Communication method and apparatus

Abstract

A communication method and apparatus, which can rationally formulate a corresponding transmission policy, and avoid the waste of communication resources. In the method, a first node can acquire first information, and send the first information to a second node, wherein the first information comprises model transfer performance indicator information of a first model, which model transfer performance indicator information is used for indicating the model transfer performance of the first model. In this way, the second node can determine the model transfer performance of the first model, and rationally plan communication resources required for model transfer, thereby avoiding the waste of communication resources, and achieving relatively high model gains.

Description

Communication methods and devices

[0001] This application claims priority to Chinese Patent Application No. 202411093840.7, filed with the State Intellectual Property Office of China on August 8, 2024, entitled "Communication Method and Apparatus", the entire contents of which are incorporated herein by reference. Technical Field

[0002] This application relates to the field of communications, specifically to a communication method and apparatus. Background Technology

[0003] In wireless communication networks, with the diversification of service requirements and the enhancement of network functions, service implementation, network planning, configuration, and resource scheduling have become increasingly complex. For example, service implementation, network planning, configuration, and resource scheduling may involve technologies such as modulation, coding, transmitters, receivers, multi-antenna technology, or positioning in wireless communication systems. Network devices can perform related operations to implement schemes such as signal modulation and demodulation, information encoding and decoding, channel state information (CSI) feedback, beam management (BM), or mobility management.

[0004] Currently, in the process of implementing the above scheme through wireless communication networks, data transmission is usually involved between communication devices. For example, network devices need to transmit the required model data to the terminal. Data transmission not only affects the actual execution effect of the terminal, but also generates communication overhead. However, in practice, it is difficult for network devices in related technologies to reasonably formulate corresponding transmission strategies, thereby affecting the normal execution of related operations or causing a waste of communication resources. Summary of the Invention

[0005] To address the aforementioned technical problems, embodiments of this application provide a communication method and apparatus that can rationally formulate corresponding transmission strategies and avoid wasting communication resources.

[0006] Firstly, a communication method is provided. This method can be executed by a first node, or by a component of the first node, such as its processor, chip, or chip system, or by a logic module or software capable of implementing all or part of the first node's functions. The following explanation uses the execution of this method by the first node as an example. The communication method includes: acquiring first information; the first information includes model transfer performance index information of a first model; wherein the first model is used to enable artificial intelligence (AI) / machine learning (ML) functions; the model transfer performance index information is used to characterize the model transfer performance of the first model; and sending the first information to a second node.

[0007] Currently, it is difficult for the network side to assess the expected performance of the terminal side after model transmission. This can lead to network devices' model transmission strategies failing to guarantee high expected model returns. For example, network devices may be limited by communication resources and only be able to transmit a portion of the models. However, if the transmitted models do not achieve good model returns, the model transmission strategy developed by the network devices will fail to achieve good model returns. Furthermore, the excessive communication resources consumed can affect the normal transmission of other models. Therefore, in this application, the first node can send relevant information representing the model transmission performance indicators of the first model to the second node. This allows the second node to fully consider the model transmission performance of the first model (also known as the cost-effectiveness of model transmission, the return value of model transmission, the target performance of model transmission, the stability of model transmission, or the robustness of model transmission) when formulating its model transmission strategy. Thus, the second node can rationally plan the communication resources required for model transmission based on the relevant information of the model transmission performance indicators, avoiding waste of communication resources while achieving higher model returns.

[0008] Secondly, a communication method is provided. This method can be executed by a second node, or by a component of the second node, such as its processor, chip, or chip system, or by a logic module or software capable of implementing all or part of the second node's functions. The following explanation uses the execution of this method by a second node as an example. The communication method includes: receiving first information from a first node; the first information includes model transfer performance index information of a first model; wherein the first model is used to enable artificial intelligence (AI) / machine learning (ML) functions; the model transfer performance index information is used to characterize the model transfer performance of the first model; and determining a model transfer strategy based on the first information.

[0009] The technical effects of the second aspect can be referred to the first aspect mentioned above, and will not be repeated here.

