated with specific configurations / conditions associated with UE capability of an AI / ML-enabled Feature / FG and additional conditions, e.g., scenarios, sites, and datasets, as determined / identified between the UE-side and NW-side. An AI / ML model identified by a model ID may be logical, and how it maps to physical AI / ML model(s) may be up to implementation.

[0017] Functional framework for AI / ML for NR air interface:

[0018] Figure 1 shows a functional framework that can be used for studying model LCM aspects for different Al for physical layer (PHY) use cases. The general framework consists of the following:

[0019] - Data Collection is a function that provides input data to the Model Training, Management, and Inference functions. o Training Data: Data needed as input for the AI / ML Model Training function. o Monitoring Data: Data needed as input for the Management of AI / ML models or AI / ML functionalities. o Inference Data: Data needed as input for the AI / ML Inference function.

[0020] - Model Training is a function that performs AI / ML model training, validation, and testing which may generate model performance metrics which can be used as part of the model testing procedure. The Model Training function is also responsible for data preparation e.g., data pre-processing and cleaning, formatting, and transformation, based on Training Data delivered by a Data Collection function, if required. o Trained / Updated Model: In case of having a Model Storage function, this is used to deliver trained, validated, and tested AI / ML models to the Model Storage function, or to deliver an updated version of a model to the Model Storage function.

[0021] - Management is a function that oversees the operation e.g., selection / (de)activation / switching / fallback, and monitoring e.g., performance, of AI / ML models or AI / ML functionalities. This function is also responsible for making decisions to ensure the proper inference operation based on data received from the Data Collection function and the Inference function. o Management Instruction: Information needed as input to manage the Inference function. Concerning information may include selection / (de)activation / switching of AI / ML models or AI / ML-based functionalities, fallback to non- AI / ML operation i.e., not relying on inference process, etc. o Model Transfer / Delivery Request: Used to request model(s) to the Model Storage function. o Performance Feedback / Retraining Request: Information needed as input for the Model Training function, e.g., for model (re)training or updating purposes.

[0022] Inference is a function that provides outputs from the process of applying AI / ML models or AI / ML functionalities, using the data that is provided by the Data Collection function i.e., Inference Data, as an input. The Inference function is also responsible for data preparation e.g., data pre-processing and cleaning, formatting, and transformation, based on Inference Data delivered by a Data Collection function, if required. o Inference Output: Data used by the Management function to monitor the performance of AI / ML models or AI / ML functionalities.

[0023] - Model Storage is a function responsible for storing trained / updated models that can be used to perform the Inference function. o Note: The Model Storage function in Figure 1 is only intended as a reference point (if any) when applicable for protocol terminations, model transfer / delivery, and related processes. It should be stressed that its purpose does not encompass restricting the actual storage locations of models. Therefore, the specification impact of all data / information / instruction flows i.e., the arrows, in Figure 1 to / from this function should be studied case by case. o Model Transfer / Delivery: Used to deliver an AI / ML model to the Inference function.

[0024] Signaling procedures for model and functionality life cycle management

[0025] The signaling procedures for different scenarios for model-ID-based management and / or functionality-based management are exemplified below for UE-side AI / ML models, where a UE-side AI / ML model is defined as an AI / ML model whose inference is at the UE. The procedures include scenarios for which the management decision is taken by the network or by the UE. For network-side decision, this can be either network-initiated, or UE-initiated and requested to the network. While for UE-side decision, this can be either event-triggered as configured by the network and where the UE’s decision is reported to the network, or UE- autonomous, with or without UE’s decision being reported to the network.

[0026] Decision by the network o Network-initiated

[0027] Figure 2 shows a signal flowchart of Network decision, network-initiated AI / ML management.

[0028] The case where the LCM decision is taken and initiated by the network is depicted in Figure 2.

[0029] Note: The Management Instruction may be a result of model / functionality performance monitoring at the network. Note: The Management Instruction may include information about the model or functionality. o UE-initiated and requested to the network.

[0030] The case where the LCM decision is taken by the network but where the request is initiated by the UE is depicted in Figure 3.

[0031] Note: The Management Request may be a result of model / functionality monitoring at the UE.

[0032] Note: In response to the Management Request, the network may send a Management

[0033] Instruction to the UE.

[0034] Note: The Management Request may include information about the model or functionality.

[0035] Note: The network may accept or reject the Management Request from the UE.

[0036] Note: The Management Request may include information related to model / functionality performance metrics.

[0037] Note: The Management Instruction may include information about the model or functionality.

[0038] Decision by the UE o Event-triggered as configured by the network, UE’s decision is reported to the network.

[0039] The case where the LCM decision is taken by the UE according to prior network configuration is depicted in Figure 4.

[0040] Note: Use case-specific events / conditions may be configured by the network for event- triggered AI / ML management at the UE.

[0041] Note: UE may send a Management Decision Report to the network following event- triggered AI / ML management at the UE.

[0042] Note: The Management Decision Report may include information about the model or functionality. o UE-autonomous, UE’s decision is reported to the network

[0043] The case where the LCM decision can autonomously be taken by the UE is depicted in

[0044] Figure 5.

[0045] Note: The UE may be configured to send a Management Decision Report to the network upon performing a model / functionality Management Decision. o UE-autonomous, UE’s decision is not reported to the network For the case where the LCM decision can autonomously be taken by the UE and where the decision is not reported to the network, the AI / ML management is transparent from a network perspective.

[0046] Time domain CSI prediction at UE

[0047] 3GPP NR Rel-18 time domain Type II CSI prediction at UE.

[0048] In 3 GPP NR Rel-18, channel measurement resource (CMR) enhancement for Type II CSI prediction at UE, a.k.a. Enhanced Type II predicted Precoding Matrix Indicator (PMI), has been introduced, see the measurement part in Figure 6, where a burst of K G {4, 8, 12} same CSI-RS resources are configured to the UE in a single CSI-RS resource set. The burst of CSI-RS resources is aperiodically (AP) triggered using a single downlink control information (DCI). The K CSI-RS resources are used for the UE to extract time domain channel properties of the channel, based on which a future CSI can be predicted. Alternatively, NW may also configure a legacy periodic (P) or semi-persistent (SP) CSI-RS resource. The CSI-RS resources are uniformly spaced in time, separated by m G {1, 2} slots, within the resource set.

[0049] For the Rel-18 Type II predicted PMI enhancement, a UE can be configured by gNB to report predicted PMIs for N4G {1, 2, 4, 8} time slots, see the Rel-18 Type II PMI part in Figure 6. Note that the prediction herein is relative to the CSI-RS reference resource. The predicted N4PMIs are supposed to reflect the channels with d G {l, m} slots separation, starting from 8 G {— nCSIre j , 0,1,2} slots into the future relative to the CSI-RS reference resource. For AP CSI-RS burst, m G {1,2}, while for P / SP CSI-RS, m is the CSI-RS periodicity. The spacing d between the N4PMIs and offset 6 relative to the CSI-RS reference resource can be configured by the gNB via RRC signaling. The N4PMIs are compressed in a beam-frequency-Doppler domain, and the compressed PMI is reported to the gNB in a single CSI report.

[0050] 3GPP NR Rel-18 time-domain CSI prediction using UE-side AI / ML model

[0051] In Rel-18, Al-based UE-side CSI prediction was introduced as an Al for PHY use case, and being studied in the study item on AI / ML for the NR air interface. One or more AI / ML models can be trained and deployed at a UE for the Al-based CSI-prediction feature. During model inference, a UE is configured by the gNB to measure a set of historical CSI-RSs and then report a predicted CSI for one or multiple future time instances using its AI / ML model(s). Model performance monitoring

[0052] There are several methods for model monitoring. Monitoring based on intermediate key performance indicators (KPI), e.g., inference accuracy, requires collecting new ground-truth data similar / identical to the training data, which is very accurate but has a high cost due to the potentially large measurement / reporting overhead. Monitoring based on data distribution of input / output data does not require any additional signaling overhead but is less accurate than monitoring based on inference accuracy since one does not retrieve the ground truth. Similarly, monitoring based on system performance does not require any additional signaling overhead, however, it can be challenging to detect that the root-cause for bad system performance is due to an inaccurate model, and not due to some other malfunctioning procedure or hardware. Monitoring based on data distribution can in contrast identify a potential problem in the model by detecting that the dataset observed during inference is not same as during training. However, it is not-trivial to define conditions and measurable data- distribution based KPIs for sounding a model failure alarm with a good trade-off between model failure detection reliability and accuracy, e.g., low false alarm rate, low missed detection rate and low latency.

[0053] To ensure reliable / accurate model performance monitoring results, sufficient monitoring data samples will be collected and used to derive the performance monitoring results. Examples of model output accuracy based performance monitoring results include intermediate KPI per monitoring data sample, intermediate KPI statistics associated to a monitoring data set, the percentage of monitoring data samples within a monitoring dataset for which the intermediate KPI fulfills a certain condition, a flag indicating whether the model is functioning ok or not. Examples of data drift based performance monitoring results include monitoring data statistics, the difference between the monitoring data statistics and the data statistics obtained in the model training stage, a flag indicating whether a data drift is detected or not.

[0054] When implementing model monitoring, the monitoring method can be selected based on UE service requirements. For example, a UE with mobile broadband (MBB) could start with a low-cost solution (e.g. system performance based), if problems are observed / predicted, then activate an inference accuracy based monitoring method associated with a higher complexity. High-complexity and signaling overhead monitoring may be required for certain UE service requirements, such as for emergency localization use cases or UEs with Ultra Reliable and Low Latency Communications (URLLC) connection. For CSI prediction using UE side AI / ML model use case studied in Rel-18 AI / ML for NR air interface study item, at least the following aspects have been proposed by companies on performance monitoring for functionality-based LCM:

[0055] Type 1 : o UE calculates the performance metric(s) o UE reports performance monitoring output that facilitates functionality fallback decision at the network

[0056] ■ Performance monitoring output details can be further defined.

[0057] ■ NW may configure threshold criterion to facilitate UE side performance monitoring (if needed). o NW makes decision(s) of functionality fallback operation (fallback mechanism to legacy CSI reporting).

[0058] Type 2: o UE reports predicted CSI and / or the corresponding ground-truth o NW calculates the performance metrics. o NW makes decision(s) of functionality fallback operation (fallback mechanism to legacy CSI reporting).

[0059] Type 3: o UE calculates the performance metric(s) o UE reports performance metric(s) to the NW o NW makes decision(s) of functionality fallback operation (fallback mechanism to legacy CSI reporting).

