Method for communication

Abstract

The embodiment of the invention relates to a communication method, equipment and a computer storage medium. The method includes receiving, at a terminal device, at least one first set of reference signals (RSs) from a network device deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each set of RSs in the at least one first set of RSs corresponding to one of the at least one AI / ML model; calculating, at the terminal device, at least one similarity based on the at least one first set of RSs; determining, at the terminal device, at least one piece of similarity information based on the calculated at least one similarity, and at least one piece of model information corresponding to the at least one piece of similarity information, the at least one piece of model information indicating an index of the at least one AI / ML model; and sending at least one of the following information to the network device: the determined at least one piece of similarity information or at least one piece of model information.

Description

Technical Field

[0001] Example embodiments of the present disclosure relate generally to the field of communication technology, and particularly to methods, devices, and computer-readable media for communication. Background Art

[0002] Artificial intelligence / machine learning (AI / ML) models are introduced for beam management (BM) in communication systems. For AI / ML model monitoring, the traditional method is to compare the predicted beam information with the actual beam information. This method requires the terminal device (also known as "user equipment", "user device" or UE) to measure a large number of beam measurement reference signals (RS) and report a large amount of beam information, which will lead to huge overhead for beam measurement and reporting. Summary of the invention

[0003] In general, example embodiments of the present disclosure provide a method, apparatus, and computer storage medium for communicating, and in particular, for reporting similarity information between trained / predicted / preconfigured beam information and field / actual beam information.

[0004] In a first aspect, a communication method is provided. The method includes: receiving at least one first group of RSs from a network device at a terminal device, the network device being deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each group of RSs in the at least one first group of RSs corresponding to one of the at least one AI / ML models; calculating at least one similarity based on the at least one first group of RSs at the terminal device; determining at least one similarity information based on the calculated at least one similarity and at least one model information corresponding to the at least one similarity information at the terminal device, the at least one model information indicating an index of the at least one AI / ML model; and sending at least one of the determined at least one similarity information or at least one model information to the network device.

[0005] In a second aspect, a communication method is provided. The method includes: sending at least one first group of RSs to a terminal device at a network device, the network device being deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each group of RSs in the at least one first group of RSs corresponding to one of the at least one AI / ML model; receiving from the terminal device at least one of the following: at least one similarity information determined, or at least one model information indicating an index of at least one AI / ML model; and determining whether the performance of at least one AI / ML model is degraded.

[0006] In a third aspect, a terminal device is provided. The terminal device includes a processor and a memory storing computer program code. The memory and the computer program code are configured to work with the processor to enable the terminal device to execute the method of the first aspect.

[0007] In a fourth aspect, a network device is provided. The network device includes a processor and a memory storing computer program code. The memory and the computer program code are configured to work with the processor to enable the network device to execute the method of the second aspect.

[0008] In a fifth aspect, a computer readable medium having instructions stored thereon is provided. The instructions, when executed by a processor of a device, cause the device to perform the method of the first aspect or the second aspect.

[0009] It should be understood that the invention summary is not intended to identify the key or essential features of the embodiments of the present disclosure, nor is it intended to be used to limit the scope of the present disclosure. Other features of the present disclosure will become readily understood through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The above and other objects, features and advantages of the present disclosure will become more apparent through a more detailed description of some example embodiments of the present disclosure in the accompanying drawings, in which:

[0011] Figure 1A illustrates an example communication system in which some embodiments of the present disclosure may be implemented;

[0012] Figure 1B A schematic diagram illustrating a set of beams according to some embodiments of the present disclosure is illustrated;

[0013] Figure 2 illustrates an example signaling diagram according to some example embodiments of the present disclosure;

[0014] Figure 3 illustrates another example signaling diagram according to some example embodiments of the present disclosure;

[0015] Figure 4 illustrates a third example signaling diagram according to some example embodiments of the present disclosure;

[0016] Figure 5 illustrates a fourth example signaling diagram according to some example embodiments of the present disclosure;

[0017] Figure 6 illustrates a fifth example signaling diagram according to some example embodiments of the present disclosure;

[0018] Figure 7 illustrates a sixth example signaling diagram according to some example embodiments of the present disclosure;

[0019] Figure 8 illustrates a seventh example signaling diagram according to some example embodiments of the present disclosure;

[0020] Fig. 9 A flowchart illustrating an example method implemented at a terminal device according to some embodiments of the present disclosure is illustrated;

[0021] Fig.10 A flowchart illustrating an example method implemented at a network device according to some embodiments of the present disclosure; and

[0022] Fig.11 A simplified block diagram of a device suitable for implementing embodiments of the present disclosure is illustrated.

[0023] Throughout the drawings, the same or similar reference numerals refer to the same or similar elements. DETAILED DESCRIPTION

[0024] The principles of the present disclosure will now be described with reference to some example embodiments. It should be understood that these embodiments are described only to illustrate and help those skilled in the art to understand and implement the present disclosure, without any limitation to the scope of the present disclosure. The embodiments described herein can be implemented in various other ways except for the way described below.

[0025] In the following description and claims, unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this disclosure belongs.

[0026] References in this disclosure to "one embodiment," "an embodiment," and "an example embodiment" indicate that the described embodiment may include a particular feature, structure, or characteristic, but not every embodiment necessarily includes the particular feature, structure, or characteristic. Furthermore, such phrases do not necessarily refer to the same embodiment. Furthermore, when a particular feature, structure, or characteristic is described in conjunction with an embodiment, it is considered to be within the knowledge of those skilled in the art to affect such feature, structure, or characteristic in conjunction with other embodiments (whether or not explicitly described).

[0027] It should be understood that although the terms "first" and "second" etc. may be used herein to describe various elements, these elements should not be limited by these terms. These terms are only used to distinguish one element from another element. For example, without departing from the scope of the example embodiments, a first element may be referred to as a second element, and similarly, a second element may be referred to as a first element. As used herein, the term "and / or" includes any and all combinations of one or more of the listed terms.

[0028] The terms used herein are only used to describe specific embodiments and are not intended to limit the exemplary embodiments. As used herein, the singular forms "a", "an", and "the" shall also include the plural forms unless the context clearly indicates otherwise. It will be further understood that the terms "comprise", "comprising", "has", "having", "includes", and / or "including" when used herein specify the presence of the features, elements, and / or components, etc., but do not exclude the presence or addition of one or more other features, elements, components, and / or combinations thereof.

[0029] In some examples, values, processes, or devices are referred to as "best," "lowest," "highest," "minimum," "maximum," etc. It should be understood that such descriptions are intended to indicate that one can choose among many functional alternatives that can be used, and that such a choice is not necessarily better, lesser, higher, or otherwise preferred over other choices.

[0030] As used herein, the term "communication network" refers to a network that complies with any suitable communication standard (such as new radio (NR), long term evolution (LTE), advanced LTE (LTE-A), wideband code division multiple access (WCDMA), high speed packet access (HSPA), narrowband Internet of Things (NB-IoT), etc.). In addition, the communication between the terminal device and the network device in the communication network can be performed according to any suitable generation of communication protocol, including but not limited to the first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G), 5.5G, 5G advanced network, or sixth generation (6G) communication protocol, and / or any other protocol currently known or to be developed in the future. The embodiments of the present disclosure can be applied to various communication systems. In view of the rapid development of communication, there will certainly be communication technologies and systems of future types that can embody the present disclosure. It should not be regarded as limiting the scope of the present disclosure to the above-mentioned systems.

[0031] As used herein, the term "terminal device" refers to any device with wireless or wired communication capabilities. Examples of terminal devices include, but are not limited to, user equipment (UE), personal computers, desktops, mobile phones, cellular phones, smart phones, personal digital assistants (PDAs), portable computers, tablets, wearable devices, Internet of Things (IoT) devices, ultra-reliable low-latency communications (URLLC) devices, Internet of Everything (IoE) devices, machine type communications (MTC) devices, vehicle-mounted devices for V2X communications (where X represents pedestrians, vehicles, or infrastructure / network), devices for integrated access and backhaul (IAB), space vehicles or air vehicles (including unmanned aerial systems (UAS)) in non-terrestrial networks (NTNs) (including satellites and high altitude platforms (HAPs)), extended reality (XR) devices (including different types of reality such as augmented reality (AR), mixed reality (MR), and virtual reality (VR)), unmanned aerial vehicles (UAVs) commonly referred to as drones (which are aircraft without any human pilots), devices on high-speed trains (HSTs), or image capture devices (such as digital cameras), sensors, gaming devices, music storage and playback devices, or Internet devices that enable wireless or wired Internet access and browsing, etc. The "terminal device" may also have "multicast / broadcast" features to support public safety and mission-critical, V2X applications, transparent IPv4 / IPv6 multicast delivery, IPTV, Smart TV, radio services, wireless software delivery, group communication, and IoT applications. It may also include one or more subscriber identity modules (SIMs) (referred to as multi-SIM). The term "terminal device" may be used interchangeably with UE, mobile station, subscriber station, mobile terminal, user terminal, or wireless device.

[0032] As used herein, the term "network equipment" refers to equipment that is capable of providing or hosting a cell or coverage area in which a terminal device can communicate. Examples of network equipment include, but are not limited to, satellites, unmanned aerial systems (UAS) platforms, Node Bs (NodeBs or NBs), evolved Node Bs (eNodeBs or eNBs), next generation Node Bs (gNBs), transmission reception points (TRPs), remote radio units (RRUs), radio heads (RHs), remote radio heads (RRHs), IAB nodes, low power nodes (such as femto nodes, pico nodes), reconfigurable smart surfaces (RISs), etc.

[0033] The communications discussed herein may conform to any suitable standard, including but not limited to new radio access (NR), long term evolution (LTE), LTE evolution, advanced LTE (LTE-A), wideband code division multiple access (WCDMA), code division multiple access (CDMA), cdma2000, and global system for mobile communications (GSM), etc. In addition, communications may be performed according to any generation of communication protocols currently known or to be developed in the future. Examples of communication protocols include but are not limited to first generation (1G), second generation (2G), 2.5G, 2.85G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G), and sixth generation (6G) communication protocols. The techniques described herein may be used for the above-mentioned wireless networks and radio technologies as well as other wireless networks and radio technologies. Embodiments of the present disclosure may be performed according to any generation of communication protocols currently known or to be developed in the future. Examples of communication protocols include, but are not limited to, first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G) communication protocols, 5.5G, 5G advanced networks, or sixth generation (6G) networks.

[0034] The terminal device or network device may have artificial intelligence (AI) or machine learning capabilities. It usually includes a model that is trained from a large amount of collected data for a specific function and can be used to predict certain information.