[0010] In conjunction with the first or second aspect described above, in one possible design, the model transfer performance information includes at least one model transfer performance indicator, which includes at least one value and / or at least one first value range. Since the model transfer performance is affected by changes in the channel environment, adjustments to communication configuration, or external physical interference, the first node can choose to report the model transfer performance information in the form of a value or a value range, depending on the current situation. This allows the second node to consider the impact of various factors when determining the model transfer strategy, ensuring the normal use of the model.

[0011] In conjunction with the first or second aspect described above, in one possible design, at least one model transfer performance metric value includes at least two model transfer performance metric values. At least one first value range includes at least two first value ranges. In this way, the first information can include multiple values / value ranges of the model transfer performance metric, giving the second node more options when determining the model transfer strategy.

[0012] In conjunction with the first or second aspect described above, in one possible design, the model transfer performance index information includes mathematical statistics used to characterize the model transfer performance of the first model. That is, this application can determine the corresponding model transfer performance index information by performing statistical analysis on the model transfer performance, so that the model transfer performance index information can more accurately or robustly reflect the model transfer performance of the first model.

[0013] In conjunction with the first or second aspect above, in one possible design, the model transfer performance index information includes mathematical statistics for characterizing the model transfer performance index of the first model.

[0014] In conjunction with the first or second aspect above, in one possible design, the first information includes at least one model transmission performance index; one or more of the at least one model transmission performance index have corresponding data features or data feature identifiers.

[0015] In this embodiment, features can be used to represent the characteristics of the signals, channels, or associated data (e.g., data that can be used to train the first model or data that can be used for inference / processing of the first model) that are received / transmitted due to one or more of the following physical factors: device configuration or environmental factors, such as node deployment, antenna configuration, and transmit / receive waveforms. Different nodes may possess different features, which may affect the training and / or inference of the first model. Therefore, in this embodiment, model transmission performance indicators can be bound to data features or data feature identifiers so that the first and second nodes can take into account the impact of device configuration or environmental factors when performing related operations.

[0016] In conjunction with the first or second aspect mentioned above, in one possible design, at least one model conveys performance index information corresponding to at least one data feature or data feature identifier.

[0017] In conjunction with the first or second aspect above, in one possible design, at least one model transfer performance index corresponds one-to-one with at least one data feature or data feature identifier.

[0018] In conjunction with the first or second aspect above, in one possible design, the first model corresponds to at least one model transmitting performance index information, the at least one model transmitting performance index information includes the first model transmitting performance index information, the first model transmitting performance index information corresponds to the first data feature or the first data feature identifier.

[0019] In conjunction with the first or second aspect described above, in one possible design, at least one model transmits performance index information, which also includes second model transmits performance index information, and the second model transmits performance index information corresponds to a second data feature or a second data feature identifier.

[0020] In conjunction with the first or second aspect above, in one possible design, the model transfer performance index information is determined by state information and / or model transfer information; the state information includes first valid state information and / or first failure state information.

[0021] In conjunction with the first or second aspect above, in one possible design, the model transfer information includes first model transfer quantization information; the first model transfer quantization information includes the number of model transfers and / or the model transfer time.

[0022] In conjunction with the first or second aspect above, in one possible design, the first valid state information corresponds to the valid state of the first model, and the valid state includes at least one of the following: the performance of the first model reaches or exceeds the first target performance index; the first model is activated; or the first model is supported / applicable / usable; and / or, the first failure state information corresponds to the failure state of the first model, and the failure state includes at least one of the following: the performance of the first model does not reach or exceed the second target performance index; the first model is inactive; or the first model is not supported / applicable / usable.

[0023] In conjunction with the first or second aspect above, in one possible design, the performance of the first model reaching or exceeding the first target performance index includes the performance of the first model reaching or exceeding the value of the first target performance index, or being within the range of the value of the first target performance index; the performance of the first model not reaching or exceeding the second target performance index includes the performance of the first model not reaching or exceeding the value of the second target performance index, or not being within the range of the value of the second target performance index.