[0060] - Functionality selection / activation / deactivation / switching as defined for other UE side use cases can be reused, if applicable.

[0061] Configuration and procedure for performance monitoring.

[0062] CSI-RS configuration for performance monitoring.

[0063] - Performance metric including at least intermediate KPI, e.g., normalized mean square error (NMSE) or squared generalized cosine similarity (SGCS).

[0064] - UE report, including periodic / semi-persistent / aperiodic reporting, and event driven report.

[0065] - Note: UE may make decisions within the same functionality on model selection, activation, deactivation, switching operation transparent to the NW.

[0066] SUMMARY As part of developing embodiments herein one or more problems were first identified.

[0067] For monitoring the performance of a UE-sided AI / ML model, if the performance monitoring is performed at the UE-side, then, a UE can report the model performance monitoring results to the NW, so that the NW takes the UE reported model performance information into account when making model level or functionality level LCM decisions e.g., fallback to non-AL / ML algorithm, functionality / model switching, etc.

[0068] However, how to configure the specific reporting configuration which will allow the network to configure a UE to report back the outcome of the performance monitoring for the AI / non-AI based CSI prediction model / algorithm, which type of physical resources can be used by the UE for reporting the performance monitoring outcome, how to encode the performance monitoring outcome into a CSI report, what signaling to use etc. have not been discussed before and are open problems.

[0069] Certain aspects of the disclosure and their embodiments may provide solutions to these or other challenges.

[0070] It is an object of embodiments herein to provide a mechanism for the network to configure a communication device to report to the network a performance monitoring report in an efficient manner. For example, a UE to report a performance related information for an AI / non-AI based CSI prediction model / algorithm, which predicts CSI for one or multiple future time instances, including configuration of physical resources for the performance monitoring report and the signaling of the configuration.

[0071] The disclosure proposes, for example, methods for configuring and / or allocating the physical resources for the communication device to transmit the performance monitoring report, according to the time domain behavior of the performance monitoring reports and the report quantities. Also, methods for reporting performance monitoring results in a 2 Part CSI framework are proposed herein.

[0072] According to an aspect of embodiments herein the object is achieved by a method performed by a communication device for handling communication in a wireless communications network. The communication device receives from a network node, a first signaling message, e.g. a configuration message, for reporting a performance related information associated to a computational model for predicting a CSI in one or multiple future time instances. The communication device sends a report on the performance related information according to the first signaling message.

[0073] According to some embodiments, the first signaling message, such as a CSI prediction performance report configuration, configures the UE to report back the performance related information for the computational model such as an AI / non-AI based CSI-prediction model / algorithm along with predicted CSI related parameters, e.g. like Channel Rank Indicator (CRI), Channel Quality Indicator (CQI), PMI, Rank Indicator (RI), Layer Indicator (LI) etc.

[0074] According to some embodiments, based on a configured time-domain behavior of a CSI report carrying the predicted CSI and / or the CSI prediction performance report configuration, the UE may feedback the CSLreport over Physical Uplink Control Channel (PUCCH) and / or Physical Uplink Shared Channel (PUSCH) which includes performance related information for the AI / non-AI based CSI-prediction model / algorithm.

[0075] According to some embodiments, the communication device may only include the performance related information for the AI / non-AI based CSI-prediction model / algorithm in the CSI report when the report is feedback periodically or semi-persistently over PUCCH / PUSCH.

[0076] According to some embodiments, the communication device may include the performance related information for the AI / non-AI based CSI-prediction model / algorithm with or without the predicted CSI related parameters, e.g. like CRI, CQI, PMI, RI, LI etc. in the CSI report, when the report is feedback aperiodically over PUSCH.

[0077] According to some embodiments, a 2-part CSI structure may be used for reporting performance monitoring results.

[0078] According to some embodiments, CSI part 1 may contain information identifying the number of time instances for which a performance monitoring metric(s) is reported.

[0079] According to some embodiments, CSI Part 2 may report the performance monitoring metric(s) for the time instances that are identified by CSI Part 1.

[0080] According to some embodiments, CSI part 1 may contain information identifying the rank or the number of layers, for which a performance monitoring metric(s) is reported.

[0081] According to some embodiments, CSI Part 2 may report the performance monitoring metric(s) for the layers that are identified by CSI Part 1.

[0082] According to some embodiments, CSI part 1 may contain information identifying the type(s) of performance monitoring metric(s) that is (are) reported.

[0083] According to some embodiments, CSI Part 2 may report the performance monitoring metric(s) of the identified types by CSI Part 1.

[0084] According to an aspect of embodiments herein the object is achieved by a method performed by a network node for handling communication in a wireless communications network. The network node sends a first signaling message to a communication device, e.g. a UE, for configuring the communication device to send a report on a performance related information associated to a computational model for predicting CSI in one or multiple future time instances; and receives a report sent from the communication device based on the first signaling message.

[0085] For example, the network node configures a communication device to report a performance monitoring result(s) associated to a UE-side AUnon-AI based CSI prediction model / algorithm that predicts CSI for one or multiple future time instances, which includes the one or more of the following:

[0086] The network node signals a communication device a CSI prediction performance monitoring report configuration, which configures the communication device to report the performance monitoring related information for an AI / non-AI based CSI-prediction model / algorithm used for generating predicted CSI.

[0087] Examples of performance monitoring related information include performance monitoring output(s) / result(s) and / or performance metric(s) associated to the AI / non-AI based CSI prediction model / algorithm.

[0088] The CSI prediction performance report configuration can be signaled via

[0089] ■ a RRC message, and / or

[0090] ■ a MAC CE, and / or

[0091] ■ a DCI.

[0092] The AI / non-AI based CSI prediction model / algorithm is associated to a feature identifier, a feature group, an AI / ML model identifier, and / or a functionality identifier.

[0093] Examples of predicted CSI include predicted PMI(s), predicted raw channel(s), predicted RI(s), predicted Reference Signal Received Power(s) (RSRP), predicted CQI(s), predicted CRI(s), predicted top-K strongest beams, predicted top-K cells.

[0094] The RRC message may be the same RRC message used for configuring the communication device to report predicted CSI, or a separate RRC message that is different from the one used for configuring UE to report predicted CSI, or the same RRC message used for configuring performance monitoring data collection at the communication device.

[0095] The RRC message indicates the format and / or content of the performance related information for the AI / non-AI based CSI-prediction model algorithm.

[0096] The RRC message configures a time window during which the performance related information shall be reported.

[0097] The MAC CE can be the same MAC CE used for triggering, requesting and / or activating the UE to report the predicted CSI, or a separate MAC CE defined for triggering and / or requesting the UE to report the performance related information.

[0098] The DCI signaling can be the same DCI used for triggering and / or requesting the communication device to report the predicted CSI, or a separate DCI signaling defined for triggering and / or requesting the communication device to report the performance related information.

[0099] Where upon decoding the received report carrying the performance related information, the network decides, either a. configure the subsequent CSI reports with the existing AI / non-AI model and / or algorithm for CSI prediction, if the performance of AI / non-AI based prediction is good enough, or b. configure the subsequent CSI reports with either a different AI / non-AI model / algorithm for CSI prediction, if the performance of AUnon-AI based prediction is not good enough, or c. configure the subsequent CSI reports with fall back to non-AI based CSI prediction, if the Al-based CSI prediction is currently configured and the performance of Al based prediction is not good enough, or d. configure the subsequent CSI reports with fall back to CSI measurement reporting without prediction.

[0100] The network may signal the communication device to stop reporting performance related information for the AI / non-AI based CSI-prediction model / algorithm. The signaling to stop the communication device reporting performance related information can be done via, e.g., a) a time window configured in the CSI prediction performance report configuration. b) using a different RRC message defined for stopping the communication device reporting the performance related information. c) using a MAC CE (the MAC CE can be the same MAC CE used for deactivating the predicted CSI report, or a separate MAC CE defined for stopping the communication device reporting the performance related information.) d) using a DCI (the DCI can be the same DCI used for triggering / requesting the communication device to report the predicted CSI, e.g., 1 bit in DCI to indicate whether to include performance related information in the CSI report or not; or a separate DCI defined for stopping the communication device reporting the performance related information.)

[0101] The network may signal the communication device to start or stop reporting based on a channel condition. Examples of such conditions may be communication device speed being above a certain threshold, or delay spread or doppler spread estimates being outside certain regions.

[0102] It is furthermore provided herein a computer program product comprising instructions, which, when executed on at least one processor, cause the at least one processor to carry out the methods herein, as performed by the communication device and the network node, respectively. It is additionally provided herein a computer-readable storage medium, having stored thereon a computer program product comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the methods herein, as performed by the communication device and the network node, respectively.

[0103] The proposed solution enables the network to configure a UE report to feedback the performance monitoring result for a computational model such as an AI / non-AI based CSI prediction model / algorithm that performs CSI prediction over one or multiple future time instances. The network can use the performance monitoring result to further take decision on activation, deactivation, fallback and / or switching of the active AI / non-AI based CSI prediction model / algorithm via 3GPP signaling e.g., RRC, MAC-CE, DCI.

[0104] The proposed solution can flexibly allocate different physical resources for performance monitoring report for CSI prediction according to the time domain behavior of the performance monitoring report. In addition, the proposed solution offers efficient encoding of CSI report into two-part CSI.

[0105] BRIEF DESCRIPTION OF THE DRAWINGS

[0106] Examples of embodiments herein are described in more detail with reference to attached drawings in which:

[0107] Figure 1 is a schematic block diagram depicting a functional framework for AI / ML for NR Air Interface;

[0108] Figure 2 is a signal flow diagram depicting network decision, network-initiated AI / ML management;

[0109] Figure 3 is a signal flow diagram depicting network decision, UE-initiated AI / ML management;

[0110] Figure 4 is a signal flow diagram depicting UE decision, event-triggered as configured by the network; Figure 5 is a signal flow diagram depicting UE autonomous, decision reported to the network;

[0111] Figure 6 is a schematic block diagram depicting CMR enhancement for Rel-18 Type II CSI prediction at UE;

[0112] Figure 7 is a schematic block diagram depicting a wireless communication system;

[0113] Figure 8 is a flowchart illustrating a method performed in a radio node according to an embodiment herein;

[0114] Figure 9 is a flowchart illustrating a method performed in a UE according to an embodiment herein;

[0115] Figure 10 is a schematic block diagram illustrating an example embodiment of a network node; and

[0116] Figure 11 is a schematic block diagram illustrating an example embodiment of a communication device.