[0035] The terminal device or network device can operate in several frequency ranges, such as FR1 (410MHz-7125MHz), FR2 (24.25GHz to 71GHz), frequency bands greater than 100GHz, and terahertz (THz). It can further operate on licensed / unlicensed / shared spectrum. In the multi-radio dual connection (MR-DC) application scenario, the terminal device can have more than one connection with the network device. The terminal device or network device can operate in full-duplex, flexible duplex, and cross-duplex modes.

[0036] The embodiments of the present disclosure may be executed in a test device, such as a signal generator, a signal analyzer, a spectrum analyzer, a network analyzer, a test terminal device, a test network device, or a channel emulator.

[0037] Embodiments of the present disclosure may be performed according to any generation of communication protocols currently known or to be developed in the future. Examples of communication protocols include, but are not limited to, first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G) communication protocols, 5.5G, 5G advanced networks, or sixth generation (6G) networks.

[0038] The term "circuitry" as used herein may refer to a combination of hardware circuits and / or hardware circuits and software. For example, a circuit system may be a combination of analog and / or digital hardware circuits and software / firmware. As another example, a circuit system may be any part of a hardware processor (including (multiple) digital signal processors), software, and (multiple) memories with software, which work together to enable devices such as terminal devices or network devices to perform various functions. In another example, a circuit system may be a hardware circuit and / or processor, such as a microprocessor or a part of a microprocessor, which requires software / firmware to operate, but when the operation is not required, the software may not exist. As used herein, the term circuit system also encompasses the implementation of only a hardware circuit or (multiple) processors or a part of a hardware circuit or (multiple) processors and their accompanying software and / or firmware.

[0039] As used herein, the singular forms "a", "an", and "the" shall also include the plural forms unless the context clearly indicates otherwise. The term "including" and its variations shall be understood as open-ended terms, meaning "including but not limited to". The term "based on" shall be understood as "based at least in part on". The terms "one embodiment" and "embodiment" shall be understood as "at least one embodiment". The term "another embodiment" shall be understood as "at least one other embodiment". The terms "first", "second", etc. may refer to different or the same objects. Other explicit and implicit definitions may be included below.

[0040] In some examples, values, processes, or devices are referred to as "best," "lowest," "highest," "minimum," "maximum," etc. It should be understood that such descriptions are intended to indicate that one can choose among many functional alternatives that can be used, and that such a choice is not necessarily better, lesser, higher, or otherwise preferred over other choices.

[0041] In some communication systems, (multiple) artificial intelligence (AI) / machine learning (ML) models are used for beam management. If the AI / ML model is deployed at the network equipment for model monitoring (or verification, testing), the following operations are performed:

[0042] S1: The network device sends set A (and set B) to the terminal device. Figure 1B Describe set A and set B.

[0043] S2: The terminal device reports beam information to the network device. Specifically, the beam information includes at least one of the following: beam information of set A, beam information of set B and set A, or beam information of set B and beam information of the first N beams of set A.

[0044] S3: Based on the reported beam information of set B, the network device uses the AI / ML model to obtain the predicted beam information. And the predicted beam information needs to be compared with the reported actual / measured beam information, such as the estimated RSRP, the top N beams.

[0045] S4: The network device determines the AI / ML model performance based on the comparison results (i.e., the difference between the predicted beam information and the actual / measured beam information).

[0046] S5: If the AI / ML model performance degrades (e.g., the difference is greater than a threshold), the network device may consider model switching / updating (e.g., fine-tuning, retraining), or may consider non-AI.

[0047] As can be seen from the above, for AI / ML model monitoring, the traditional method is to compare the predicted beam information with the actual beam information. In addition, the above method requires the terminal device to measure a large number of beam measurement RSs and report a large amount of beam information, which will result in huge overhead for beam measurement and reporting.

[0048] Generally speaking, in the above-mentioned AI / ML model monitoring scenario, the AI / ML model learns "(historical) experience". "Experience" consists of tens of thousands of training data. The training data covers all possible realities as much as possible. However, due to the variability and randomness of the real environment, the AI / ML model still cannot handle some "unexpected" situations. Specifically, the data (i.e., the field data corresponding to the input of the AI / ML model) may be "strange" to the AI / ML model. In other words, the AI / ML model may not have learned the "experience" corresponding to the field data, or the field data may not be similar to the (original) training data. In this case, the performance (e.g., generalization) of the AI / ML model will decrease (or even deteriorate), which will reduce the accuracy of the AI / ML model reasoning. Therefore, the similarity between the field data and the training data can be used to indirectly determine the performance (e.g., generalization) of the AI / ML model. In other words, the similarity between the field data and the training data can be used as a metric that indirectly reflects the performance (e.g., generalization) of the AI / ML model.

[0049] For model monitoring based on beam prediction, instead of the traditional method (i.e., comparing predicted beam information and actual beam information), we can determine the similarity between the field set B and the training set B to determine the performance of the AI / ML model. Unlike AI / ML models, the training data does not involve the relevant algorithms and therefore has no privacy. Therefore, the relevant training data at the network device can be shared with the terminal device.

[0050] Then, the terminal device can obtain the training data set (ie, training set B) through network equipment, core network or edge cloud (or server, device) configuration or provision.

[0051] Considering that the beam information is mainly RSRP (i.e., real number), the algorithm for calculating the similarity between the live set B and the training set B is simple and easy to implement (the specific algorithm depends on the implementation of the terminal device). For beam prediction based on AI / ML models, the terminal device can determine the similarity between the live set B and the training set B.

[0052] In summary, at least for AI / ML model monitoring, in order to reduce the overhead of beam measurement and reporting, it can be considered that the terminal device only reports the similarity information between the live set B and the training set B to the network device. In this case, the terminal device can only measure the beam corresponding to set B (instead of set A) (which can mean that the network device only needs to send set B). And, the terminal device can only report the similarity information without reporting the beam information of the beam corresponding to set B and set A or part of set A.

[0053] Figure 1A An example communication system 100 is shown in which some embodiments of the present disclosure may be implemented. The communication system 100, as part of a communication network, includes a network device 110 and a terminal device 120.

[0054] Network device 110 may provide services to terminal device 120, and network device 110 and terminal device 120 may communicate data and control information to each other. In some embodiments, network device 110 and terminal device 120 may communicate via a direct link / channel.

[0055] In system 100, the link from network device 110 to terminal device 120 is called downlink (DL), and the link from terminal device 120 to network device 110 is called uplink (UL). In the downlink, network device 110 is a sending (TX) device (or transmitter), and terminal device 120 is a receiving (RX) device (or receiver). In the uplink, terminal device 120 is a sending (TX) device (or transmitter), and network device 110 is an RX device (or receiver). It should be understood that network device 110 can provide one or more service cells. As shown in Figure 1, network device 110 provides a service cell 102, and terminal device 120 resides on service cell 102. In some embodiments, network device 110 can provide multiple service cells. It should be understood that the number of (multiple) service cells shown in Figure 1 is for illustrative purposes only and does not represent any limitation.

[0056] The communications in the communication system 100 may conform to any suitable standard, including but not limited to Long Term Evolution (LTE), LTE Evolution, Advanced LTE (LTE-A), Wideband Code Division Multiple Access (WCDMA), Code Division Multiple Access (CDMA), and Global System for Mobile Communications (GSM), etc. In addition, the communications may be performed according to any generation of communication protocols currently known or to be developed in the future. Examples of communication protocols include but are not limited to first generation (1G), second generation (2G), 2.5G, 2.75G, third generation (3G), fourth generation (4G), 4.5G, fifth generation (5G), 5.5G, 5G advanced network or sixth generation (6G) communication protocols.

[0057] It should be understood that the number of devices and their connection relationships and types shown in Figure 1 are for illustration purposes only and do not represent any limitation. Communication system 100 may include any suitable number of devices suitable for implementing embodiments of the present disclosure.

[0058] It should be noted that:

[0059] The term "beam of a target signal" in the present disclosure refers to, for example, a QCL-Type D (source) RS of a target signal.

[0060] The term "beam information" in the present disclosure refers to, for example, a beam identifier (ID) or (and) beam quality.

[0061] The term "beam ID" in the present disclosure refers to, for example, a CSI-RS resource indicator (CRI) or a SS / PBCH block resource indicator (SSBRI).

[0062] The term "beam quality" in the present disclosure refers to, for example, layer 1 reference signal received power (L1-RSRP), layer 1 signal to interference and noise ratio (L1-SINR), RSRP or SINR. Here, L1-RSRP may be equivalent to RSRP, and L1-SINR may be equivalent to SINR.

[0063] The term "QCL-Type D" refers to, for example, spatial Rx parameters.

[0064] The term "similarity" between A and B in the present disclosure refers to, for example, a measure reflecting the distance or correlation or similarity (e.g., Euclidean distance, Minkowski distance, cosine similarity, Pearson correlation) between A and B. Typically, for example, the similarity can be a number between 0 and 1, such as 0, 0.1, 0.2, 0.3, ..., 0.8, 0.9, and 1.

[0065] It should be noted that in the present disclosure, "similarity" can be determined based on "dissimilarity" (or "diversity", "difference"). Generally, the relationship between "similarity" and "dissimilarity" satisfies: similarity = 1-dissimilarity.

[0066] It should be noted that in the present disclosure, "similarity" or "similarity information" can be replaced by "dissimilarity" or "dissimilarity information". Therefore, the reporting and interpretation of "dissimilarity information" by terminal device 120 and network device 110 should be opposite. For example, in some example embodiments, the indication can be used to indicate whether set B is dissimilar to training set B.

[0067] Figure 1B A schematic diagram of a set of beams according to some embodiments of the present disclosure is illustrated. For discussion purposes only, process 200 will be described with reference to FIG1 . Process 200 may involve terminal device 120 and network device 110 .

[0068] like Figure 1B As shown, the AI / ML model and set A are deployed at the network device 110. "Set A" refers to a set of RSs (hereinafter also referred to as "a set of beams") deployed in the network device 110, and it includes a total of 16 beams. "Set B" refers to another set of beams that will be used for field measurements at the terminal device 120 for AI / ML model monitoring. Figure 1B In the example shown, set B includes 4 beams in set A and is therefore a subset of set A. In a beam management scenario, the network device 110 may send set B to the terminal device 120 to obtain field / actual measurements. Such field measurements are used to select the top N beams from set A to improve communication quality and system performance.

[0069] It should be noted that although Figure 1B It is shown in FIG. 1 that set B of field measurements for AI / ML model monitoring is a subset of set A, but in other examples, set B may be a set of RSs not included in set A.

[0070] Figure 2 An example signaling diagram of a communication process 200 according to some example embodiments of the present disclosure is illustrated. For discussion purposes only, the process 200 will be described with reference to FIG. 1 . The process 200 may involve a terminal device 120 and a network device 110 .