[0024] Since the performance of the first model is often affected by various factors during actual use, the embodiments of this application can be divided into multiple intervals according to the value or value range. For example, the performance interval where the performance of the first model reaches or exceeds the value of the first target performance index is the interval corresponding to the effective state, or the performance interval within the value range of the first target performance index is the interval corresponding to the effective state; the performance interval where the performance of the first model does not reach or exceed the value of the second target performance index is the interval corresponding to the failed state, or the performance interval within the value range of the second target performance index is the interval corresponding to the failed state. In this way, this application can determine the current state of the first model based on the interval in which the performance of the first model falls.

[0025] In conjunction with the first or second aspect described above, in one possible design, the first effective state information includes the first effective quantification information of the first model, which includes effective time information and / or effective count information, wherein the effective time information includes the length of time the first model is in an effective state, and the effective count information includes the number of times the first model is in an effective state; and / or, the first failure state information includes the first failure quantification information of the first model, which includes failure time information and / or failure count information, wherein the failure time information includes the length of time the first model is in a failed state, and the failure count information includes the number of times the first model is in a failed state. Thus, this application can quantify the effective state information and failure state information according to different dimensions, such as time dimension, statistical frequency dimension, etc., to facilitate the evaluation of the model transfer performance index information of the first model.

[0026] In conjunction with the first or second aspect above, in one possible design, the model transmits performance index information including at least one of the following: first proportional information, first difference information, and first probability statistics information.

[0027] In conjunction with the first or second aspect above, in one possible design, the first ratio information includes: the ratio of the first effective quantization information to the model transfer information, or the ratio of the first failure quantization information to the model transfer information, or the ratio of the model transfer information to the first effective quantization information, or the ratio of the model transfer information to the first failure quantization information.

[0028] In conjunction with the first or second aspect described above, in one possible design, the first valid state information is the first valid state information determined within at least one first time period; and / or, the first failure state information is the first failure state information determined within at least one second time period. Since the model transmission performance of the first model is affected by changes in the channel environment, adjustments to communication configuration, or external physical interference, its model transmission performance changes over time. Therefore, in this embodiment, the first valid state information and the first failure state information can be statistically analyzed in time segments to further determine the model transmission performance indicators of the first model at different times.

[0029] In conjunction with the first or second aspect above, in one possible design, at least one first time period includes at least one of the following: at least one time period for performance monitoring of the first model; at least one time period for the first model to be in an active state; at least one time period for the first model to be in a supportable / applicable / available state; at least one time period indicated by a second node; or, at least one time period determined by a first node; and / or, at least one second time period includes at least one of the following: at least one time period for performance monitoring of the first model; at least one time period for the first model to be in an active state; at least one time period for the first model to be in a supportable / applicable / available state; at least one time period indicated by a second node; or, at least one time period determined by a first node.

[0030] In conjunction with the first or second aspect above, in one possible design, at least one first time period is included within at least one monitoring time window, at least one activation time window, at least one applicable time window, or at least one available time window, and at least one monitoring time window, at least one activation time window, at least one applicable time window, or at least one available time window is earlier than the first time unit for transmitting performance metrics of the model; and / or, at least one second time period is included within at least one monitoring time window, at least one activation time window, at least one applicable time window, or at least one available time window, and at least one monitoring time window, at least one activation time window, at least one applicable time window, or at least one available time window is earlier than the second time unit for transmitting performance metrics of the model.

[0031] In conjunction with the first or second aspect above, in one possible design, the first time unit is not located within the monitoring time window, or the first time unit is not located within the activation time window, or the first time unit is not located within the applicable time window, or the first time unit is not located within the available time window; and / or, the second time unit is not located within the monitoring time window, or the second time unit is not located within the activation time window, or the second time unit is not located within the applicable time window, or the second time unit is not located within the available time window.

[0032] Thirdly, a communication device is provided for implementing various methods. The communication device includes modules, units, or means corresponding to the implementation of the methods, wherein the modules, units, or means can be implemented in hardware, software, or by hardware executing corresponding software. The hardware or software includes one or more modules or units corresponding to the functions.