[0117] DETAILED DESCRIPTION

[0118] It should be understood by the skilled in the art that “communication device” is a nonlimiting term which means any terminal, wireless communication terminal, user equipment, Machine Type Communication (MTC) device, Device to Device (D2D) terminal, or node e.g. smart phone, laptop, mobile phone, sensor, relay, mobile tablets or even a small base station communicating within a cell.

[0119] The terms “communication device”, “wireless device”, “UE”, “user equipment”, “terminal equipment”, “wireless terminal” and “terminal” may be used interchangeably herein.

[0120] A network node may be a RAN node, a gNB, an eNB, an en-gNB, a ng-eNB, a gNB- CU, a gNB-CU-CP, a gNB-CU-UP, an eNB-CU, an eNB-CU-CP, an eNB-CU-UP, an IAB- node, an lAB-donor DU, an lAB-donor-CU, an IAB-DU, an IAB-MT, an O-CU, an O-CU- CP, an O-CU-UP, an O-DU, an O-RU, an O-eNB, a Non-Real Time RAN Intelligent Controller (Non-RT RIC), a Real-Time RAN Intelligent Controller (RT-RIC), an 0AM node, a Core Network node / function, a Cloud-based network function, a Cloud-based centralized training node, a node hosting NR PDCP etc.

[0121] Embodiments herein relate to communication networks in general. Figure 7 is a schematic overview depicting a communication network or system 700. The communication network 700 may be a wireless communication network comprising one or more RANs, and one or more CNs. The communication network 700 may use a number of different RATs, such as Wi-Fi, Long Term Evolution (LTE), LTE-Advanced, NR, Wideband Code Division Multiple Access (WCDMA), Global System for Mobile communications / enhanced Data rate for GSM Evolution (GSMZEDGE), Worldwide Interoperability for Microwave Access (WiMax), or Ultra Mobile Broadband (UMB), NR etc. just to mention a few possible implementations.

[0122] In the wireless communication network 700, one or more wireless communication devices 730, 731 such as a UE, a mobile station or a wireless terminal communicates via one or more Radio Access Networks (RAN) to one or more core networks (CN).

[0123] Network nodes operate in the wireless communication network 700 such as a first network node 711, a second network node 712. The first and second network nodes 711, 712 may be any of RAN node, such as gNB, eNB, en-gNB, ng-eNB, gNB etc. The first network node 711 provides radio coverage over a geographical area, a service area 11, which may also be referred to as a beam or a beam group where the group of beams is covering the service area of a first radio access technology (RAT), such as 5G, LTE, Wi-Fi or similar. The second network node 712 provides radio coverage over a geographical area, a service area 12, which may also be referred to as a beam or a beam group where the group of beams is covering the service area of a second radio access technology (RAT), such as 5G, LTE, Wi-Fi or similar. It should be noted that a network node may be a RAN node, a CN node or an 0AM node.

[0124] The first / second network nodes 711 / 712 may be a transmission and reception point e.g. a radio access network node such as a Wireless Local Area Network (WLAN) access point or an Access Point Station (AP STA), an access controller, a base station, e.g. a radio base station such as a NodeB, a gNB, an evolved Node B (eNB, eNode B), a base transceiver station, a radio remote unit, an Access Point Base Station, a base station router, a transmission arrangement of a radio base station, a stand-alone access point or any other network unit capable of communicating with a wireless communication device within the service area served by the respective first / second network nodes 711 / 712 depending e.g. on the radio access technology and terminology used. The first and the second network nodes 711 / 712 may be referred to as a source and a target network node, respectively, and may communicate with the wireless communication device 730, 731 with Downlink (DL) transmissions to the wireless communication device 730, 731 and Uplink (UL) transmissions from the wireless communication device 730, 731.

[0125] A network node (NW) such as the network node 711 or a core network node, or a computational model node, transmits to a UE, e.g. the communication device 730, a first signaling message for reporting a performance monitoring result associated to a computational model for predicting a CSI in one or multiple future time instances.

[0126] • The concept “network” may refer to one of a generic network node, such as the radio network node 12, e.g., a gNB (or the corresponding node in a 6G network), a base station, transmission and reception point (TRP), a unit within the base station to handle at least some ML operation, a relay node, a core network node, a core network node that handle at least some ML operations, or a device supporting device to device (D2D) communication.

[0127] • A computational model may refer to an ML-based model, a configuration of an ML-based model, a non-ML-based functionality, or a configuration of a non-ML- based functionality.

[0128] • The terms “ML-model” and “Al-model” are interchangeable. An AI / ML model can be defined as a functionality or be part of a functionality that is deployed / implemented in a first node. This first node can receive a message from a second node indicating that the functionality is not performing correctly. Further, an AI / ML model can be defined as a feature or part of a feature that is implemented / supported in a first node. This first node can indicate the feature version to a second node. If the ML-model is updated, the feature version maybe changed by the first node.

[0129] For example, the network node 711, signals a CSI prediction performance report configuration for an AI / non-AI based CSI prediction model / algorithm to a UE, e.g. the communication device 730, based on which the UE derives one or more performance monitoring result(s), and reports the derived performance monitoring result(s) to the network node 711. The AI / non-AI based CSI prediction model / algorithm is implemented at the communication device 730, and it can be an Al-based scheme e.g., a neural network model trained by using a large dataset and then deployed at the communication device, or non-AI based scheme e.g., an Auto-Regression based algorithm, or a Kalman-filer based algorithm.

[0130] The network node 711 could refer to the gNB, e.g. the gNB-central unit (CU) or the gNB-distributed unit (DU), the Operation, administration and management (0AM), or a core network node, e.g. the Network Data Analytics Function (NWDAF). In the following the term network node can refer to any of the aforementioned entities. The CSI prediction model / algorithm is not limited to predicting the PMIs or raw channels for one or multiple future time instances. Examples of predicted CSI include predicted PMI(s), predicted raw channel(s), predicted RI(s), predicted RSRP(s), predicted CQI(s), predicted CRI(s), predicted top-K strongest beams, predicted top-K cells etc.

[0131] The disclosure describes methods for reporting configuration for performance monitoring outcome of an AI / non-AI based CSI prediction, either as a standalone report, or as part of a CSI report that contains predicted CSI and methods for signaling the proposed configurations. Embodiments herein may disclose methods for configuring and / or allocating the physical resources for the communication device 730 to transmit the performance monitoring report, according to the time domain behavior of the performance monitoring reports and the report quantities. Also, methods for reporting performance monitoring results in a 2 Part CSI framework are proposed.

[0132] Identification of the AI / non-AI based CSI prediction model(s) / algorithm(s) / feature(s) / functionality(ies) whose performance is / are to be monitored.

[0133] In an embodiment, the CSI-prediction model, algorithm, functionality, and / or feature whose performance to be monitored is indicated and signaled from NW to communication device via one or more of the following:

[0134] • A feature / feature-group ID (e.g., the feature / feature-group ID for Rel-18 type II CSI (eType2Doppler-r 18, eType2DopplerN4-r 18), the feature / feature-group ID for Rel-19 AI-CSI prediction), and / or

[0135] • A functionality ID (e.g., a functionality ID associated to the Rel-19 AI-CSI prediction feature), and / or

[0136] • A model ID (e.g., a logic AI-CSI prediction model for the Rel-19 AI-CSI prediction feature).

[0137] UE reporting performance monitoring results to NW according to NW configuration.

[0138] The following embodiments are described where the communication device 730 reports capability of supporting AI / non-AI based CSI prediction and the network configures to the communication device 730 the use of AI / non-AI based CSI prediction to predict the channel(s) in one or more future time instances and report the predicted information about the channel(s) with Rel-18 CSI reporting framework (i.e., the network configures the CSI-ReportConfig with codebookType set to ‘ typell-Doppler-r 18 ’).

[0139] • In one embodiment, the communication device 730 reports capability of AI / non-AI based CSI prediction and the network configures the use of AI / non-AI based CSI prediction to predict the channel(s) in one or more future time instances and report the predicted information about the channel(s) with a new Rel-19 CSI reporting framework, i.e., the network configures the CSI-ReportConfig with a new codebookType, e.g., say ‘typeII-Doppler-rl9 ’

[0140] Moreover, it is assumed that either the performance monitoring output (Type 1 performance monitoring) or the performance metrics (Type 3 performance monitoring) that is reported by the UE to the network has been quantized into bits, which can be included in the CSI report.

[0141] In the following, the term performance monitoring outcome refers to the performance monitoring output (Type 1 performance monitoring) or the performance metrics (Type 3 performance monitoring) of the AI / non-AI based CSI prediction for a configured AI / non-AI model / algorithm, which is reported back to the network.

[0142] Embodiments related to configuration for performance monitoring of AI / non-AI based CSI reporting.

[0143] Method 1: adding new value(s) for parameter reportQuantity in the CSI-ReportConfig IE, only one value can be configured for reportQuantity.

[0144] In a general embodiment, one or more new value(s) for parameter reportQuantity Ca Q included in the CSI-ReportConfig IE to configure the communication device 730 to include or not include the performance monitoring outcome from the performance evaluation of the AEnon-AI based CSI prediction in the CSI report, where the reporting can be periodic, semi-persistent or aperiodic (event triggered) configured with reportConfigType parameter.

[0145] • In a related embodiment, a new parameter, say, reportConfigType _predPM, is used for configuring the time domain behavior of reporting the performance monitoring outcome. Hence, the reporting of performance monitoring outcome and the reporting of the predicted CSI can be configured to have different time domain behaviors. E.g., Periodic reporting predicted CSI and aperiodic reporting the performance monitoring outcome. • In another related embodiment, the same reportConfigType parameter is used for configuring the reporting of performance monitoring outcome and predicted CSI. If both performance monitoring outcome and predicted CSI are carried in the same CSI report, then, they shall have the same time domain behavior.

[0146] In one dependent embodiment, a new value to signal only the performance monitoring outcome, say predPM, is included for the parameter reportQuantity in the CSI- ReportConfig IE. Accordingly, the network can configure one or more CSI-RS resource set(s) for performance monitoring (or data collection for training / fme-tuning), which can be used to compute the performance of the AI / non-AI based CSI prediction model / algorithm and include the performance monitoring output outcome in the CSI report.