[0071] In some example embodiments, network device 110 is deployed with at least one AI / ML model. Terminal device 120 receives at least one first group of RSs from network device 110, and each group of RSs in at least one first group of RSs corresponds to one AI / ML model in at least one AI / ML model. Then, terminal device 120 calculates at least one similarity based on at least one first group of RSs. Then, terminal device 120 determines at least one similarity information based on the calculated at least one similarity, and sends at least one of the following to network device 110: the determined at least one similarity information, or at least one model information indicating an index of at least one AI / ML model.

[0072] exist Figure 2 In the example shown, the network device 110 sends 220 a set of site beams (denoted as "site set B") 201 to the terminal device 120. Here, the site set B 201 is as shown in FIG. Figure 1B A set of RSs of “Set B” is explained.

[0073] On the other side of the communication, the terminal device 120 receives 222 at least one first set of RSs (in Figure 2 Then, the terminal device 120 calculates 224 at least one similarity based on at least one first group of RSs.

[0074] For example, the terminal device 120 calculates the similarity based on the field set B 201. In this case, the network device 110 may also send a signal that is consistent with the first group of RSs (in Figure 2In the present invention, the terminal device 120 receives training information of a second group of RSs (hereinafter also referred to as "training set B") associated with the field set B 201). The second group of RSs associated with the first group of RSs may have the same or similar beam parameters (e.g., beam, beam direction, etc.) as the first group of RSs, but measurements are performed in a training environment to obtain training information (also referred to as "training data set"). On the other side of the communication, the terminal device 120 receives training information from the network device 110. The terminal device 120 measures the field set B to obtain a field measurement, which includes at least one of the following: layer 1 reference signal received power (L1-RSRP), layer 1 signal to interference and noise ratio (L1-SINR), RSRP, SINR, RSRQ, or CINR. For example, the field measurement of the field set B may include L1-RSRP. Then, the terminal device 120 calculates the Euclidean distance of the quality metric between the field measurement of at least one first group of RSs and the training information of at least one second group of RSs to obtain multiple similarities, and takes the maximum similarity among the multiple similarities as the determined similarity. Here, the quality metric may be one of the following: L1-RSRP, L1-SINR, RSRP, SINR, reference signal received quality (RSRQ), or carrier-to-interference-plus-noise ratio (CINR). For example, the terminal device 120 calculates the field measurement of the first set of RSs (in Figure 2 In the example, the Euclidean distance of the L1-RSRP between the site set B 201 and the training information of the second group of RSs (training set B) is calculated to obtain the similarity between the two. For example, it is assumed that the training information includes the L1-RSRP of 100,000 samples of the training set B associated with the site set B 201. Each sample is used to calculate the Euclidean distance, as shown in Equation 1.

[0075]

[0076] In Equation 1, xi represents the value of the field measurement of the L1-RSRP of the site set B 201, and yi represents the value of the L1-RSRP in the training information of the training set B associated with the site set B 201. Therefore, the Euclidean distance d(x, y) of the L1-RSRP between the field measurement of the site set B 201 and the training information of the training set B associated with the site set B 201 is calculated. Then, the similarity sim(x, y) between the field measurement of the site set B 201 and the training information of the training set B associated with the site set B 201 is obtained, which is the inverse of the sum of d(x, y) and 1.

[0077] Since the training set B associated with the field set B 201 has 100,000 samples, 100,000 Euclidean distances d(x, y) and 100,000 sim(x, y) may be obtained. The terminal device 120 may then use the maximum similarity among the 100,000 similarities as the calculated similarity.

[0078] If there are multiple field sets B, for each group of RSs in at least one first group of RSs, at least one similarity based on at least one first group of RSs can be calculated by calculating the similarity between the measured beam quality of each group of RSs in the first group of RSs and the beam quality corresponding to each group of RSs in the training data set of the corresponding AI / ML model based on each group of RSs in the first group of RSs. In other words, for each group of RSs in at least one first group of RSs, a method for calculating similarity can be applied to obtain corresponding at least one (multiple) similarities.

[0079] It should be noted that the similarity between the field set B and the training set B is used as an example only. It may generally refer to the similarity between the first group of RSs and the second group of RSs associated with the first group of RSs, or the similarity between the measurements corresponding to the first group of RSs (e.g., RSRP / L1-RSRP / RSRQ / CIR / CINR) and the measurements corresponding to the second group of RSs, or the similarity between the measured values ​​of the first group of reference signals and the corresponding / related values ​​in the training data set. Alternatively, it may refer to the similarity between the first group of RSs and a set of predefined / configured values ​​associated with the first group of reference signals. The terminal device 120 may obtain training information (training data set) of at least one second group of reference signals associated with at least one first group of reference signals from the configuration or supply of the network device 110, the core network or the edge cloud (or a server, other devices via a side chain).

[0080] Then, based on the calculated similarities, the terminal device 120 determines 230 at least one similarity information based on the calculated at least one similarity and at least one model information corresponding to the at least one similarity information. The at least one model information indicates an index of at least one AI / ML model.

[0081] In one example, the similarity information is a first indication indicating a first state and a second state. If the calculated similarity is greater than or equal to a first threshold, the first indication indicates the first state, and if the calculated similarity is less than the first threshold, the first indication indicates the second state. In other words, the first state indicates that the field measurement of at least one first group of reference signals is similar to the training information of at least one second group of reference signals associated with the at least one first group of reference signals 201, and the second state indicates that the field measurement of at least one first group of reference signals is not similar to the training information of at least one second group of reference signals associated with the at least one first group of reference signals 201. Since only two different results need to be indicated, the bit width (or payload size) of the first indication can be 1 bit. For example, "1" can be used to indicate at least one first group of reference signals (e.g., Figure 2 The field set B 201 shown is similar to at least one second set of reference signals (e.g., training set B) associated with at least one first set of reference signals, and "0" may be used to indicate at least one first set of reference signals (e.g., Figure 2 The field set B 201 shown in FIG. 1 is not similar to at least one second set of reference signals (e.g., training set B) associated with at least one first set of reference signals 201. Based on the determined similarity sim(x, y) being greater than or equal to a predefined first threshold, the terminal device 120 may determine that at least one first set of reference signals 201 (e.g., Figure 2 Based on the determined similarity sim(x,y) being less than a predefined first threshold, the terminal device 120 may determine that at least one first set of reference signals 201 (e.g., Figure 2 The field measurements of the field set B) shown are not similar to the training information of at least one second set of reference signals. The predefined first threshold can be specified by the network device 110 and can be fixed and unchanged. Alternatively, the predefined first threshold can be AI / ML model specific and can be configured by the network device 110 through RRC / MAC-CE / DCI signaling.

[0082] In doing so, the network device 110 can indirectly determine the AI / ML model performance. In other words, the network device 110 can determine whether the AI / ML model is applicable or suitable for the current communication environment.

[0083] In some example embodiments, the similarity information is a second indication indicating a similarity level corresponding to the calculated similarity, and the payload size of the similarity information is determined based on the number of similarity levels. In one example, the bit width (payload size) of the second indication may be M (M ≥ 1) bits, where the value of M depends on the number of similarity levels, e.g. here, Represents an upper limit function, which is used to round the calculated log2 (number of levels) value. For example, assume that the number of levels is 4. In this case, the bit width (payload size) of the second indication is 2 (=log24) bits. And, it can correspond to 4 similarity levels: "Level 0: Dissimilar", "Level 1: Low Similarity", "Level 2: High Similarity" and "Level 3: Completely Similar". The terminal device 120 can determine the level of the determined similarity by comparing the determined similarity with a predefined threshold. For example, for 4 levels, 3 predefined thresholds can be specified, such as T1, T2 and T3.

[0084] T1, T2, and T3 may be specified, fixed, or constant. Alternatively, they may be AI / ML model specific and may be configured by network device 110 via RRC / MAC-CE / DCI signaling. The intervals between levels (i.e., the range of similarity) may be the same, and may be different, and may be specified / fixed / constant or configured.

[0085] Table 1

[0086] Second indicator Similarity level Similarity range 00 Not similar 0≤similarity≤T1 01 Low similarity T1<similarity≤T2 10 High similarity T2<similarity≤T3 11 Completely similar T3 < Similarity ≤ 1

[0087] When the second indication indicates that it is not "completely similar", the network device 110 may send additional training information to the terminal device 120 to update at least one AI / ML model. On the other side of the communication, the terminal device 120 may receive additional training information to update at least one AI / ML model. Specifically, if the terminal device 120 indicates "dissimilar" to the network device 110, in other words, if the network device 110 is indicated as the second indication "dissimilar", the network device 110 may discard the currently used AI / ML model. Alternatively or additionally, the network device 110 may perform model retraining / switching. Alternatively or additionally, the network device 110 may decide not to use any AI / ML model for beam management. If the terminal device indicates "low similarity" or "high similarity" to the network device 110, in other words, if the network device 110 is indicated as "low similarity" or "high similarity" of the second indicator, the network device 110 may further collect new training data to perform model updates (e.g., fine-tuning). In particular, for "low similarity", more new training data may be required because a large number of parameters of the AI / ML model need to be updated. For "high similarity", less new training data may be required because only a few parameters of the AI / ML model need to be updated. If the terminal device 120 indicates "complete similarity" to the network device 110, in other words, if the network device 110 is indicated by the terminal device 120 as "complete similarity" of the second indicator, the network device 110 can continue to use the currently used AI / ML model, and the AI / ML model does not need to be changed.

[0088] In some example embodiments, the similarity information is a third indication of a digital value indicating the calculated similarity. The digital value of the calculated similarity is a real number between 0 and 1, with a scale of 0.1. The payload size of the similarity information is determined based on the number of scales of the similarity. For example, the third indication may indicate a similarity (e.g., 0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1) between at least one first group of reference signals 201 and at least one second group of reference signals associated with at least one first group of reference signals 201. For example, the third indication can be the value of the calculated similarity sim(x, y) itself, ranging from 0 to 1, with a scale of 0.1, which means that although the actual calculated similarity sim(x, y) may not be exactly one of {0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1}, the actual calculated similarity sim(x, y) can be rounded to {0, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1}, with a scale of 0.1. In this case, the similarity includes 11 different values, so the bit width (payload size) of the third indication can be determined based on log211, i.e., 4 bits. For example, here, "0000" may mean "similarity equals 0", "0001" may mean "similarity equals 0.1", "0010" may mean "similarity equals 0.2", ... "1000" may mean "similarity equals 0.8", "1001" may mean "similarity equals 0.9", and "1010" may mean "similarity equals 1". In addition, "1011", "1100", "1101", "1110" and "1111" may be reserved.

[0089] In doing so, the network device 110 can obtain a more accurate similarity. Therefore, the network device 110 can more accurately determine the size of the new training data set required for model update (eg, fine-tuning).