[0033] In some possible designs, the communication device may include a processing module and a transceiver module. The processing module can be used to implement the processing functions in any of the above aspects and any possible implementations thereof. The transceiver module may include a receiving module and a transmitting module, respectively used to implement the receiving function and the transmitting function in any of the above aspects and any possible implementations thereof.

[0034] In some possible designs, the transceiver module can consist of transceiver circuits, transceivers, transceivers, or communication interfaces.

[0035] Fourthly, a communication device is provided, comprising: a processor and a memory; the memory being used to store computer instructions that, when executed by the processor, cause the communication device to perform the method described in any of the above aspects and any possible design thereof.

[0036] Fifthly, a communication device is provided, comprising: a processor and a communication interface; the communication interface being used to communicate with a module outside the communication device; the processor being used to execute computer programs or instructions to cause the communication device to perform the methods described in any of the above aspects and any possible designs thereof.

[0037] A sixth aspect provides a communication device comprising: at least one processor; said processor being configured to execute a computer program or instructions stored in a memory to cause the communication device to perform the methods described in any of the foregoing aspects and any possible designs thereof. The memory may be coupled to the processor, or may be independent of the processor.

[0038] In a seventh aspect, a communication device (e.g., a chip or chip system) is provided, the communication device including a processor for implementing the functions involved in any of the above aspects and any possible designs thereof.

[0039] In some possible designs, the communication device includes a memory for storing necessary program instructions and data.

[0040] In some possible designs, when the device is a chip system, it can be composed of chips or contain chips and other discrete components.

[0041] The communication device described in the third to seventh aspects may be the first node in the first aspect, or a device included in the first node, such as a chip or chip system; or the communication device may be the second node in the second aspect, or a device included in the second node, such as a chip or chip system.

[0042] Eighthly, a communication device is provided, which may be a first node, or a module or unit (e.g., a chip, a chip system, or a circuit) in the first node that performs the methods / operations / steps / actions described in the first aspect, or a module or unit that can be used in conjunction with the first node; or, the communication device may be a second node, or a module or unit (e.g., a chip, a chip system, or a circuit) in the second node that performs the methods / operations / steps / actions described in the second aspect, or a module or unit that can be used in conjunction with the second node.

[0043] It is understandable that when the communication device provided by any of the third to eighth aspects is a chip, the sending action / function of the communication device can be understood as outputting information, and the receiving action / function of the communication device can be understood as inputting information.

[0044] Ninthly, a computer-readable storage medium is provided that stores a computer program or instructions that, when executed on a communication device, enable the communication device to perform the methods described in any of the preceding aspects and any possible designs thereof.

[0045] In a tenth aspect, a computer program product containing instructions is provided, which, when run on a communication device, enables the communication device to perform the methods described in any of the foregoing aspects and any possible design thereof.

[0046] The technical effects of any of the design methods in aspects three through ten can be found in the technical effects of different design methods in aspects one or two, and will not be repeated here. Attached Figure Description

[0047] Figure 1 is a schematic diagram of the structure of a communication system provided in this application;

[0048] Figure 2 is a schematic diagram of another communication system provided in this application;

[0049] Figure 3 is a schematic diagram of an O-RAN system provided in this application;

[0050] Figure 4 is a flowchart illustrating a model transfer method provided in this application;

[0051] Figure 5 is a flowchart illustrating one model transfer method provided in this application;

[0052] Figure 6 is a flowchart illustrating a communication method provided in this application;

[0053] Figure 7 is a schematic diagram of a scenario for model transfer performance provided in this application;

[0054] Figure 8 is a schematic diagram of another scenario for model transfer performance provided in this application;

[0055] Figure 9 is a schematic diagram of a time window provided in this application;

[0056] Figure 10 is a schematic diagram of another time window provided in this application;

[0057] Figure 11 is a flowchart illustrating a communication method provided in this application;

[0058] Figures 12-14 are schematic diagrams of the communication device provided in this application. Detailed Implementation

[0059] In the description of this application, unless otherwise stated, " / " indicates that the objects before and after are in an "or" relationship. For example, A / B can mean A or B. "And / or" in this application is merely a description of the relationship between the related objects, indicating that there can be three relationships. For example, A and / or B can mean: A exists alone, A and B exist simultaneously, and B exists alone. A and B can be singular or plural.