[0147] • In a related embodiment, a new value to signal only the performance monitoring outcome, say predPM, is included for the parameter reportQuantity. In an optional embodiment, the CSI-reportConfig IE when configured with reportQuantity parameter set to value predPM may not be associated with any CSI-RS resource set(s) for resourcesForChannelMeasurement and / or csi-IM- ResourcesForlnterference. Accordingly, when the network configures reportQuantity = predPM, the UE computes the performance of the AI / non-AI based CSI prediction model / algorithm and included performance monitoring outcome based on predicted CSIs and the associated ground-truth labels from one or more CSI-RS resources sets for performance monitoring (or data collection for training / fme-tuning) in the CSI report. Note that when reportQuantity parameter in a CSI reporting configuration is set to value predPM, the resulting report is a standalone report that only carries the performance monitoring outcome. An example ASN.l change with respect to 3 GPP TS 38.331 VI 8.0.0 for introducing the new value for the reportQuantity as proposed in this embodiment is shown below. Changes over 3GPP TS 38.331 V18.0.0 are highlighted in bold text. In this embodiment, reportQuantity-r 19 is present, UE shall ignore reportQuantity (without suffix).

[0148] - ASN1 START

[0149] - TAG-CSI-REPORTCONFIG-START

[0150] CSI-ReportConfig ::= SEQUENCE { reportConfigld CSI-ReportConfigld, carrier ServCelllndex OPTIONAL, - Need S resourcesForChannelMeasurement CSI-ResourceConfigld, csi-IM-ResourcesForlnterference CSI-ResourceConfigld OPTIONAL, - Need R nzp-CSI-RS-ResourcesForlnterference CSI-ResourceConfigld OPTIONAL, - Need R reportQuantity CHOICE { none NULL, cri-RI-PMI-CQI NULL, cri-RI-il NULL, cri-RI-il-CQI SEQUENCE { pdsch-BundleSizeForCSI ENUMERATED {n2, n4} OPTIONAL - Need S

[0151] }, cri-RI-CQI NULL, cri-RSRP NULL, ssb-Index-RSRP NULL, cri-RI-LI-PMI-CQI NULL

[0152] [[ report Qu antity-rl9 predPM-r!9 OPTIONAL, - Need R

[0153] ]]

[0154] - TAG-CSI-REPORTCONFIG-STOP

[0155] - ASN1STOP

[0156] • In a related embodiment, when the network configures the communication device 730 to report only the performance monitoring outcome, the network can further configure a subset of subbands in the bandwidth part for which the performance monitoring outcome is computed by configuring the subset of subbands through csi-ReportingBand parameter. This can be used by the network to reduce the computing overhead of the communication device 730, where the computation of the performance monitoring outcome over a subset of subbands can provide an indication for the performance of the AI / non-AI based CSI prediction model / algorithm.

[0157] • In another dependent embodiment, a new value to signal the performance monitoring outcome along with the legacy parameters CRI, RI, PMI, CQI, say cri- RI-PMI-CQI-predPM, is included for the parameter reportQuantity . Accordingly, the network can configure one or more CSI-RS resource sets through the legacy parameters resourcesForChannelMeasurement and / or csi-IM- ResourcesForlnterference for channel measurement and / or interference (in CSI- reportConfig), where the communication device 730 can include the performance monitoring outcome from the performance evaluation of AI / non-AI based CSI prediction model / algorithm based on predicted channel from one or more past CSI- RS resources sets for performance monitoring (or data collection for training / fine- tuning) in the CSI report. When the reportQuantity parameter is set to the new value (e.g., cri-RI-PMI-CQI-predPM) , UE shall report both the performance monitoring outcome together with the predicted CSI related info (e.g., CRI, RI. PMI, CQI) in the periodic / semi-persistent / aperiodic CSI report. An example ASN.l change with respect to 3GPP TS 38.331 V18.0.0 for introducing the new value for the reportQuantity as proposed in this embodiment is shown below. Changes over 3GPP TS 38.331 V18.0.0 are highlighted in bold text. In this embodiment, reportQuantity-r 19 is present, UE shall ignore reportQuantity (without suffix).

[0158] CSI-ReportConfig information element

[0159] - ASN1 START

[0160] - TAG-CSI-REPORTCONFIG-START CSI-ReportConfig : := SEQUENCE { reportConfigld CSI-ReportConfigld, carrier ServCelllndex OPTIONAL, — Need S resourcesForChannelMeasurement CSI-ResourceConfigld, csi-IM-ResourcesForlnterference CSI-ResourceConfigld OPTIONAL, — Need R nzp-CSI-RS-ResourcesForlnterference CSI-ResourceConfigld OPTIONAL, - Need R reportQuantity CHOICE { none NULL, cri-RI-PMI-CQI NULL, cri-RI-il NULL, cri-RI-il-CQI SEQUENCE { pdsch-BundleSizeForCSI ENUMERATED {n2, n4} OPTIONAL - Need S cri-RI-CQI NULL, cri-RSRP NULL, ssb-Index-RSRP NULL, cri-RI-LI-PMI-CQI NULL

[0161] [[ reportQuantity-rl9 cri-RI-PMI-CQI-predPM-rl9 OPTIONAL, — Need R

[0162] ]]

[0163] }

[0164] - TAG-CSI-REPORTCONFIG-STOP

[0165] - ASN1STOP

[0166] • In another dependent embodiment, when the veportQiiantity parameter is set to “none”, the communication device 730 shall not feedback CSI report to the network. Hence, no performance monitoring outcome nor predicted CSI related info will be feedback from the UE to the network. However, the communication device 730 can still monitor the performance of its CSI prediction model(s) / algorithm(s) using the CSI-RS resource sets configured by the NW.

[0167] Method 2: adding new value(s) for parameter reportQuantity in the CSI-ReportConfig IE, more than one value can be configured for reportQuantity.

[0168] In another dependent embodiment, the parameter reportQuantity is configured with one or two values, while the first value indicates the report quantity for CSI related information, e.g., PMI, RI, CQI, CRI, SSBRI, Ll-RSRP, etc., and the second value indicates the report quantity for the performance monitoring outcome, e.g., predPM- SGCS, predPM-NMSE, etc.. The candidate of report quantities for performance monitoring outcome can be defined based on the performance monitoring result format / content. Examples include per sample or statistics of SGCS, NMSE, prediction accuracy, confidence level, etc. An example ASN.1 change with respect to 3GPP TS 38.331 V18.0.0 for introducing the new value for the reportQuantity as proposed in this embodiment is shown below. Changes over 3GPP TS 38.331 V18.0.0 are highlighted in bold text. In this embodiment, a UE may be configured with both reportQuantity and reportQuantity-r 19 in a CSI reporting configuration.

[0169] CSI-ReportConfig information element

[0170] - ASN1 START

[0171] - TAG-CSI-REPORTCONFIG-START

[0172] CSI-ReportConfig : := SEQUENCE { reportConfigld CSI-ReportConfigld, carrier ServCelllndex OPTIONAL, — Need S resourcesForChannelMeasurement CSI-ResourceConfigld, csi-IM-ResourcesForlnterference CSI-ResourceConfigld OPTIONAL, — Need R nzp-CSI-RS-ResourcesForlnterference CSI-ResourceConfigld OPTIONAL, - Need R reportQuantity CHOICE { none NULL, cri-RI-PMI-CQI NULL, cri-RI-il NULL, cri-RI-il-CQI SEQUENCE { pdsch-BundleSizeForCSI ENUMERATED {n2, n4} OPTIONAL - Need S

[0173] }, cri-RI-CQI NULL, cri-RSRP NULL, ssb-Index-RSRP NULL, cri-RI-LI-PMI-CQI NULL

[0174] [[ reportQuantity2-rl9 CHOICE { none NULL, predPM-SGCS NULL, predPM-NMSE NULL, predPM-accuracy NULL, predPM-confidencelevel NULL

[0175] }, OPTIONAL, - Need R

[0176] ]] - TAG-CSI-REPORTCONFIG-STOP

[0177] - ASN1STOP

[0178] • If the first value (e.g., the value of reportQuantity in the above example) is set to “none ”, then, UE shall not include CSI related parameters in its CSI report.

[0179] • If the second value (e.g., the value of reportQuantity2 in the above example) is set to “none ”, then, the communication device 730 shall not include performance monitoring outcome in its CSI report.

[0180] • If both values (e.g., the values of reportQuantity and reportQuantity2 in the above example) are set to “none ”, then, the communication device 730 shall not transmit the CSI report.

[0181] Method 3: adding new report quantity param eter(s) in the CSI-ReportConfig IE for configuring the report quantity for performance monitoring outcome.

[0182] In another embodiment, a new parameter, say reportQuantity PM, is added in the CSI-reportConfig RRC IE to configure the report quantity for the performance monitoring outcome. Candidate values for the new parameter reportQuantity PM defined based on the performance monitoring result format / content. Examples include per sample or statistics of SGCS, NMSE, prediction accuracy, confidence level, etc.

[0183] • If the legacy parameter reportQuantity is set to “none ”, then, the communication device 730 shall not include CSI related parameters in its CSI report.

[0184] • If the new parameter reportQuantity PM is set to “none ”, then, the communication device 730 shall not include performance monitoring outcome in its CSI report.

[0185] • If both parameters are set to “none ”, then, the communication device 730 shall not transmit the CSI report.

[0186] Applicable for all three methods:

[0187] In an embodiment, a second DL control signaling, e.g., a RRC message, a MAC CE and / or a DCI signaling, which is different from the CSI-reportConfig RRC message, is sent from the NW to the communication device 730, where this second DL control signaling together with the CSI-reportConfig RRC signaling are used by the communication device 730 to determine whether it shall only report the predicted CSI related info in the CSI report, or only report the performance monitoring outcome in the CSI report, or report both in the CSI report.

[0188] • As an example, the communication device 730 is configured with the method(s) described above to include both predicted CSI and the performance monitoring outcome in its CSI report, and the reportConfigType parameter is set to aperiodic. A DCI signaling used for scheduling the CSI report includes a bit-field, which indicates whether the communication device 730 shall include performance monitoring outcome into this CSI report or not.

[0189] • As another example, the communication device 730 is configured with the method(s) described above to include both predicted CSI and the performance monitoring outcome in its CSI report, and the reportConfigType parameter is set to periodic. By default, the communication device 730 reports predicted CSI together with the performance monitoring outcome in the CSI report periodically. An RRC / MAC-CE / DCI signaling from the network to the communication device 730 can be used for requesting the communication device 730 to stop reporting the performance monitoring outcome in its periodic CSI report. When receiving the RRC / MAC-CE / DCI signaling from the network, the communication device 730 shall not include performance monitoring outcome in its proceeding periodic CSI report, or the communication device 730 shall not include performance monitoring outcome in its proceeding periodic CSI report for a certain duration (if a time window is configured).