[0090] After determining 230 at least one similarity information and corresponding at least one model information, the terminal device 120 may send 240 at least one of the determined at least one similarity information or at least one model information 202 to the network device 110, and the at least one model information indicates an index of at least one AI / ML model. On the other side of the communication, the network device 110 may receive 242 at least one of the similarity information or the model information 202 from the terminal device 120. For example, the terminal device 120 may send 240 at least one of the similarity information or the model information 202 to the network device 110 in a channel state information (CSI) report.

[0091] The network device 110 may determine 570 performance degradation of at least one AI / ML model through at least one of the similarity information or model information received from the terminal device 120. Then, as described above, the network device 110 may discard the AI / ML model currently in use. Alternatively or additionally, the network device 110 may perform model retraining / switching. Alternatively or additionally, the network device 110 may decide not to use any AI / ML model for beam management. If the terminal device 120 indicates "completely similar" to the network device 110, the network device 110 may continue to use the AI / ML model currently in use, and the AI / ML model does not need to be changed.

[0092] In some example embodiments, the bit width (payload size) of the model information is determined based on the number of at least one AI / ML model. For example, assuming that there are L (L is a positive integer) AI / ML models, that is, the number of at least one AI / ML model is L, the bit width (payload size) of the model information can be determined based on L. For a specific example, the bit width (payload size) of the model information is determined as "┌log2L┐". For example, assuming L=6, the bit width (payload size) of the model information is determined to be 3.

[0093] Reference now Figure 3 , which illustrates a signaling diagram of a communication process 300 according to some example embodiments of the present disclosure. For discussion purposes only, reference will be made to FIG. 1 and FIG. Figure 2 The process 300 is described. The process 300 may involve the terminal device 120 and the network device 110. For the same or similar operation(s) as in the process 200, reference may be made to the description of the process 200, and thus the details will be omitted.

[0094] In some example embodiments, the terminal device 120 sends at least one of the similarity information or the model information 202 to the network device 110 in the following manner: receiving configuration information from the network device 110, the configuration information being used for the terminal device 120 to report at least one of the following in a first channel state information (CSI) report: at least one determined similarity information or at least one model information; generating a CSI report, the CSI report including at least one determined similarity information or at least one model information; and sending the CSI report to the network device 110.

[0095] For example, Figure 3 As shown, the network device 110 sends 220 at least one first group of RSs (such as Figure 2 Before the scene set B 201 shown, the network device 110 can send configuration information 301 to the terminal device 120. On the other side of the communication, the terminal device 120 receives 312 the configuration information 301.

[0096] In one example, the configuration information 301 may include a high-level parameter for enabling the terminal device 120 to perform AI / ML model monitoring. In another example, the configuration information 302 may be sent by the network device 110 and received by the terminal device 120 via at least one of the following: remote resource control (RRC) signaling, media access control control element (MAC-CE) signaling, or downlink control information (DCI) signaling. In this case, the configuration information 301 may configure the terminal device 120 to report an event indicating performance degradation of at least one AI / ML model.

[0097] In another example, the configuration information 301 may further include a parameter indicating that the CSI report is used to report at least one similarity information or at least one model information, and the configuration information 301 may configure the terminal device 120 to report a CSI report including at least one item 202 of the similarity information or the model information. Specifically, the configuration information 301 may configure the terminal device 120 to report a CSI report including a new report item. The new report item may be named, for example, "similarity", and at least one item 202 of the similarity information or the model information may be included in the new report item in the CSI report to be reported to the network device 110.

[0098] Through the configuration information 301, the terminal device can be configured to perform AI / ML model monitoring. In other words, the terminal device can be configured to have the ability to perform AI / ML model monitoring using the configuration information 301.

[0099] After being configured to perform AI / ML model monitoring and before performing AI / ML model monitoring, the terminal device 120 may report capability information to the network device 110, and the capability information is used to indicate at least one of the following: the terminal device 120 supports AI / ML model monitoring, the terminal device 120 supports measuring / calculating / reporting similarity, or the training data set corresponding to the AI / ML model is deployed at the terminal device 120. On the other side of the communication, the network device 110 may receive capability information from the terminal device 120.

[0100] In some example embodiments, after the terminal device 120 determines 230 the similarity, the terminal device 120 may generate a CSI report including at least one 202 of the similarity information or the model information based on the configuration information 301, and send 240 the CSI report carrying the determined similarity information to the network device 110 in the allocated PUCCH / PUSCH resources, as already referred to Figure 2 As described above, on the other side of the communication, the network device 110 may receive 242 a CSI report.

[0101] In some example embodiments, after network device 110 receives 242 at least one item of similarity information or model information 202 from terminal device 120, network device 110 may determine 250 the performance of the AI / ML model based on at least one item of similarity information or model information 202 included in the received CSI report.

[0102] Reference now Figure 4 , which illustrates a third example signaling diagram of a communication process 400 according to some example embodiments of the present disclosure. For discussion purposes only, reference will be made to FIG. Figure 3 The process 400 is described. The process 400 may involve the terminal device 120 and the network device 110. For the same or similar operation(s) as in the process 300, reference may be made to the description of the process 300, and thus the details will be omitted.

[0103] In some example embodiments, in order to predict the beam information of set A, multiple AI / ML models may be deployed at network device 110. Specifically, these AI / ML models may correspond to different beam patterns or groups, for example, set B1, set B2, ... For simplicity, these sets B1, set B2, etc. are referred to as "multiple sets B". In order to perform model monitoring on multiple AI / ML models, network device 110 may send 410 a complete set A (or a union of multiple sets B) 401 to terminal device 120.

[0104] Then, the terminal device 120 can use the reference Figure 2 and Figure 3 The same method as described above is used to determine the similarity corresponding to each AI / ML model, that is, the similarity between the field set B corresponding to the AI / ML model and the training set B corresponding to the AI / ML model. Here, "field set B corresponding to the AI / ML model" means that the measured beam quality (e.g., L1-RSRP) of the first group of RSs is used as the input of the AI / ML model. The terminal device 120 can report the similarities corresponding to multiple AI / ML models at the same time.

[0105] In some example embodiments, network device 110 may send an AI / ML model list to terminal device 120. The AI / ML model list may include at least one AI / ML model ID (hereinafter also referred to as an "AI / ML model index" or simply "index"). At least one AI / ML model ID may correspond to at least one first group RS. On the other side of the communication, terminal device 120 may receive the AI / ML model list from network device 110.

[0106] For example, the terminal device 120 may be configured with an AI / ML model (or beam pattern) list including a plurality of AI / ML model IDs, and each AI / ML model ID corresponds to a specific beam pattern (ie, set B) sent from the network device 110.

[0107] Table 2: List of AI / ML models

[0108]

[0109] As shown in Table 2, there are (N+1) AI / ML models, each corresponding to a unique model ID. Each AI / ML model is associated with a beam set (i.e., RS set), as described above with reference to Figure 2 and Figure 3 The field set B. For example, the AI / ML model with AI / ML model ID=0 corresponds to the beam set {beam 0, beam 3, beam 12, beam 15}, the AI / ML model with AI / ML model ID=1 corresponds to the beam set {beam 5, beam 6, beam 9, beam 10}, the AI / ML model with AI / ML model ID=2 corresponds to the beam set {beam 0, beam 15}, ..., the AI / ML model with AI / ML model ID=N corresponds to the beam set {beam 3, beam 12}. Such beam sets can be regarded as examples of set B1, set B2, ..., which are generally referred to as "multiple sets B" in the present disclosure.

[0110] In this case, the terminal device 120 may report multiple indications to the network device 110. Each indication (i.e., similarity information) in the multiple indications may correspond to a specific AI / ML model in the multiple AI / ML models. Each indication in the multiple indications may be used to indicate any one of the following: whether the field measurement of at least one first group of RSs is similar to the training information of at least one second group of RSs, the level or state of the determined similarity between the field measurement of at least one first group of RSs and the training information of at least one second group of RSs, and the similarity value between the field measurement of at least one first group of RSs and the training information of at least one second group of RSs, the similarity value being a real value between 0 and 1. In other words, as described above, each indication in the multiple indications may be a first indication, a second indication, or a third indication.

[0111] In some example embodiments, the CSI report sent from the terminal device 120 to the network device 110 does not include any of the model IDs of the multiple AI / ML models, and the mapping order of the reported at least one determined similarity information in the CSI report is determined based on the index of the at least one AI / ML model corresponding to the at least one determined similarity information, for example, in ascending or descending order of the AI / ML model ID, as shown in Table 3.

[0112] Table 3: Similarity information included in the CSI report

[0113]

[0114] As shown in Table 3, in the CSI report, the CSI report number (i.e., the number of the indication) is indicated in the "CSI report number" field. In this example, as shown in the "CSI field", there is one indication for each AI / ML model, and the AI / ML model ID ranges from 0 to N. In other words, there are (N+1) indications for (N+1) AI / ML models. In this case, the CSI report number can be determined as (N+1). Therefore, if the indications of the (N+1) AI / ML models are sorted in ascending order based on the corresponding AI / ML model ID as shown in Table 3, then after receiving the CSI report including the information described in Table 3, the network device 110 can know that there are (N+1) indications for the AI / ML model, and can understand that each indication corresponds to the AI / ML model in ascending order of the AI / ML model ID. Since the CSI report includes (N+1) indications, the network device 110 can understand that the indication with index 0 corresponds to the AI / ML model with AI / ML model ID=0, the indication with index 1 corresponds to the AI / ML model with AI / ML model ID=1, and so on.

[0115] In some example embodiments, the CSI report may include a first part and a second part. The first part may include a fourth indication and a fifth indication, the fourth indication indicating the first state or the second state, and the fifth indication indicating the number of such AI / ML models: the corresponding similarity information of the AI / ML model indicates the first state or the second state. And, when the fifth indication is non-zero, the second part may include at least one of the following: model information, the model information indicating the index of such AI / ML model: the corresponding similarity information of the AI / ML model indicates the first state or the second state; or corresponding similarity information. The first part may be a fixed payload size, and the second part may be a non-fixed payload size.

[0116] For example, when the number of AI / ML models is large (in the examples shown in Tables 2 and 3, (N+1)), CSI part 1 (fixed payload size) and CSI part 2 (non-fixed payload size) can be applied to report similarity information corresponding to multiple (i.e., (N+1)) AI / ML models. Specifically, CSI part 1 includes at least 2 new indications, such as indication 0 and indication 1. Indication 0 is used to indicate whether the field set B corresponding to at least one AI / ML model is dissimilar (or similar) to the training set B corresponding to (multiple) AI / ML models. Indication 1 is used to indicate the number of such AI / ML models: the corresponding field set B of the AI / ML model is dissimilar (or similar) to the corresponding training set B. The bit width (payload size) of indication 0 is 1 bit, and the bit width (payload size) of indication 1 can be determined based on the following formula: CSI part 2 may include the AI / ML model ID. CSI part 2 may also include corresponding similarity information.