[0060] In the description of this application, unless otherwise stated, "multiple" means two or more. "At least one of the following" or similar expressions refer to any combination of these items, including any combination of a single item or a plurality of items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or multiple.

[0061] Furthermore, to facilitate a clear description of the technical solutions in the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, and the terms "first" and "second" are not necessarily different.

[0062] In the embodiments of this application, the terms "exemplary" or "for example" are used to indicate that something is an example, illustration, or description. Any embodiment or design that is described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design. Specifically, the use of terms such as "exemplary" or "for example" is intended to present the relevant concepts in a specific manner to facilitate understanding.

[0063] It is understood that the term "embodiment" used throughout the specification means that a specific feature, structure, or characteristic related to an embodiment is included in at least one embodiment of this application. Therefore, various embodiments throughout the specification do not necessarily refer to the same embodiment. Furthermore, these specific features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. It is understood that in the various embodiments of this application, the sequence number of each process does not imply the order of execution; the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0064] It is understood that in this application, "...when" and "if" both refer to the corresponding processing that will be carried out under certain objective circumstances, and are not limited to a specific time, nor do they require a judgment action to be performed during implementation, nor do they imply any other limitations.

[0065] It is understood that some optional features in the embodiments of this application can be implemented independently in certain scenarios without relying on other features, such as the current solution on which they are based, to solve the corresponding technical problems and achieve the corresponding effects. Alternatively, they can be combined with other features as needed in certain scenarios. Correspondingly, the apparatus given in the embodiments of this application can also implement these features or functions, which will not be elaborated here.

[0066] In this application, unless otherwise specified, the same or similar parts between the various embodiments can be referred to each other. In the various embodiments of this application, unless otherwise specified or there is a logical conflict, the terminology and / or descriptions between different embodiments are consistent and can be mutually referenced. Technical features in different embodiments can be combined to form new embodiments based on their inherent logical relationships. The following descriptions of the embodiments of this application do not constitute a limitation on the scope of protection of this application.

[0067] To facilitate understanding of the technical solutions of the embodiments of this application, a brief introduction to the relevant technologies of this application is given below.

[0068] In wireless communication networks, with the diversification of service requirements and the enhancement of network functions, service implementation, network planning, configuration, and resource scheduling have become increasingly complex. For example, service implementation, network planning, configuration, and resource scheduling may involve technologies such as modulation, coding, transmitters, receivers, multi-antenna technology, or positioning in wireless communication systems. Network devices can perform related operations to implement schemes such as signal modulation and demodulation, information encoding and decoding, CSI feedback, BM, or mobility management.

[0069] For example, communication devices can execute the technical solutions in the aforementioned business scenarios using traditional algorithms, or they can apply artificial intelligence (AI) / machine learning (ML) technologies to these solutions. AI / ML technology refers to training a model using relevant data, thereby using the trained model to achieve a specific purpose. The purpose that the model can achieve is related to the data used during training.

[0070] Taking AI / ML-based CSI feedback as an example, this scenario includes AI / ML-based CSI compression and AI / ML-based CSI prediction. AI / ML-based CSI compression refers to the terminal compressing the downlink CSI (measured by the terminal) using AI / ML technology. The terminal then transmits the compressed CSI to the network device over the air interface. The network device then restores (decompresses) the CSI using AI / ML. Compared to traditional compression algorithms, AI / ML-based compression algorithms have higher compression ratios and better CSI restoration capabilities. Therefore, the terminal can feed back more CSI with less air interface overhead, enabling the network device to perform more accurate downlink precoding. AI / ML-based CSI prediction refers to the network device using AI / ML technology to predict the downlink CSI for future times based on the current / historical downlink CSI, and then performing precoding according to the predicted CSI. This scheme predicts CSI that better matches the channel state when downlink data is scheduled, thus overcoming the channel aging problem and achieving more accurate downlink precoding. The training data for models involved in business scenarios based on AI / ML CSI feedback includes CSI.