[0190] In another embodiment, time restriction on the number of CSI prediction measurements to generate the performance monitoring outcome is configured by including a new parameter in CSI-ReportConfig IE.

[0191] • In a related embodiment, the new parameter, say timeRestrictionForPredictionPerformanceMonitoring, takes integer values defining the number of past CSI prediction measurements used to compute the performance metrics for AI / non-AI based CSI prediction. For example, the parameter can take values of { 1, 2, 4, 8}, and if the network configures the communication device 730 with a value, say 4, then the communication device 730 uses the last 4 predicted measurements to compute the performance metrics for AI / non-AI based CSI prediction. The possible value that can be configured for the UE may depend on the UE capability. For example, the communication device 730 with a first capability may be configured with one (or more) value from a first set of values and the communication device 730 with a second capability may be configured with one (or more) value from a second set of values. The UE capability related to this aspect may be an independent parameter in the CSI prediction capability report or may be derived from the UE capability report of the inference operation.

[0192] Aperiodic CSI reporting in PUSH

[0193] In one case, performance outcome is included with legacy parameters like CRI, RI, PMI and CQI in CSI report, for example, when reportQuantity = cri-RI-PMI-CQI-predPM.

[0194] In a general embodiment, when the performance monitoring outcome is configured to be aperiodically reported along with one or more legacy parameters like CRI, RI, PMI and CQI (with Rel-18 Type II framework) computed from the CSI-RS for channel measurement and interference measurement resource, where one or more CSI prediction measurements is (are) used for calculating the performance metrics for AI / non-AI based CSI prediction model / algorithm, the performance monitoring outcome is included in the CSI report carrying the CQI, RI, estimated PMI and CQI. The CSI prediction measurement(s) used for calculating the performance metrics may be, a. derived from an aperiodic CSI-RS resource set triggered for performance monitoring with current CSI-ReportConfig, or, b. consists of one or more past CSI measurement s) derived and stored from CSI-RS resource sets configured, periodically, semi-persistently, or aperiodically, for performance monitoring or data collection, or, c. derived from one or more CSI measurements derived from both the above methods.

[0195] Note that for Rel-18 Type II predicted PMI feedback on PUSCH, a CSI report comprises of two parts. Part 1 has a fixed payload size and is used to identify the number of information bits in Part 2. Part 1 shall be transmitted in its entirety before Part 2. For Rel-18 Type II CSI feedback, Part 1 contains RI (if reported), CRI (if reported), CQI for the first codeword (if reported and RI < 4), and the total number of reported non-zero amplitude coefficients across layers. The fields of Part 1, i.e., RI (if reported), CQI, and the total number of reported non-zero amplitude coefficients across layers, are separately encoded. Part 2 contains PMI (if reported) and contains the CQI for the second codeword (if reported) when RI is larger than 4. Part 1 and 2 are separately encoded. In one embodiment, since the performance monitoring outcome can be used by the network to firstly decide how accurate the legacy channel parameters (RI, CQI and PMI) reported in the CSI report, the performance monitoring outcome is included in the Part 1 of the CSI report (and is separately encoded w.r.t. other fields in Part 1). Accordingly, the network can firstly decode the Part 1 of the CSI report and based on the performance monitoring outcome for the AI / non-AI based CSI prediction, decide either,

[0196] ■ In one sub-embodiment, to use the legacy CSI parameters reported in the CSI report, if the performance of AI / non-AI based prediction is good enough, or configure / trigger another aperiodic CSI report, if the performance of AI / non-AI based prediction is not good enough, with either a different Al model for CSI prediction or fall back to non- AI based CSI prediction or a different non-AI based CSI prediction algorithm.

[0197] ■ In another sub-embodiment, to use the legacy CSI parameters reported in the CSI report, irrespective of performance of AI / non-AI based prediction, and configure subsequent CSI reports, if the performance of AI / non-AI based prediction is not good enough, with either a different Al model for CSI prediction or fall back to non- AI based CSI prediction or a different non-AI based CSI prediction algorithm.

[0198] In another embodiment, the performance monitoring outcome is included in the Part 2 of the CSI report, with highest priority in CSI Part 2 and is separately encoded with fields in same priority level, i.e., wideband CSI parameters and / or second CQI if RI > 4. Accordingly, the network can decode the performance of the AI / non-AI based CSI prediction model / algorithm from Part 2 of the CSI report and based on the performance monitoring outcome for the AI / non-AI based CSI prediction, decide either,

[0199] ■ In one sub-embodiment, to use the legacy CSI parameters reported in the CSI report if the performance of AI / non-AI based prediction is good enough or configure another aperiodic CSI report if the performance of AI / non-AI based prediction is not good enough, with either a different Al model for CSI prediction or fall back to non- AI based CSI prediction or a different non-AI based CSI prediction algorithm.

[0200] ■ In another sub-embodiment, to use the legacy CSI parameters reported in the CSI report irrespective of performance of AI / non-AI based prediction and configure subsequent CSI reports if the performance of AI / non-AI based prediction is not good enough, with either a different Al model for CSI prediction or fall back to non-AI based CSI prediction or a different non-AI based CSI prediction algorithm.

[0201] In this case, the CSI Part 1 includes a new field to indicate the size for reporting the performance monitoring outcome in Part 2 of the CSI report.

[0202] In another case, performance outcome is only included in CSI report, for example, when reportQuantity =predPM.

[0203] In a related embodiment, when the performance outcome is configured to be aperiodically reported (in PUSCH) without any legacy parameters like CRI, RI, PMI and CQI, where one or more CSI prediction measurements is (are) used for calculating the performance metrics for AI / non-AI based CSI prediction model / algorithm, the performance monitoring outcome in included in the CSI report which consist of a single part. The CSI prediction measurement(s) used for calculating the performance metrics can either be, a. derived from an aperiodic CSI-RS resource set triggered for performance monitoring with current CSI-ReportConfig, or, b. consists of one or more past CSI measurement s) derived (and stored) from CSI-RS resource sets configured (periodically, semi-persistently, or aperiodically) for performance monitoring or data collection, or, c. derived from one or more CSI measurements derived from both the above methods.

[0204] In this embodiment, the performance outcome is included in a single part CSI report. Accordingly, the network can decode the performance of the AI / non-AI based CSI prediction model / algorithm from the CSI report and based on the performance monitoring outcome for the AI / non-AI based CSI prediction, decide either to configure the subsequent aperiodic CSI report with the existing AI / non-AI model / algorithm for CSI prediction if the performance of AI / non-AI based prediction is good enough, or configure the subsequent CSI reports with either a different Al model for CSI prediction or fall back to non-AI based CSI prediction or a different non-AI based CSI prediction algorithm if the performance of AI / non-AI based prediction is not good enough.

[0205] Periodic CSI reporting in PUCCH

[0206] In one embodiment, when the communication device 730 is configured with a periodic CSI-RS resource set for performance monitoring and / or data collection for Al-model training, only the performance monitoring outcome is configured to be periodically reported in PUCCH, where one or more CSI prediction measurements is (are) used for calculating the performance metrics for AI / non-AI based CSI prediction. The CSI prediction measurement(s) used for calculating the performance metrics can consist of one or more past CSI measurement(s) derived and stored from CSI-RS resource sets configured periodically for performance monitoring or data collection.

[0207] Note that only performance monitoring outcome is configured to be reported to limit the payload since the periodic CSI report is feedback in PUCCH. Further, the periodic CSI- RS resource set for performance monitoring and / or data collection for training and the CSI report with only performance monitoring outcome are always assumed to be present and active once configured by RRC.

[0208] Accordingly, the network upon decoding the performance of the AI / non-AI based CSI prediction model / algorithm from the CSI report, decide either to configure the subsequent periodic CSI report with the existing AI / non-AI model / algorithm for CSI prediction if the performance of AI / non-AI based prediction is good enough, or configure the subsequent CSI reports with either a different Al model for CSI prediction or fall back to non-AI based CSI prediction or a different non-AI based CSI prediction algorithm if the performance of AI / non- AI based prediction is not good enough.

[0209] Semi-persistent CSI reporting in PUCCH / PUSH

[0210] In a general embodiment, when the communication device 730 is configured with a periodic / semi-persistent CSI-RS resource set for performance monitoring and / or data collection for Al-model training, the performance monitoring outcome is configured to be semi-persistently reported in PUCCH / PUSCH, where one or more CSI prediction measurements is (are) used for calculating the performance metrics for AI / non-AI based CSI prediction model / algorithm. The CSI prediction measurement(s) used for calculating the performance metrics can consists of one or more past CSI measurement s) derived and stored from past CSI-RS resource sets configured periodically and / or semi-persistently for performance monitoring or data collection.

[0211] Note that semi-persistent CSI reporting can either be PUCCH-based or PUSCH-based, where each option is targeting different use cases. PUCCH-based semi -persistent reporting is more akin to periodic CSI reporting but with the additional functionality to be activated / deactivated with MAC CE and is intended for a low-payload CSI. PUSCH-based semi-persistent reporting on the other hand can be seen as a multi-shot aperiodic CSI report, which can encompass more high-payload CSI, such as Type II CSI.

[0212] ■ In one embodiment, for the PUCCH-based semi-persistent reporting activated / deactivated with MAC CE message, only the performance monitoring outcome based on one or more past CSI prediction measurements may be included in the CSI report to have low-payload over PUCCH.

[0213] ■ In another embodiment, for the PUSCH-based semi-persistent reporting activated / deactivated with MAC CE message, the performance monitoring outcome based on one or more past CSI prediction measurements may be included along with legacy CSI parameters like CQI, RI, PMI, LI and CQI, in the CSI report.

[0214] Accordingly, the network upon decoding the performance of the AI / non-AI based CSI prediction model / algorithm from the CSI report, decide either to configure the subsequent semi-persistent CSI report with the existing AI / non-AI model / algorithm for CSI prediction if the performance of AI / non-AI based prediction is good enough, or configure the subsequent CSI reports with either a different Al model for CSI prediction or fall back to non-AI based CSI prediction or a different non-AI based CSI prediction algorithm if the performance of AI / non-AI based prediction is not good enough.

[0215] Embodiments related to omission of part of the performance monitoring output from the CSI report.

[0216] In one embodiment, where only the performance monitoring output corresponding to the AI / non-AI based CSI prediction is included in the CSI report and when the length of corresponding bits in the CSI report is > 1, the communication device 730 can decide to omit certain bits from the CSI report to fit the PUCCH / PUSCH payload.