[0117] For example, assume that the indication 1 used to indicate the number of AI / ML models whose corresponding set B is not similar (or completely similar) to the corresponding training set B is 4. In this case, if the first indication as described above is used to indicate that the field set B is not similar to the corresponding training set B of its corresponding AI / ML model, the CSI part 2 can be composed of 4 AI / ML model IDs. This means that the set B corresponding to the 4 AI / ML models is not similar to the corresponding training set B. If the above-mentioned second indication is used to indicate that the field set B is not similar to the corresponding training set B of its corresponding AI / ML model, the CSI part 2 can be composed of 4 AI / ML model IDs and corresponding similarity information (i.e., similarity level). This can indicate that the levels corresponding to the 4 AI / ML models are not "completely similar", in other words, the levels corresponding to the 4 AI / ML models can be "Level 0: Dissimilar", "Level 1: Low Similarity" or "Level 2: High Similarity".

[0118] If the third indication is used to indicate that the live set B is not similar to the corresponding training set B of its corresponding AI / ML model, the CSI part 2 may consist of 4 AI / ML model IDs and corresponding similarity information (i.e., similarity values). This may indicate that the similarity values ​​corresponding to the 4 AI / ML models are less than a predefined threshold (e.g., value 1).

[0119] In some example embodiments, if a CSI report carrying similarity information conflicts with another CSI report carrying information other than the similarity information, the CSI report carrying the similarity is prioritized. Specifically, if a CSI report carrying similarity information conflicts with another CSI report carrying other information other than the similarity information, if the similarity information in the CSI-RS report carrying the similarity information indicates the second state, the CSI report carrying the similarity information is prioritized. The reasons for proposing such (multiple) operations and the beneficial effects of such (multiple) operations are explained below.

[0120] AI / ML model monitoring can be a periodic behavior, so the (time domain) type of the CSI report carrying similarity can be periodic. A CSI report carrying similarity may conflict with another periodic / semi-persistent / aperiodic CSI report (e.g., L1-RSRP / L1-SINR or CSI) that carries information other than similarity. According to existing specifications, the terminal device 120 will give priority to sending semi-persistent / aperiodic CSI reports. However, if the AI / ML model performance deteriorates at this time and the corresponding similarity information is not reported, the network device 110 will continue to apply the currently used AI / ML model whose performance has deteriorated, which is undesirable for model reasoning. In addition, the priority of CSI reports carrying similarity and CSI reports carrying L1-RSRP / L1-SINR or CSI is unclear. It should be noted here that if the time occupancy of the physical channel scheduled to carry the CSI report overlaps in at least one OFDM symbol and is sent on the same carrier, the two CSI reports are said to conflict.

[0121] To address this issue, in some example embodiments, if a first CSI report carrying similarity-related information conflicts with another CSI report carrying information other than similarity information, the first CSI report is prioritized. For example, when a CSI report carrying similarity conflicts with another CSI report carrying L1-RSRP / L1-SINR or CSI, the terminal device 120 may prioritize sending the CSI report carrying similarity, i.e., the terminal device 120 may first send the CSI report carrying similarity (before sending another CSI report carrying L1-RSRP / L1-SINR or CSI).

[0122] Specifically, in some example embodiments, if the first CSI report conflicts with another CSI report carrying information other than similarity information, the first CSI report is prioritized if the similarity information in the first CSI report indicates the second state. For example, the terminal device 120 may prioritize sending the CSI report carrying similarity over sending another CSI report carrying L1-RSRP / L1-SINR or CSI when at least one of the following conditions is met: the reported similarity information indicates at least one "0" (the second state, which indicates "dissimilar"), the reported similarity information indicates at least one of "level 0: dissimilar", "level 1: low similarity" or "level 2: high similarity", or the reported similarity information indicates at least one value less than or equal to a predefined threshold (e.g., value 1). In other words, when the reported similarity information indicates that the field measurement is not completely similar to the training information, the terminal device 120 may prioritize sending the CSI report including the similarity information (first sending the CSI report including the similarity information, and then sending the other CSI report not including the similarity information).

[0123] In some example embodiments, in order to monitor multiple AI / ML models simultaneously (i.e., multiple sets B in the same set A), the network device 110 may send set A (or the union of multiple sets B) to the terminal device 120 instead of one set B, such as Figure 4 In this case, if the terminal device 120 only reports (multiple) similarities, it will cause a waste of beam measurement resources, so in addition to (multiple) similarities, the terminal device 120 can also obtain actual / measured best beam information and report it to the network device 110. In other words, the terminal device 120 reports the similarity information and the actual / measured beam information to the network device 110 at the same time.

[0124] In this case, the CSI report also includes beam information indicating multiple RSs, and in the third group of RSs, multiple RSs have higher beam quality than other RSs, and the third group of RSs consists of at least one first group of RSs. For example, for AI / ML model monitoring, the terminal device 120 can simultaneously report (multiple) indications and the first K beams in set A. For example, the first K beams can be K beams with higher CRI / SSBRI and / or L1-RSRP / L1-SINR values ​​than other beams. Here, K can be specified, fixed and unchanged, for example, K=1. In this case, only the best beam will be reported to the network device 110. K can also be any other positive integer. In this case, multiple beams will be reported to the network device 110. Alternatively, the value of K can be indicated by a high-level parameter "nrofReportedRS".

[0125] Reporting similarity information to the network device 110 may also be event driven. In this case, when a predefined condition is met, the terminal device 120 will report an event to the network device 110. This will refer to Figure 5 Provide detailed explanation.

[0126] Figure 5 A fourth example signaling diagram of a communication process 500 according to some example embodiments of the present disclosure is illustrated. For discussion purposes only, reference will be made to FIG. Figure 3 The process 500 is described. The process 500 may involve the terminal device 120 and the network device 110. For the same or similar operation(s) as those in the process 300 and the process 400, reference may be made to the description of the process 300 and the process 400, and thus the details will be omitted.

[0127] In some example embodiments, the terminal device 120 may determine that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold value for P1 consecutive times within the first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold value for P2 consecutive times within the second duration, where P1 and P2 are positive integers; and then send a scheduling request (SR) message to the network device 110, the SR message indicating that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold value for P1 consecutive times within the first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold value for P2 consecutive times within the second duration. The sixth indication indicates whether the calculated similarity corresponding to the at least one AI / ML model is greater than or equal to the fifth threshold value.

[0128] For example, when the following predefined conditions are met, the terminal device 120 will send a specific event to the network device 110: the determined similarity corresponding to the AI / ML model currently applied by the network device 110 is less than or equal to a predefined threshold for P consecutive times (e.g., 1, 2, or any other positive integer) within a predefined duration. The current AI / ML model, the predefined threshold, P, and the predefined duration can be configured by the network device 110 through RRC / MAC-CE / DCI signaling / messages. Figure 5 In the example shown, P is set to 2. The terminal device 120 receives 222 the scene set B from the network device 110 and determines 530 that the predefined condition is satisfied for the first time. Then, the terminal device 120 receives 542 the scene set B from the network device 110 and determines 550 that the predefined condition is satisfied for the second time. Upon determining that the predefined condition is satisfied for the second time (P=2), an event is triggered.

[0129] The event may be indicated by a new dedicated (or specified) scheduling request (SR) from the terminal device 120 to the network device 110. For example, the configuration of the PUCCH transmission may be provided to the terminal device 120 by an ID of the dedicated SR, and when the event is triggered, the terminal device 120 sends 560 an SR 501 indicating the event to the network device 110 on the PUCCH. In other words, the terminal device 120 sends a PUCCH carrying a new dedicated SR corresponding to the ID of the new dedicated SR configured by the network device 110. On the other side of the communication, the network device 110 receives 562 an SR 501 indicating the event from the terminal device 120. The new dedicated SR 501 is used by the network device 110 to indicate the performance degradation of the current AI / ML model, i.e., the above-mentioned predefined condition is satisfied. In other words, the network device 110 uses the new dedicated SR 501 to determine 570 the performance degradation of the current AI / ML model.

[0130] In some example embodiments, when the predefined conditions are met, the terminal device 120 will report the event to the network device 110, as described above with reference to Figure 5 In addition, the terminal device 120 will report similarity information to the network device 110 in the scheduled PUSCH resources. This will refer to Figure 6 Provide detailed explanation.

[0131] Figure 6 A fifth example signaling diagram of a communication process 600 according to some example embodiments of the present disclosure is illustrated. For discussion purposes only, reference will be made to FIGS. 1 and 2. Figure 5 The process 600 is described. The process 600 may involve the terminal device 120 and the network device 110. For the same or similar operation(s) as in the process 500, reference may be made to the description of the process 500, and thus the details will be omitted.

[0132] In some example embodiments, terminal device 120 sends a media access control element (MAC-CE) message to network device 110, the MAC-CE message including at least one of the following: at least one similarity information determined, at least one model information, or a sixth indication. The sixth indication indicates whether the calculated similarity corresponding to the AI / ML model indicated by the model information in the MAC-CE message is greater than or equal to a fifth threshold.

[0133] For example, after receiving 562 the new dedicated SR 501, the network device 110 schedules PUSCH resources for the terminal device 120 to report similarity information corresponding to the AI / ML model, and sends 610 a scheduled DCI 601 to the terminal device 120, such as Figure 6The scheduling DCI 601 is used to schedule PUSCH resources for the terminal device 120 to report similarity information corresponding to (multiple) new (candidate) AI / ML models. On the other side of the communication, the terminal device 120 receives 612 the scheduled UL resources 601 from the network device 110.

[0134] A new MAC-CE message is introduced to report similarity information. Figure 6 In the example shown, after receiving 612 the scheduling DCI 601, the terminal device 120 sends 620 a new MAC-CE message 602 carrying (or including) similarity information to the network device 110 via the scheduling DCI 601 on the PUSCH resources scheduled by the terminal device 110. On the other side of the communication, the network device 110 receives 622 the new MAC-CE message 602 from the terminal device 120. The new MAC-CE message 602 includes at least similarity information, such as a level or value of similarity.

[0135] In some example embodiments, the event is not for the currently used AI / ML model, but for at least one (multiple) AI / ML model. In this case, instead of a single field set B, a field set A or a union of multiple sets B is sent from the network device 110 to the terminal device 120 so that the terminal device 120 determines whether the predefined condition for triggering the event is satisfied. This will refer to Figure 7 Provide detailed explanation.

[0136] Figure 7 A sixth example signaling diagram of a communication process 700 according to some example embodiments of the present disclosure is illustrated. For discussion purposes only, reference will be made to FIGS. 1 and 2. Figure 6 The process 700 is described. The process 700 may involve the terminal device 120 and the network device 110. For the same or similar operation(s) as in the process 600, reference may be made to the description of the process 600, and thus the details will be omitted.