[0217] For example, consider the performance monitoring output is a bitmap b = ... , btW4] of length N4, where N4is the length of the future CSI predicted by the Al / non- AI based model / algorithm, such that each the t -th bit in the bitmap corresponds to performance of the AI / non-AI model / algorithm in predicting the j-th future CSI.

[0218] Accordingly, the omission rule can be specified such that the communication device 730 can omit reporting bits from b in the increasing order of priority, where the highest priority can be given to bits corresponding to the nearest predicted CSI and lowest priority to the farthest predicted CSI.

[0219] The following are some embodiments related to prediction monitoring reporting in 2-part CSI.

[0220] In one embodiment, CSI Part 1 may contain information identifying the number of time instances, for which a performance monitoring metric(s) is reported. For example, the communication device 730 may be configured to monitor the prediction performance in terms of SGCS for 4 time instances, then the UE may report only a subset, say 2 time instances, where the SGCS values are below a threshold.

[0221] In one embodiment, CSI Part 2 may report the performance monitoring metric(s) for the time instances that are identified by CSI Part 1.

[0222] In one embodiment, CSI Part 1 may contain information identifying the number of layers or rank, for which a performance monitoring metric(s) is reported.

[0223] For example, the communication device 730 may be configured to monitor the prediction performance in terms of SGCS for up to rank 4, then the communication device 730 may report only for rank 2 if there is no additional performance gain for higher ranks. Or, the communication device 730 may only report for layer 2 if rank 2 is selected by the communication device 730, as only the SGCS for layer 2 falls below a threshold.

[0224] In one embodiment, CSI Part 2 may report the performance monitoring metric(s) for the layers that are identified by CSI Part 1.

[0225] In another embodiment, CSI Part 1 may contain information identifying the types of performance monitoring reported by the UE.

[0226] For example, the communication device 730 may be configured with {SGCS, confidence, NMSE, . . . } for performance monitoring report, where the communication device 730 can choose to report one or multiple of these e.g., the communication device 730 only reports those metrics that are below a threshold configured by the NW.

[0227] In one embodiment, CSI Part 2 may report the performance monitoring metric(s) of the identified types by CSI Part 1.

[0228] A method performed by a network node e.g. the network node 711 for configurating physical resources for a UE to report a performance monitoring report in a wireless communication system 300 will be described with reference to Figure 8. The performance monitoring report may comprise performance related information monitoring output / performance metrics for the AI / non-AI based CSI-prediction model / algorithm. The method comprises the following actions which may be performed in any suitable order.

[0229] Action 810

[0230] The network node 711 sends a first signaling message to the communication device 730, for configuring the communication device 730 to send a report on a performance related information associated to a computational model for predicting Channel State Information, CSI, in one or multiple future time instances. The first signaling message may be a configuration message for configurating physical resources for the communication device 730 to transmit or report performance related information, e.g. for monitoring output / performance metrics for the AI / non-AI based CSI-prediction model / algorithm. The first signaling message may be a CSI prediction performance monitoring report configuration for configuring the communication device 730 to report the performance monitoring results for an Al or non- Al based CSI-prediction model used for generating a predicted CSI. The first signaling message may be sent using an RRC message, and / or using an MAC CE, and / or as a type of uplink control information (UCI) on PUSCH / PUCCH. The performance related information may be defined as a new UCI type, or as part of the CSI report carrying the predicted CSI.

[0231] The first signaling message may comprise any one or a combination of the following: a) a CSI prediction performance monitoring report configuration for configuring the communication device (730) to report the performance monitoring results for an Artificial Intelligence, Al, or non- Al based CSI-prediction model used for generating a predicted CSI; b) a physical resource configuration for configuring physical resources for the communication device (730) to transmit a performance monitoring report.

[0232] The first signaling message is sent by using any one or a combination of the following: a) a Radio Resource Control (RRC) message; b) a Media Access ControlControl Element (MAC CE); c) a Downlink Control Information (DCI) signaling.

[0233] The performance related information may comprise at least one of performance monitoring result(s), performance metric(s), associated to an Al or non-AI based CSI prediction model. The Al or non-AI based CSI prediction model may be associated to any one of a feature identifier, a feature group, an AI / ML model identifier, a functionality identifier.

[0234] The RRC message may be any one of the following: a. the same RRC message used for configuring the communication device 730 to report predicted CSI, b. a separate RRC message that is different from the one used for configuring the communication device 730 to report predicted CSI, c. the same RRC message used for configuring performance monitoring data collection at the communication device 730. The RRC message may indicate at least one of: a format of the performance related information, a content of the performance related information, for the Al or non- Al based CSI-prediction model.

[0235] The RRC message may comprise a time window configuration during which the performance related information shall be reported.

[0236] The MAC CE may be any one of the following: a. the same MAC CE used for requesting the communication device 730 to report the predicted CSI, b. a separate MAC CE defined for requesting the communication device 730 to report the performance related information.

[0237] The DCI signaling may be any one of the following: a. the same DCI signaling used for requesting the communication device 730 to report the predicted CSI, b. a separate DCI signaling defined for requesting the communication device 730 to report the performance related information.

[0238] Action 820

[0239] The network node 711 receives a report message with related performance information from the UE 730 based on the first signaling message, e.g. the configuration message.

[0240] According to some embodiments herein, upon decoding the received report carrying the performance related information, the method may further comprise at least one of the following actions: a. sending a second signaling message to the communication device 730 for configuring subsequent CSI reports with the existing Al or non- Al model for CSI prediction, if the performance of the Al or non-AI based prediction fulfils a performance criterion; b. sending a second signaling message to the communication device 730 for configuring subsequent CSI reports with a different Al or non-AI model for CSI prediction if the performance of the Al or non-AI based prediction does not fulfil a performance criterion; c. sending a second signaling message to the communication device 730 for configuring subsequent CSI reports with fall back to non-AI based CSI prediction if an Al-based CSI prediction is currently configured and the performance of the Al based CSI prediction does not fulfil a performance criterion; d. sending a second signaling message to the communication device 730 for configuring subsequent CSI reports with fall back to CSI measurement reporting without prediction.

[0241] According to some embodiments herein, the method may further comprise signaling to the communication device 730 to stop reporting performance related information for an Al or non- Al based CSI-prediction model. The signaling to the communication device 730 to stop reporting performance related information may be implemented by any one of the following: a. using a time window configured in a CSI prediction performance report configuration message. b. using a different RRC message defined for stopping the communication device 730 reporting the performance related information. c. using the same MAC CE used for deactivating a predicted CSI report, or a separate MAC CE defined for stopping the communication device 730 reporting the performance related information. d. using the same DCI used for requesting the communication device 730 to report a predicted CSI, or a separate DCI defined for stopping the communication device 730 reporting the performance related information.

[0242] According to some embodiments herein, the configuration for the communication device 730 to start or stop reporting may be based on a channel condition. The channel condition may be any one of a speed of the communication device 730 being above a certain threshold, a delay spread estimate being outside a certain region, a doppler spread estimate being outside a certain region.

[0243] According to some embodiments herein, the physical resource configuration for configuring physical resources for the communication device 730 to transmit a performance monitoring report may be based on time domain behavior of the performance monitoring reports and quantities of the reports.

[0244] According to some embodiments herein, the performance related information may be configured to be reported by any one of the following: a. using an RRC message; b. using an MAC CE; c. using a type of Uplink control information (UCI) on physical uplink control channel (PUCCH), or Physical Uplink Shared Channel (PUSCH), wherein the performance related information is defined as a new UCI type, or the performance related information is defined as part of the CSI report carrying the predicted CSI.

[0245] According to some embodiments herein, the CSI prediction performance report configuration may configure the communication device 730 to report back the performance related information for an Al or non-AI based CSI-prediction model along with predicted CSI related parameters of any one of Channel Rank Indicator (CRI), Channel Quality Indicator (CQI), Precoding Matrix Indicatorm (PMI), Rank Indicator (RI), Layer Indicator (LI).

[0246] A method performed by the communication device 730 e.g. a UE, for handling communication, e.g. reporting a performance monitoring report, in a wireless communication system 700 will be described with reference to Figure 9. The performance monitoring report may comprise performance related information monitoring output and / or performance metrics for the Al and / or non-AI based CSI-prediction model / algorithm. The method comprises the following actions which may be performed in any suitable order.

[0247] Action 910

[0248] The communication device 730 receives the first signaling message from the network node 711 for reporting performance related information associated to the computational model for predicting CSI in the one or multiple future time instances. The first signaling message may be the physical resource configuration for configuring the physical resources for the communication device 730 to transmit the performance monitoring report.

[0249] The first signaling message may be the CSI prediction performance monitoring report configuration message for configuring the communication device 730 to, report the performance monitoring output and / or performance metrics for the Al and / or non-AI based CSI-prediction model and / or algorithm. The first signaling message may be sent using an RRC message, and / or using an MAC CE, and / or as a type of UCI on PUSCH / PUCCH. The performance related information may be defined as a new UCI type, or as part of the CSI report carrying the predicted CSI.

[0250] The first signaling message may comprise any one or a combination of the following: a. a CSI prediction performance monitoring report configuration for configuring the communication device 730 to report the performance monitoring results for an Artificial Intelligence, Al, or non-AI based CSI-prediction model used for generating a predicted CSI; b. a physical resource configuration for configuring the physical resources for the communication device 730 to transmit a performance monitoring report.

[0251] The first signaling message may be received via any one or a combination of the following: a) a Radio Resource Control (RRC) message; b) a Media Access Control Control Element (MAC CE); c) a Downlink Control Information (DCI) signaling.

[0252] The RRC message may be any one of the following: a. the same RRC message used for configuring the communication device 730 to report predicted CSI, b. a separate RRC message that is different from the one used for configuring the communication device 730 to report predicted CSI, c. the same RRC message used for configuring performance monitoring data collection at the communication device 730.

[0253] The RRC message may indicate at least one of: a format of the performance related information, a content of the performance related information, for the Al or non-AI based CSI-prediction model.

[0254] The RRC message comprises a time window configuration during which the performance related information shall be reported.

[0255] The MAC CE may be any one of the following: a. the same MAC CE used for requesting the communication device 730 to report the predicted CSI, b. a separate MAC CE defined for requesting the communication device 730 to report the performance related information.

[0256] The DCI signaling may be any one of the following: a. the same DCI signaling used for requesting the communication device 730 to report the predicted CSI, b. a separate DCI signaling defined for requesting the communication device 730 to report the performance related information.

[0257] The performance related information may comprise at least one of performance monitoring result(s) and performance metric(s) associated to an Al or non- Al based CSI prediction model.

[0258] The Al or non-AI based CSI prediction model may be associated to any one of a feature identifier, a feature group, an AI / ML model identifier, and a functionality identifier.

[0259] Action 920

[0260] The UE 730 sends a report message with related performance information to the network node based on the first signaling message.

[0261] The performance related information may be reported by using any one of the following: a. a Radio Resource Control (RRC) message; b. a Media Access Control Control Element, (MAC CE); c. a type of Uplink control information (UCI) on physical uplink control channel (PUCCH) or Physical Uplink Shared Channel (PUSCH). The performance related information may be defined as a new UCI type, or the performance related information may be defined as part of the CSI report carrying the predicted CSI.

[0262] The performance monitoring results for an Al or non-AI based CSI-prediction model may be sent along with the predicted CSI related parameters of any one of CSI reference signal Resource Indicator (CRI), Channel Quality Indicator (CQI), Precoding Matrix Indicator (PMI) Rank Indicator (RI), Layer Indicator (LI).

[0263] The report on the performance related information may include a performance related information for an Al or non-AI based CSI-prediction model in a CSI report when the CSI report is sent periodically or semi-persistently over a physical uplink control channel ( PUCCH), or a Physical Uplink Shared Channel (PUSCH).

[0264] The performance related information may include the performance related information for an Al or non-AI based CSI-prediction model with or without the predicted CSI related reporting quantities in the CSI report when the report is sent aperiodically over a Physical Uplink Shared Channel, PUSCH.

[0265] A 2 part CSI structure may be used for reporting performance monitoring results, where CSI part 1 may comprise information identifying the number of time instances, for which a performance monitoring metric(s) is reported, and CSI Part 2 may comprise the performance monitoring metric(s) for the time instances that are identified by the CSI Part 1. Alternatively, CSI part 1 may comprise information identifying the rank or the number of layers, for which a performance monitoring metric(s) is reported, and CSI Part 2 may comprise the performance monitoring metric(s) for the layers that are identified by CSI Part 1. Alternatively, CSI part 1 may comprise information identifying the type(s) of performance monitoring metric(s) that is (are) reported, and CSI Part 2 comprises the performance monitoring metric(s) for the identified types by CSI Part 1.

[0266] According to some embodiments herein, the method may further comprise receiving a second signaling message to stop reporting performance related information for an Al or non- AI based CSI-prediction model.

[0267] The second signaling message to stop reporting performance related information may be implemented by any one of the following: a. a time window configured in a CSI prediction performance report configuration message. b. a different RRC message defined for stopping the communication device (730) reporting the performance related information. c. the same MAC CE used for deactivating a predicted CSI report, or a separate MAC CE defined for stopping the communication device (730) reporting the performance related information. d. the same DCI used for requesting the communication device (730) to report a predicted CSI, or a separate DCI defined for stopping the communication device (730) reporting the performance related information.

[0268] To perform the method in the network node 711, the network node 711 may comprise modules or functions as shown in Figure 10. The network node 711 may comprise a receiving module 1010, a transmitting module 1020, a determining module 1030, a processing module 1040, a memory 1050 etc.

[0269] The network node 711 is configured to perform any one of the Actions 810-820 described above.

[0270] The network node 711 is configured to send a first signaling message to the communication device 730, for configuring the communication device 730 to send a report on a performance related information associated to a computational model for predicting Channel State Information, CSI, in one or multiple future time instances. The first signaling message may be a configuration message for configurating physical resources for the communication device 730 to transmit or report performance related information, e.g. for monitoring output / performance metrics for the AI / non-AI based CSI-prediction model / algorithm. The first signaling message may be a CSI prediction performance monitoring report configuration for configuring the communication device 730 to report the performance monitoring results for an Al or non-AI based CSI-prediction model used for generating a predicted CSI. The first signaling message may be sent using an RRC message, and / or using an MAC CE, and / or as a type of uplink control information (UCI) on PUSCH / PUCCH. The performance related information may be defined as a new UCI type, or as part of the CSI report carrying the predicted CSI.

[0271] The network node 711 is configured to receive a report message with related performance information from the UE 730 based on the first signaling message, e.g. the configuration message.

[0272] The methods according to embodiments herein may be implemented through one or more processors, such as the processor 1060 in the network node 711 together with computer program code for performing the functions and actions of the embodiments herein. The program code mentioned above may also be provided as a computer program product, for instance in the form of computer readable medium or a data carrier 1080 carrying computer program code 1070, as shown in Figure 10, for performing the embodiments herein when being loaded into the network node 711. One such carrier may be in the form of a CD ROM disc. It is however feasible with other data carriers such as a memory stick. The computer program code may furthermore be provided as pure program code on a server or a cloud and downloaded to the network node 711.

[0273] To perform the method in the communication device 730, e.g. UE, the communication device 730 may comprise modules or functions as shown in Figure 11. The communication device 730 may comprise a receiving module 1110, a transmitting module 1120, a determining module 1130, a processing module 1140, a memory 1150 etc.

[0274] The communication device 730 is configured to perform any one of the Actions 910- 920 described above.

[0275] The communication device 730 is configured to receive the first signaling message from the network node 711 for reporting performance related information associated to the computational model for predicting CSI in the one or multiple future time instances. The first signaling message may be the physical resource configuration for configuring the physical resources for the communication device 730 to transmit the performance monitoring report. The first signaling message may be the CSI prediction performance monitoring report configuration message for configuring the communication device 730 to, report the performance monitoring output and / or performance metrics for the Al and / or non-AI based CSI-prediction model and / or algorithm. The first signaling message may be sent using an RRC message, and / or using an MAC CE, and / or as a type of UCI on PUSCH / PUCCH. The performance related information may be defined as a new UCI type, or as part of the CSI report carrying the predicted CSI.

[0276] The communication device 730 is configured to send a report message with related performance information to the network node based on the first signaling message.

[0277] The methods according to embodiments herein may be implemented through one or more processors, such as the processor 1160 in the communication device 730 together with computer program code for performing the functions and actions of the embodiments herein. The program code mentioned above may also be provided as a computer program product, for instance in the form of computer readable medium or a data carrier 1180 carrying computer program code 1170, as shown in Figure 11, for performing the embodiments herein when being loaded into the communication device 730. One such carrier may be in the form of a CD ROM disc. It is however feasible with other data carriers such as a memory stick. The computer program code may furthermore be provided as pure program code on a server or a cloud and downloaded to the communication device 730.

[0278] EXAMPLE EMBODIMENTS

[0279] 1. A method for configurating physical resources for transmitting performance related information monitoring output / performance metrics for the AI / non-AI based CSI- prediction model / algorithm, wherein the performance related information can be configured to be reported, e.g., using an RRC message, and / or using an MAC CE, and / or as a type of UCI on PUSCH / PUCCH, and wherein the performance related information is defined as a new UCI type, or the performance related information is defined as part of the CSI report carrying the predicted CSI.

[0280] 2. The method according to embodiment 1, wherein the CSI prediction performance report configuration configures the UE to report back the performance related information for the AI / non-AI based CSI-prediction model / algorithm along with predicted CSI related parameters like CRI, CQI, PMI, RI, LI.

[0281] 3. The method according to embodiment 1, based on the configured time-domain behavior of the CSI report carrying the predicted CSI and / or the CSI prediction performance report configuration, the UE feedback the CSI-report over PUCCH / PUSCH which includes performance related information for the AI / non-AI based CSI-prediction model / algorithm.

[0282] 4. The method according to embodiment 1, where in UE only includes the performance related information for the AI / non-AI based CSI-prediction model / algorithm in the CSI report when the report is feedback periodically or semi-persistently over PUCCH / PUSCH.

[0283] 5. The method according to embodiment 1, where in UE includes the performance related information for the AI / non-AI based CSI-prediction model / algorithm with or without the predicted CSI related parameters (like CRI, CQI, PMI, RI, LI) in the CSI report when the report is feedback aperiodically over PUSCH.

[0284] 6. The method according to any one of embodiments 1-5, wherein a 2 part CSI structure is used for reporting performance monitoring results.

[0285] 7. The method according to embodiment 6, wherein CSI part 1 contains information identifying the number of time instances, for which a performance monitoring metric(s) is reported.

[0286] 8. The method according to embodiment 7, wherein CSI Part 2 reports the performance monitoring metric(s) for the time instances that are identified by CSI Part 1.

[0287] 9. The method according to embodiment 6, wherein CSI part 1 contains information identifying the rank or the number of layers, for which a performance monitoring metric(s) is reported.

[0288] 10. The method according to embodiment 9, wherein CSI Part 2 reports the performance monitoring metric(s) for the layers that are identified by CSI Part 1.

[0289] 11. The method according to embodiment 6, wherein CSI part 1 contains information identifying the type(s) of performance monitoring metric(s) that is (are) reported.

[0290] 12. The method according to embodiment 11, wherein CSI Part 2 reports the performance monitoring metric(s) of the identified types by CSI Part 1.

[0291] 13. A network node (311) configured to perform the method according to any one of the embodiments 1-12.

[0292] 14. A communication device (UE 330) configured to perform the method according to any one of the embodiments 1-12.

[0293] Abbreviation Explanation SGCS Squared generalized cosine similarity

[0294] To A Time of arrival

[0295] LMF Location management function

[0296] KPI Key performance indicator

[0297] CSI Channel state information

[0298] RSRP Reference signal received power

[0299] NMSE Normalized mean square error

[0300] MDT Minimization Drive Test

[0301] CU Central Unit

[0302] DU Distributed Unit

[0303] NWDAF Network Data Analytics Function

Claims

Claims1. A method performed by a communication device (730) for handling communication in a wireless communications network (700), the method comprising: receiving (910) a first signaling message from a network node (711) for reporting a performance related information associated to a computational model for predicting a Channel State Information, CSI, in one or multiple future time instances; sending (920) a report on the performance related information according to the first signaling message.

2. The method according to claim 1, wherein the first signaling message comprises any one or a combination of the following: a. a CSI prediction performance monitoring report configuration for configuring the communication device (730) to report the performance monitoring results for an Artificial Intelligence, Al, or non-AI based CSI-prediction model used for generating a predicted CSI; b. a physical resource configuration for configuring the physical resources for the communication device (730) to transmit a performance monitoring report.