[0137] exist Figure 7 In the example shown, when the predefined condition is met, the terminal device 120 will report the event to the network device 110. In addition, the terminal device 120 will report (multiple) AI / ML models and corresponding similarity information to the network device 110 in the scheduled PUSCH resources.

[0138] This example is similar to Figure 6 The example shown differs in that Figure 6 In the example, the field set B is used to determine the conditions for triggering an event, and the event is for an AI / ML model currently applied by the network device 110, and in Figure 7In the embodiment, the field set A or the union of multiple sets B is used to determine the condition for triggering an event, and the event is for at least one (multiple) AI / ML model deployed at the network device 110. In the following description, only the differences will be described in detail. Figure 6 For the same and similar operations, please refer to Figure 6 and its description.

[0139] and Figure 6 Similar to the example shown, events can be indicated by new dedicated SRs. Figure 6 Similar to the example shown, the terminal device 120 sends 560 a specific event to the network device 110 via the new dedicated SR 501 when the following predefined condition is met: the determined similarity corresponding to at least one AI / ML model is less than or equal to a predefined threshold for P consecutive times (e.g., 1, 2, and any other positive integer) within a predefined duration.

[0140] At least one AI / ML model may also be configured by the network device 110 via RRC / MAC-CE / DCI signaling / messages.

[0141] and Figure 6 Unlike the example shown, the network device 110 sends 410 the scene set A (or the union of multiple scene sets B) 401 to the terminal device 120, and sends 710 the scene set A (or the union of multiple scene sets B) 401 to the terminal device 120. On the other side of the communication, the terminal device 120 receives 412 the scene set A (or the union of multiple scene sets B) 401 from the network device 110, and determines 530 that the predefined condition is satisfied, that is, the determined similarity corresponding to at least one AI / ML model is less than or equal to the predefined threshold for the first time. The terminal device 120 receives 712 the scene set A (or the union of multiple scene sets B) 401 from the network device 110, and determines 550 that the predefined condition is satisfied, that is, the determined similarity corresponding to at least one AI / ML model is less than or equal to the predefined threshold for the second time. After determining that the predefined condition is satisfied for the second time (P=2), an event is triggered, and the terminal device 120 sends 560 a specific event to the network device 110 via the new dedicated SR 501.

[0142] On the other side of the communication, the network device 110 receives 562 a new dedicated SR 501 from the terminal device 120. This means that the network device 110 can determine 570 performance degradation of at least one AI / ML model based on the new dedicated SR 501.

[0143] Likewise, similar to Figure 6In the illustrated example, after receiving 562 the new dedicated SR, the network device 110 determines 570 an event of performance degradation of at least one (or more) AI / ML model.

[0144] In order to know which (which) AI / ML models have performance degradation, the network device 110 schedules PUSCH resources and sends 610 scheduling DCI 601 to the terminal device 120. Here, the scheduling DCI 601 is used for the terminal device 120 to report similarity information corresponding to at least one (multiple) AI / ML models with performance degradation.

[0145] On the other side of the communication, the terminal device 120 receives 612 the scheduling DCI 601 from the network device 110, and then sends 620 a new MAC-CE message 602 carrying (or including) similarity information to the network device 110 on the PUSCH resources scheduled by the terminal device 110 via the scheduling DCI 601. On the other side of the communication, the network device 110 receives 622 the new MAC-CE message 602 from the terminal device 120. The new MAC-CE message 602 includes at least AI / ML model information (e.g., AI / ML model ID). And, the new MAC-CE message 602 may also include corresponding similarity information (e.g., a level or value of similarity).

[0146] The (multiple) AI / ML model corresponding to the reported AI / ML model information (e.g., AI / ML model ID or AI / ML model index) included in the new MAC-CE message 602 refers to the (multiple) AI / ML model with degraded performance. Therefore, the network device 630 can determine 720 the (multiple) AI / ML model with degraded performance (and (multiple) corresponding similarities) based on the received MAC-CE message 602.

[0147] In some example embodiments, the event may involve at least one candidate (multiple) AI / ML model. When the predefined condition is met, the terminal device 120 may report the event to the network device 110. In addition, the terminal device 120 may report (multiple) new AI / ML models and corresponding similarity information to the network device 110 in the scheduled PUSCH resources. In this case, the field set B corresponding to the AI / ML model currently used by the network device 110 and the union of the set A or multiple sets B are sent from the network device 110 to the terminal device 120 so that the terminal device 120 determines whether the predefined condition for triggering the event is met. This will refer to Figure 8 Provide detailed explanation.

[0148] Figure 8 A seventh example signaling diagram of a communication process 800 according to some example embodiments of the present disclosure is illustrated. For discussion purposes only, reference will be made to FIGS. 1 and 2. Figure 6 The process 800 is described. The process 800 may involve the terminal device 120 and the network device 110. For the same or similar operation(s) as in the process 600, reference may be made to the description of the process 600, and thus the details will be omitted.

[0149] In some example embodiments, the terminal device 120 may determine that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold value for P1 consecutive times within the first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold value for P2 consecutive times within the second duration. P1 and P2 are positive integers. The terminal device 120 may also send a scheduling request (SR) message to the network device 110, indicating that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold value for P1 consecutive times within the first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold value for P2 consecutive times within the second duration.

[0150] exist Figure 8 In the example shown, when the predefined condition is met, the terminal device 120 will report the event to the network device 110. In addition, the terminal device 120 will report at least one (multiple) candidate AI / ML models and corresponding similarity information to the network device 110 in the scheduled PUSCH resources.

[0151] This example is similar to Figure 6 One of the differences between the examples shown is that in Figure 8 , the predefined conditions for triggering an event to report a new dedicated SR message to the network device 110 include predefined condition 1 and predefined condition 2. Therefore, the predefined conditions for the terminal device 120 to send a specific event to the network device 110 via a new dedicated SR message may be: the determined similarity corresponding to the AI / ML model currently applied by the network device 110 is less than or equal to the second predefined threshold value (predefined condition 1) for P1 consecutive times (for example, 1, 2, etc.) within a first duration, and among a group of candidate AI / ML models, the determined similarity corresponding to the candidate AI / ML model is greater than or equal to the third threshold value (predefined condition 2) for P2 consecutive times (for example, 1, 2, etc.) within a second duration. Figure 8 In the example shown, P1=2, P2=1.

[0152] In the following description, only the differences will be described in detail. Figure 6 For the same or similar operations, please refer to Figure 6 and its description.

[0153] and Figure 6 Similar to the example shown, the event may be indicated by a new dedicated SR.

[0154] and Figure 6 Similar to the example shown, the network device 110 sends 220 the field set B 201 to the terminal device 120, and sends 540 the field set B 201 to the terminal device 120. On the other side of the communication, the terminal device 120 receives 222 the field set B 201 from the network device 110, and determines 530 that the predefined condition 1 is satisfied, that is, the determined similarity corresponding to the AI / ML model currently applied by the network device 110 is less than or equal to the predefined second threshold (predefined condition 1) for the first time. The terminal device 120 receives 542 the field set B 201 from the network device 110, and determines 550 that the predefined condition 1 is satisfied, that is, the determined similarity corresponding to the AI / ML model currently applied by the network device 110 is less than or equal to the predefined second threshold (predefined condition 1) for the second time.

[0155] The network device 110 sends 410 a field set A (or a union of multiple field sets B) 401 to the terminal device 120, and on the other side of the communication, the terminal device 120 receives 412 a set A (or a union of multiple field sets B) 401 from the network device 110, and determines 810 that the predefined condition 2 is satisfied, i.e., the determined similarity corresponding to at least one (multiple) new (candidate) AI / ML model is less than or equal to a predefined third threshold for the first time.

[0156] After determining that predefined condition 1 is satisfied for the second time (P1=2) and then predefined condition 2 is satisfied for the first time (P2=1), an event corresponding to the predefined condition is triggered, and the terminal device 120 sends 560 a specific event to the network device 110 via the new dedicated SR 501. The new dedicated SR 501 is used to indicate that the current AI / ML model performance is degraded, that is, the above-mentioned predefined condition is satisfied. Figure 8 As shown, the network device 110 determines 570 an event of performance degradation of the AI / ML model currently applied by the network device 110 .

[0157] With the above Figure 6 Similar to the example shown, after receiving the new dedicated SR 501, the network device 110 schedules PUSCH resources for the terminal device 120 to report similarity information corresponding to one (or more) new AI / ML models, and sends 610 a scheduled DCI 601 to the terminal device 120. Figure 8 As shown, the (multiple) new AI / ML models need to meet the predefined condition 2. The scheduling DCI 601 is used to schedule PUSCH resources for the terminal device 120 to report similarity information corresponding to the (multiple) new AI / ML models. On the other side of the communication, the terminal device 120 receives 612 the scheduled UL resources 601 from the network device 110.

[0158] A new MAC-CE message is introduced to report the similarity information of candidate AI / ML models. Figure 8 In the example shown, after receiving 612 the scheduling DCI 601, the terminal device 120 sends 620 a new MAC-CE message 602 carrying (or including) similarity information of the candidate AI / ML model to the network device 110 via the scheduling DCI 601 on the PUSCH resources scheduled by the terminal device 110. On the other side of the communication, the network device 110 receives 622 the new MAC-CE message 602 from the terminal device 120.

[0159] The new MAC-CE message 602 may include at least candidate AI / ML model information (e.g., AI / ML model ID or AI / ML model index), and may also include corresponding similarity information (e.g., level or value of similarity). In addition, the new MAC-CE message 602 may also include at least an indication. The indication may be used to indicate whether there is a new AI / ML model in the candidate AI / ML model set, that is, whether condition 2 is satisfied. For example, the indication may include 1 bit. If no AI / ML model in the candidate AI / ML model set satisfies condition 2, the indication may be "0". In this case, after receiving the MAC-CE message 602, the network device 110 may know that the performance of the currently used AI / ML has degraded, and the performance of the reported new AI / ML model is good enough to be used thereafter. The network device 110 may then perform an AI / ML model switch.

[0160] In some example embodiments, (multiple) AI / ML models may be deployed at the terminal device 120, rather than at the network device 110 as described above. In this case, at the terminal device 120, the lower layer (e.g., PHY) may report similarity information to the higher layer (e.g., RRC, NAS). As a result, the higher layer will make decisions about model management, such as whether to continue to apply the currently used AI / ML model, perform model switching, or model update. Specifically, the higher layer may provide instructions about the decision to the lower layer to assist the lower layer in performing model management.

[0161] Fig. 9 A flow chart of an example method 900 implemented at a terminal device according to some embodiments of the present disclosure is illustrated. For discussion purposes, the method 900 will be described from the perspective of the terminal device 120 with reference to FIG.