3. The method according to any one of claims 1-2, wherein the first signaling message is received via any one or a combination of the following: a) a Radio Resource Control, RRC, message; b) a Media Access Control, MAC, Control Element, MAC CE; c) a Downlink Control Information, DCI, signaling.

4. The method according to claim 3, wherein the RRC message is any one of the following: a. the same RRC message used for configuring the communication device (730) to report predicted CSI, b. a separate RRC message that is different from the one used for configuring the communication device (730) to report predicted CSI, c. the same RRC message used for configuring performance monitoring data collection at the communication device (730).

5. The method according to claim 4, wherein the RRC message indicates at least one of: a format of the performance related information, a content of the performance related information, for the Al or non-AI based CSI-prediction model.

6. The method according to claim 4, wherein the RRC message comprises a time window configuration during which the performance related information shall be reported.

7. The method according to claim 3, wherein the MAC CE is any one of the following: a. the same MAC CE used for requesting the communication device (730) to report the predicted CSI, b. a separate MAC CE defined for requesting the communication device (730) to report the performance related information.

8. The method according to claim 3, wherein the DCI signaling is any one of the following: a. the same DCI signaling used for requesting the communication device (UE 730) to report the predicted CSI, b. a separate DCI signaling defined for requesting the communication device (730) to report the performance related information.

9. The method according to any one of claims 1-8, wherein the performance related information comprises at least one of performance monitoring result(s) and performance metric(s) associated to an Al or non-AI based CSI prediction model.

10. The method according to claim 2-9, wherein the Al or non-AI based CSI prediction model is associated to any one of a feature identifier, a feature group, an AI / ML model identifier, and a functionality identifier.

11. The method according to any one of claims 1-10, wherein the performance related information is reported by using any one of the following: a. a Radio Resource Control, RRC, message; b. a Media Access Control, MAC, Control Element, MAC CE;c. a type of Uplink control information, UCI, on physical uplink control channel, PUCCH, or Physical Uplink Shared Channel, PUSCH, wherein the performance related information is defined as a new UCI type, or the performance related information is defined as part of the CSI report carrying the predicted CSI.

12. The method according to any one of claims 2-11, the performance monitoring results for an Al or non-AI based CSI-prediction model is sent along with the predicted CSI related parameters of any one of CSI reference signal Resource Indicator, CRI, Channel Quality Indicator, CQI, Precoding Matrix Indicator, PMI, Rank Indicator, RI, Layer Indicator, LI.

13. The method according to any one of claims 1-12, wherein the report on the performance related information includes a performance related information for an Al or non-AI based CSI-prediction model in a CSI report when the CSI report is sent periodically or semi-persistently over a physical uplink control channel, PUCCH, or a Physical Uplink Shared Channel, PUSCH.

14. The method according to claim 13, wherein the report on the performance related information includes the performance related information for an Al or non-AI based CSI-prediction model with or without the predicted CSI related reporting quantities in the CSI report when the report is sent aperiodically over a Physical Uplink Shared Channel, PUSCH.

15. The method according to any one of claims 2-14, wherein a 2 part CSI structure is used for reporting performance monitoring results.

16. The method according to claim 15, wherein CSI part 1 comprises information identifying the number of time instances, for which a performance monitoring metric(s) is reported.

17. The method according to claim 16, wherein CSI Part 2 comprises the performance monitoring metric(s) for the time instances that are identified by the CSI Part 1.

18. The method according to claim 15, wherein CSI part 1 comprises information identifying the rank or the number of layers, for which a performance monitoring metric(s) is reported.

19. The method according to claim 18, wherein CSI Part 2 comprises the performance monitoring metric(s) for the layers that are identified by CSI Part 1.

20. The method according to claim 15, wherein CSI part 1 comprises information identifying the type(s) of performance monitoring metric(s) that is (are) reported.

21. The method according to claim 20, wherein CSI Part 2 comprises the performance monitoring metric(s) for the identified types by CSI Part 1.

22. The method according to any one of claims 1-21, further comprising: receiving a second signaling message to stop reporting performance related information for an Al or non-AI based CSI-prediction model.

23. The method according to claim 22, wherein the second signaling message to stop reporting performance related information is implemented by any one of the following: a. a time window configured in a CSI prediction performance report configuration message. b. a different RRC message defined for stopping the communication device (730) reporting the performance related information. c. the same MAC CE used for deactivating a predicted CSI report, or a separate MAC CE defined for stopping the communication device (730) reporting the performance related information. d. the same DCI used for requesting the communication device (730) to report a predicted CSI, or a separate DCI defined for stopping the communication device (730) reporting the performance related information.

24. A method performed by a network node (711) for handling communication in a wireless communications network (700), the method comprising: sending (810) a first signaling message to a communication device (730), for configuring the communication device (730) to send a report on a performance related information associated to a computational model for predicting Channel State Information, CSI, in one or multiple future time instances; and receiving (820) a report sent from the communication device (730) based on the first signaling message.

25. The method according to claim 24, wherein the first signaling message comprises any one or a combination of the following: a) a CSI prediction performance monitoring report configuration for configuring the communication device (730) to report the performance monitoring results for an Artificial Intelligence, Al, or non- Al based CSI-prediction model used for generating a predicted CSI; b) a physical resource configuration for configuring physical resources for the communication device (730) to transmit a performance monitoring report.

26. The method according to any one of claims 24-25, wherein the first signaling message is sent by using any one or a combination of the following: a) a Radio Resource Control, RRC, message; b) a Media Access Control, MAC, Control Element, MAC CE; c) a Downlink Control Information, DCI, signaling.

27. The method according to any one of claims 25-26, wherein the performance related information comprises at least one of performance monitoring result(s), performance metric(s), associated to an Al or non- Al based CSI prediction model.

28. The method according to any one of claims 25-27, wherein the Al or non-AI based CSI prediction model is associated to any one of a feature identifier, a feature group, an AI / ML model identifier, a functionality identifier.

29. The method according to any one of claims 26-28, wherein the RRC message is any one of the following: a. the same RRC message used for configuring the communication device (730) to report predicted CSI, b. a separate RRC message that is different from the one used for configuring the communication device (730) to report predicted CSI, c. the same RRC message used for configuring performance monitoring data collection at the communication device (730).

30. The method according to claim 29, wherein the RRC message indicates at least one of: a format of the performance related information, a content of the performance related information, for the Al or non-AI based CSI-prediction model.

31. The method according to claim 29, wherein the RRC message comprises a time window configuration during which the performance related information shall be reported.

32. The method according to any one of claims 26-28, wherein the MAC CE is any one of the following: a. the same MAC CE used for requesting the communication device (730) to report the predicted CSI, b. a separate MAC CE defined for requesting the communication device (730) to report the performance related information.

33. The method according to any one of claims 26-28, wherein the DCI signaling is any one of the following: a. the same DCI signaling used for requesting the communication device (730) to report the predicted CSI, b. a separate DCI signaling defined for requesting the communication device (730) to report the performance related information.

34. The method according to any one of claims 25-33, wherein upon decoding the received report carrying the performance related information, the method further comprises at least one of the following actions: a. sending a second signaling message to the communication device (730) for configuring subsequent CSI reports with the existing Al or non- Al model for CSI prediction, if the performance of the Al or non-AI based prediction fulfils a performance criterion; b. sending a second signaling message to the communication device (730) for configuring subsequent CSI reports with a different Al or non-AI model for CSI prediction if the performance of the Al or non-AI based prediction does not fulfil a performance criterion; c. sending a second signaling message to the communication device (730) for configuring subsequent CSI reports with fall back to non-AI based CSI prediction if an Al-based CSI prediction is currently configured and the performance of the Al based CSI prediction does not fulfil a performance criterion; d. sending a second signaling message to the communication device (730) for configuring subsequent CSI reports with fall back to CSI measurement reporting without prediction.

35. The method according to any one of claims 25-34, further comprising: signaling to the communication device (730) to stop reporting performance related information for an Al or non-AI based CSI-prediction model.

36. The method according to claim 35, wherein signaling to the communication device (730) to stop reporting performance related information is implemented by any one of the following: a. using a time window configured in a CSI prediction performance report configuration message. b. using a different RRC message defined for stopping the communication device (730) reporting the performance related information.c. using the same MAC CE used for deactivating a predicted CSI report, or a separate MAC CE defined for stopping the communication device (730) reporting the performance related information. d. using the same DCI used for requesting the communication device (730) to report a predicted CSI, or a separate DCI defined for stopping the communication device (730) reporting the performance related information.

37. The method according to any one of claims 35-36, wherein the configuration for the communication device (730) to start or stop reporting is based on a channel condition.

38. The method according to claim 37, wherein the channel condition is any one of a speed of the communication device (730) being above a certain threshold, a delay spread estimate being outside a certain region, a doppler spread estimate being outside a certain region.

39. The method according to any one of claims 26-38, wherein the physical resource configuration for configuring physical resources for the communication device (730) to transmit a performance monitoring report is based on time domain behavior of the performance monitoring reports and quantities of the reports.

40. The method according to any one of claims 25-39, wherein the performance related information is configured to be reported by any one of the following: a. using an RRC message; b. using an MAC CE; c. using a type of Uplink control information, UCI, on physical uplink control channel, PUCCH, or Physical Uplink Shared Channel, PUSCH, wherein the performance related information is defined as a new UCI type, or the performance related information is defined as part of the CSI report carrying the predicted CSI.

41. The method according to claim 40, wherein the CSI prediction performance report configuration configures the communication device (730) to report back the performance related information for an Al or non-AI based CSI-prediction modelalong with predicted CSI related parameters of any one of Channel Rank Indicator, CRI, Channel Quality Indicator, CQI, Precoding Matrix Indicator, PMI, Rank Indicator, RI, Layer Indicator, LI.

42. A communication device (730) configured to perform the method according to any one of claims 1-23.

43. A network node (711) configured to perform the method according to any one of claims 24-41.

44. A computer program product (1070) comprises instructions, which, when executed on at least one processor (1060), cause the at least one processor to carry out the methods according to any one of claims 1-23 as performed by the communication device.

45. A computer-readable storage medium (1080), having stored a computer program product (1070) comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the methods according to any one of claims 1-23 as performed by the communication device.

46. A computer program product (1170) comprising instructions, which, when executed on at least one processor (1160), cause the at least one processor to carry out the methods according to any one of claims 24-41 as performed by the network node.

47. A computer-readable storage medium (1180), having stored a computer program product (1170) comprising instructions which, when executed on at least one processor, cause the at least one processor to carry out the methods according to any one of claims 24-41 as performed by the network node.

Images