[0162] At box 910, the terminal device 120 receives at least one first group of RSs from the network device 110 at the terminal device 120, and the network device 110 is deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each group of RSs in the at least one first group of RSs corresponds to one AI / ML model in the at least one AI / ML model. At box 920, the terminal device 120 calculates at least one similarity based on the at least one first group of RSs. At box 930, the terminal device 120 determines at least one similarity information based on the calculated at least one similarity, and at least one model information corresponding to the at least one similarity information, and the at least one model information indicates an index of the at least one AI / ML model. At least one model information indicates an index of the at least one AI / ML model. At box 940, the terminal device 120 sends at least one of the determined at least one similarity information or at least one model information to the network device 110.

[0163] In some example embodiments, the payload size of the model information is determined based on the number of at least one AI / ML model.

[0164] In some example embodiments, calculating at least one similarity based on at least one first group of RSs includes: for each group of RSs in the at least one first group of RSs, calculating, based on each group of RSs in the first group of RSs, a similarity between a measured beam quality of each group of RSs in the first group of RSs and a beam quality corresponding to each group of RSs in the first group of RSs in a training data set of a corresponding AI / ML model.

[0165] In some example embodiments, the similarity information includes a first indication indicating a first state or a second state; wherein if the calculated similarity is greater than or equal to a first threshold, the first indication indicates the first state; and if the calculated similarity is less than the first threshold, the first indication indicates the second state.

[0166] In some example embodiments, the payload size of the similarity information is 1 bit.

[0167] In some example embodiments, the similarity information includes a second indication indicating a similarity level corresponding to the calculated similarity.

[0168] In some example embodiments, the payload size of the similarity information is determined based on the number of similarity levels.

[0169] In some example embodiments, the similarity information includes a third indication of a numerical value indicative of the calculated similarity.

[0170] In some example embodiments, the payload size of the similarity information is determined based on the number of scales of similarity.

[0171] In some example embodiments, the sending includes: receiving configuration information at a terminal device from a network device, the configuration information being used for the terminal device to report at least one of the following in a first channel state information (CSI) report: at least one determined similarity information or at least one model information; generating a first CSI report at the terminal device, the first CSI report including at least one determined similarity information or at least one model information; and sending the first CSI report to the network device.

[0172] In some example embodiments, the configuration information further includes a parameter indicating that the first CSI report is used to report at least one similarity information or at least one model information.

[0173] In some example embodiments, a mapping order of the determined at least one similarity information in the first CSI report is determined based on an index of at least one AI / ML model corresponding to the determined at least one similarity information.

[0174] In some example embodiments, the first CSI report includes CSI part 1 and CSI part 2; wherein CSI part 1 includes at least a fourth indication and a fifth indication, the fourth indication indicating the first state or the second state, the fifth indication indicating the number of such AI / ML models: the corresponding similarity information of the AI / ML model indicates the first state or the second state, and when the fifth indication indicates non-zero, CSI part 2 includes at least one of the following: model information, the model information indicating the index of such AI / ML model: the corresponding similarity information of the AI / ML model indicates the first state or the second state; or corresponding similarity information.

[0175] In some example embodiments, if the first CSI report conflicts with another CSI report carrying information other than similarity information, the first CSI report is prioritized.

[0176] In some example embodiments, if the first CSI report collides with another CSI report carrying information other than similarity information, the first CSI report is prioritized if the similarity information in the first CSI report indicates the second state.

[0177] In some example embodiments, the first CSI report further includes beam information indicating a plurality of RSs in a second group of RSs having higher beam quality than other RSs, wherein the second group of RSs is composed of at least one first group of RSs.

[0178] In some example embodiments, the terminal device 120 further determines that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold for P1 consecutive times within a first duration, or the calculated similarity corresponding to at least one AI / ML model is less than a third threshold for P2 consecutive times within a second duration, where P1 and P2 are positive integers; and sends a scheduling request (SR) message to the network device, the SR message indicating that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold for P1 consecutive times within the first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold for P2 consecutive times within the second duration, in which case the sending includes: sending a media access control control element (MAC-CE) message to the network device, the MAC-CE message including at least one similarity information, at least one model information, or at least one of the sixth indications determined.

[0179] In some example embodiments, the calculated similarity corresponding to the AI / ML model indicated by the model information in the MAC-CE message is greater than or equal to a fourth threshold.

[0180] In some example embodiments, the sixth indication indicates whether the calculated similarity corresponding to the at least one AI / ML model is greater than or equal to a fifth threshold.

[0181] In some example embodiments, the terminal device 120 also reports capability information to the network device, wherein the capability information is used to indicate at least one of the following: the terminal device supports AI / ML model monitoring, the terminal device supports similarity measurement, or a training data set corresponding to the AI / ML model is deployed at the terminal device.

[0182] Fig.10 A flow chart of an example method 1000 implemented at a network device according to some embodiments of the present disclosure is illustrated. For discussion purposes, the method 1000 will be described from the perspective of the network device 110 with reference to FIG.

[0183] At block 1010, the network device 110 sends at least one first group of RSs to the terminal device 120. The network device 110 is deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each group of RSs in the at least one first group of RSs corresponds to one AI / ML model in the at least one AI / ML model. At block 1020, the network device 110 receives at least one of the following from the terminal device 120: at least one similarity information determined or at least one model information indicating an index of the at least one AI / ML model. At block 1030, the terminal device 120 determines whether the performance of the at least one AI / ML model is degraded.

[0184] Fig.11 A simplified block diagram of a device 1100 suitable for implementing an embodiment of the present disclosure is illustrated. The device 1100 may be viewed as another example implementation of the terminal device 120 and / or the network device 110 as shown in FIG1 . Thus, the device 1100 may be implemented at or as at least a part of the terminal device 120 or the network device 110.

[0185] As shown, the device 1100 includes a processor 1110, a memory 1120 coupled to the processor 1110, a suitable transmitter (TX) and a receiver (RX) 1140 coupled to the processor 1110, and a communication interface coupled to the TX / RX 1140. The memory 1120 stores at least a portion of the program 1130. The TX / RX 1140 is used for bidirectional communication. The TX / RX 1140 has at least one antenna to facilitate communication, but in fact, the access node mentioned in the present disclosure may have multiple. The communication interface may represent any interface required for communication with other network elements, such as an X2 interface for bidirectional communication between eNBs, an S1 interface for communication between a mobility management entity (MME) / serving gateway (S-GW) and an eNB, a Un interface for communication between an eNB and a relay node (RN), or a Uu interface for communication between an eNB and a terminal device.

[0186] Assume that the program 1130 includes program instructions that, when executed by the associated processor 1110, enable the device 1100 to operate in accordance with embodiments of the present disclosure, as referred to herein. Figures 2 to 10 The embodiments herein may be implemented by computer software executable by a processor 1110 of the device 1100, or by hardware, or by a combination of software and hardware. The processor 1110 may be configured to implement various embodiments of the present disclosure. In addition, the combination of the processor 1110 and the memory 1120 may form a processing component 1550 suitable for implementing various embodiments of the present disclosure.

[0187] The memory 1120 may be of any type suitable for the local technology network and may be implemented using any suitable data storage technology, such as, as non-limiting examples, non-transitory computer-readable storage media, semiconductor-based memory devices, magnetic memory devices and systems, optical memory devices and systems, fixed memory, and removable memory. Although only one memory 1120 is shown in the device 1100, there may be several physically different memory modules in the device 1100. The processor 1110 may be of any type suitable for the local technology network and, as non-limiting examples, may include one or more of a general-purpose computer, a special-purpose computer, a microprocessor, a digital signal processor (DSP), and a processor based on a multi-core processor architecture. The device 1100 may have multiple processors, such as a dedicated integrated circuit chip that is time-slave to a clock synchronized with a main processor.

[0188] In summary, the embodiments of the present disclosure may provide the following solutions.

[0189] The present disclosure provides a communication method, the method comprising: receiving at least one first group of RSs from a network device at a terminal device, the network device being deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each group of RSs in the at least one first group of RSs corresponds to one AI / ML model in the at least one AI / ML model; calculating at least one similarity based on the at least one first group of RSs at the terminal device; determining at least one similarity information based on the calculated at least one similarity and at least one model information corresponding to the at least one similarity information at the terminal device, the at least one model information indicating an index of the at least one AI / ML model; and sending at least one of the following to the network device: the determined at least one similarity information or the at least one model information.

[0190] In one embodiment, as described above, the payload size of the model information is determined based on the number of at least one AI / ML model.

[0191] In one embodiment, the method as described above, calculating at least one similarity based on at least one first group of RS includes: for each group of RS in at least one first group of RS, calculating the similarity between the measured beam quality of each group of RS in the first group of RS and the beam quality corresponding to each group of RS in the first group of RS in the training data set of the corresponding AI / ML model based on each group of RS in the first group of RS.

[0192] In one embodiment, as described above, the similarity information includes a first indication indicating a first state or a second state; wherein if the calculated similarity is greater than or equal to a first threshold, the first indication indicates the first state; and if the calculated similarity is less than the first threshold, the first indication indicates the second state.

[0193] In one embodiment, in the method described above, the payload size of the similarity information is 1 bit.

[0194] In one embodiment, according to the method described above, the similarity information includes a second indication indicating a similarity level corresponding to the calculated similarity.

[0195] In one embodiment, in the method described above, the payload size of the similarity information is determined based on the number of similarity levels.

[0196] In one embodiment, in the method as described above, the similarity information includes a third indication indicating a numerical value of the calculated similarity.

[0197] In one embodiment, in the method described above, the payload size of the similarity information is determined based on the number of scales of similarity.

[0198] In one embodiment, the method as described above, the sending includes: receiving configuration information from a network device at a terminal device, the configuration information being used for the terminal device to report at least one of the following in a first channel state information (CSI) report: at least one determined similarity information or at least one model information; generating a first CSI report at the terminal device, the first CSI report including at least one of the following: at least one determined similarity information or at least one model information; and sending the first CSI report to the network device.

[0199] In one embodiment, according to the method described above, the configuration information further includes a parameter, where the parameter indicates that the first CSI report is used to report at least one similarity information or at least one model information.

[0200] In one embodiment, according to the method described above, the mapping order of the determined at least one similarity information in the first CSI report is determined based on the index of at least one AI / ML model corresponding to the determined at least one similarity information.

[0201] In one embodiment, according to the method described above, the first CSI report includes CSI part 1 and CSI part 2; wherein CSI part 1 includes at least a fourth indication and a fifth indication, the fourth indication indicates the first state or the second state, and the fifth indication indicates the number of such AI / ML models: the corresponding similarity information of the AI / ML model indicates the first state or the second state, and when the fifth indication indicates non-zero, CSI part 2 includes at least one of the following: model information, the model information indicates the index of such AI / ML model: the corresponding similarity information of the AI / ML model indicates the first state or the second state; or corresponding similarity information.

[0202] In one embodiment, according to the method described above, if the first CSI report conflicts with another CSI report carrying information other than similarity information, the first CSI report is given priority.

[0203] In one embodiment, according to the method described above, if the first CSI report conflicts with another CSI report carrying information other than similarity information, the first CSI report is prioritized if the similarity information in the first CSI report indicates the second state.

[0204] In one embodiment, according to the method described above, the first CSI report further includes beam information, the beam information indicating a plurality of RSs in a second group of RSs having higher beam quality than other RSs, wherein the second group of RSs is composed of at least one first group of RSs.

[0205] In one embodiment, the above method also includes: determining that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold for P1 consecutive times within a first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold for P2 consecutive times within a second duration, P1 and P2 are positive integers; and sending a scheduling request (SR) message to the network device, the SR message indicating that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold for P1 consecutive times within the first duration, or the calculated similarity corresponding to at least one AI / ML model is less than the third threshold for P2 consecutive times within the second duration, wherein the sending includes: sending a media access control control element (MAC-CE) message to the network device, the MAC-CE message including at least one of the following: at least one similarity information determined, at least one model information, or a sixth indication.

[0206] In one embodiment, according to the method described above, the calculated similarity corresponding to the AI / ML model indicated by the model information in the MAC-CE message is greater than or equal to the fourth threshold.

[0207] In one embodiment, as described above, the sixth indication indicates whether the calculated similarity corresponding to at least one AI / ML model is greater than or equal to a fifth threshold.

[0208] In one embodiment, the above method also includes: reporting capability information to the network device, wherein the capability information is used to indicate at least one of the following: the terminal device supports AI / ML model monitoring, the terminal device supports similarity measurement, or the training data set corresponding to the AI / ML model is deployed at the terminal device.

[0209] The present disclosure provides a method for communication, the method comprising: sending at least one first group of RSs to a terminal device at a network device, the network device being deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each group of RSs in the at least one first group of RSs corresponding to one AI / ML model in the at least one AI / ML model; receiving at least one of the following from the terminal device: at least one similarity information determined, or at least one model information indicating an index of the at least one AI / ML model; and determining whether performance of the at least one AI / ML model is degraded.

[0210] The present disclosure provides a terminal device, which includes: a processor; and a memory storing a computer program code; the memory and the computer program code are configured to, together with the processor, cause the terminal device to execute the method implemented at the terminal device 120 .

[0211] The present disclosure provides a network device, which includes: a processor; and a memory storing a computer program code; the memory and the computer program code are configured to work with the processor to enable the network device to execute the above method implemented at the network device 110 .

[0212] The present disclosure provides a computer-readable medium having instructions stored thereon, which, when executed by a processor of a device, causes the device to perform the above-mentioned method implemented at the terminal device 120 or the network device 110 .

[0213] Generally, various embodiments of the present disclosure may be implemented using hardware or dedicated circuits, software, logic, or any combination thereof. Some aspects may be implemented using hardware, while other aspects may be implemented using firmware or software that may be executed by a controller, microprocessor, or other computing device. Although various aspects of the embodiments of the present disclosure are illustrated and described as block diagrams, flow charts, or using some other graphical representations, it should be understood that, as non-limiting examples, the blocks, devices, systems, techniques, or methods described herein may be implemented using hardware, software, firmware, dedicated circuits or logic, general-purpose hardware or controllers or other computing devices, or some combination thereof.

[0214] The present disclosure also provides at least one computer program product tangibly stored on a non-transitory computer-readable storage medium. The computer program product includes computer executable instructions, such as instructions included in a program module, which are executed in a device on a target real or virtual processor to perform the above referenced Figure 6-Figure 2 0. Typically, a program module includes a routine, program, library, object, class, component, data structure, etc. that performs a specific task or implements a specific abstract data type. In various embodiments, the functions of the program modules can be combined or split between program modules as needed. The machine executable instructions of the program modules can be executed in local or distributed devices. In distributed devices, program modules can be located in both local and remote storage media.

[0215] The program code for executing the method of the present disclosure can be written in any combination of one or more programming languages. These program codes can be provided to a processor or controller of a general-purpose computer, a special-purpose computer or other programmable data processing device so that the program code is implemented when executed by the processor or controller in the flow chart and / or the block diagram. The program code can be executed completely on the machine, partially on the machine, as an independent software package, partially on the machine and partially on a remote machine, or completely on a remote machine or server.

[0216] The above program code can be embodied in a machine-readable medium, which can be any tangible medium that can contain or store a program for use by an instruction execution system, device or equipment or used in combination with it. The machine-readable medium can be a machine-readable signal medium or a machine-readable storage medium. The machine-readable medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared or semiconductor systems, devices or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium will include an electrical connection with one or more wires, a portable computer floppy disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0217] In addition, although the operations are described in a particular order, this should not be understood as requiring such operations to be performed in the particular order shown or in order or performing all the operations shown to obtain the desired result. In some cases, multitasking and parallel processing may be advantageous. Similarly, although several specific implementation details are included in the above discussion, these should not be interpreted as limitations on the scope of the present disclosure, but rather descriptions of features that may be specific to a particular embodiment. Certain features described in the context of a separate embodiment may also be implemented in combination in a single embodiment. On the contrary, the various features described in the context of a single embodiment may also be implemented in multiple embodiments individually or in any suitable sub-combination.

[0218] Although the present disclosure has been described in language specific to structural features and / or methodological acts, it should be understood that the present disclosure defined in the appended claims is not necessarily limited to the specific features or acts described above. Instead, the specific features and acts described above are disclosed as example forms of implementing the claims.

Claims

1. A method for communication, comprising: At a terminal device, receiving at least one first set of reference signals (RS) from a network device, the network device being deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each set of RSs in the at least one first set of RSs corresponding to one of the at least one AI / ML models; At the terminal device, calculating at least one similarity based on the at least one first group of RSs; At the terminal device, determining at least one similarity information based on the calculated at least one similarity and at least one model information corresponding to the at least one similarity information, the at least one model information indicating an index of at least one AI / ML model; as well as Send at least one of the following to the network device: the determined at least one similarity information, or the at least one model information.

2. The method of claim 1, wherein a payload size of the model information is determined based on the number of the at least one AI / ML model.

3. The method according to claim 1, wherein calculating the at least one similarity based on the at least one first group of RSs comprises: For each group of RSs in the at least one first group of RSs, Based on each RS in the first group of RSs, the similarity between the measured beam quality of each RS in the first group of RSs and the beam quality corresponding to each RS in the first group of RSs in the training data set of the corresponding AI / ML model is calculated.

4. The method according to claim 1, wherein The similarity information includes a first indication indicating a first state or a second state; wherein if the calculated similarity is greater than or equal to a first threshold, the first indication indicates the first state; and If the calculated similarity is less than the first threshold, the first indication indicates the second state. The method according to claim 4 , wherein the payload size of the similarity information is 1 bit.

6. The method according to claim 1, wherein the similarity information comprises: A second indication indicating a similarity level corresponding to the calculated similarity. The method of claim 6 , wherein a payload size of the similarity information is determined based on the number of the similarity levels.

8. The method according to claim 1, wherein the similarity information comprises: A third indication indicating a numerical value of the calculated similarity. 9 . The method of claim 8 , wherein a payload size of the similarity information is determined based on the number of scales of the similarity.

10. The method according to any one of claims 1 to 9, wherein the sending comprises: At the terminal device, receiving configuration information from the network device, the configuration information being used for the terminal device to report at least one of the following in a first channel state information (CSI) report: the determined at least one similarity information, or the at least one model information; At the terminal device, generating the first CSI report, the first CSI report including at least one of the following: the determined at least one similarity information, or the at least one model information; Sending the first CSI report to the network device. 11 . The method according to claim 10 , wherein the configuration information further comprises a parameter indicating that the first CSI report is used to report the at least one similarity information or the at least one model information.

12. The method according to claim 10, wherein a mapping order of the determined at least one similarity information in the first CSI report is determined based on the index of the at least one AI / ML model corresponding to the determined at least one similarity information.

13. The method according to claim 10, The first CSI report includes CSI part 1 and CSI part 2; wherein the CSI part 1 includes at least a fourth indication and a fifth indication, the fourth indication indicating the first state or the second state, the fifth indication indicating the number of the following AI / ML models: the corresponding similarity information of the AI / ML models indicates the first state or the second state, and When the fifth indication indicates non-zero, the CSI part 2 includes at least one of the following: model information indicating the index of the following AI / ML model: the corresponding similarity information of the AI / ML model indicates the first state or the second state, or Corresponding similarity information.

14. The method of claim 10, wherein if the first CSI report conflicts with another CSI report carrying information other than similarity information, the first CSI report is prioritized.

15. The method of claim 14, wherein if the first CSI report collides with another CSI report carrying information other than similarity information, the first CSI report is prioritized if the similarity information in the first CSI report indicates the second state. 16 . The method according to claim 10 , wherein the first CSI report further includes beam information indicating a plurality of RSs having higher beam quality than other RSs in a second group of RSs, wherein the second group of RSs is composed of the at least one first group of RSs.

17. The method according to any one of claims 1 to 9, further comprising: Determine that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold value for P1 consecutive times within the first duration, or that the calculated similarity corresponding to at least one AI / ML model is less than the third threshold value for P2 consecutive times within the second duration, where P1 and P2 are positive integers; and Sending a scheduling request (SR) message to the network device, the SR message indicating that the calculated similarity corresponding to the currently applied AI / ML model is less than the second threshold for P1 consecutive times within the first duration, or the calculated similarity corresponding to the at least one AI / ML model is less than the third threshold for P2 consecutive times within the second duration, The sending includes: A media access control element (MAC-CE) message is sent to the network device, wherein the MAC-CE message includes at least one of the following: the determined at least one similarity information, the at least one model information, or a sixth indication.

18. The method of claim 17, wherein the calculated similarity corresponding to the AI / ML model indicated by the model information in the MAC-CE message is greater than or equal to a fourth threshold.

19. The method of claim 17, wherein the sixth indication indicates whether the calculated similarity corresponding to at least one AI / ML model is greater than or equal to a fifth threshold.

20. The method of claim 1, further comprising: reporting capability information to the network device, The capability information is used to indicate at least one of the following: The terminal device supports AI / ML model monitoring, The terminal device supports similarity measurement, or The training data set corresponding to the AI / ML model is deployed at the terminal device.

21. A method for communication, comprising: At a network device, at least one first set of reference signals (RS) is sent to a terminal device, wherein the network device is deployed with at least one artificial intelligence / machine learning (AI / ML) model, and each RS in the at least one first set of RSs corresponds to one AI / ML model in the at least one AI / ML model; receiving from the terminal device at least one of the following: the determined at least one similarity information, or at least one model information indicating an index of at least one AI / ML model; and Determine whether performance of at least one AI / ML model has degraded.

Images