[0071] Taking a business scenario based on AI / ML (Beam Builder) as an example, network devices and terminal devices can use AI / ML technology to predict transmit and / or receive beams. For instance, AI / ML can be used to infer the optimal beam from a small number of beam scan results. Compared to traditional solutions that require scanning a large number of beams to obtain the optimal beam, AI / ML-based beam prediction reduces the processing overhead of beam scanning. For example, a terminal can scan a small number of beams and then use an AI / ML model to predict the optimal beam from a large number of candidate beams, thus avoiding the need to scan all candidate beams and reducing overhead. The small number of beams scanned by the terminal can be sparse or wide beams, while the candidate beams can be dense or narrow beams. The terminal can input the beam information scanned at the current / historical time into the model to predict the optimal beam at future time points, thus avoiding the need to perform beam scanning again at future time points and improving beam scanning efficiency. In AI / ML-based business scenarios involving BM, the training data for the models includes beam information, such as beam ID and / or the reference signal receiving power (RSRP) of the corresponding beam.

[0072] Taking AI / ML-based positioning as an example, the communication device inputs channel information into the AI / ML model to infer the intermediate parameters required for positioning, or directly obtains the location coordinates. Compared to traditional positioning algorithms, the intermediate parameters or positioning coordinates obtained based on AI / ML are more accurate. The training data for the model involved in AI / ML-based positioning scenarios includes channel information and / or location information. Channel information includes power information, phase information, delay information, distance information, velocity information, channel scattering information, and channel-related information such as line-of-sight (LOS) / non-line-of-sight (NLOS) information.

[0073] In implementing the above scheme through wireless communication networks, data transmission is typically involved between communication devices. For example, network devices need to transmit the required model data to the terminal. Data transmission not only affects the actual execution performance of the terminal but also incurs communication overhead. However, in practice, it is difficult for the network side to assess the expected performance of the terminal after model transmission. This can lead to the network device's model transmission strategy failing to guarantee high expected model returns. For instance, due to communication resource limitations, the network device can only transmit a portion of the model, but the transmitted model does not achieve good model returns. This results in the network device's model transmission strategy failing to achieve good model returns, and the normal transmission of other models is affected due to the large amount of communication resources consumed. Based on this, the first node in this application can send relevant information about the model transfer performance index representing the first model to the second node, so that the second node can fully consider the model transfer performance of the first model (also known as the cost-effectiveness of model transfer, or the benefit value of model transfer, or the target performance of model transfer, or the stability of model transfer, or the robustness of model transfer) when formulating the model transfer strategy. Therefore, the second node can reasonably plan the communication resources required for model transfer based on the relevant information about the model transfer performance index, avoid waste of communication resources, and achieve higher benefits for the model.

[0074] The technical solutions of this application embodiment can be used in various communication systems, including third-generation partnership project (3GPP) communication systems, such as fourth-generation (4G) systems like Long Term Evolution (LTE), fifth-generation (5G) systems like New Radio (NR), LTE and 5G hybrid networking systems, integrated communication and sensing systems, non-terrestrial networks (NTN), device-to-device (D2D) communication systems, vehicle-to-everything (V2X) communication systems, machine-type communication (MTC) systems, Internet of Things (IoT) systems, or other future communication systems. The communication system can also be a non-3GPP communication system; there is no limitation on this.

[0075] The communication systems described above are merely illustrative examples, and are not limited to those described herein. The communication systems provided in this application do not impose any limitations on the solutions described herein. This will be explained uniformly here and will not be repeated below.

[0076] Figure 1 illustrates a possible, non-limiting system diagram. As shown in Figure 1, the communication system 10 includes a radio access network (RAN) 100 and a core network (CN) 200. RAN 100 includes at least one RAN node (110a and 110b in Figure 1, collectively referred to as 110) and at least one terminal (120a-120j in Figure 1, collectively referred to as 120). RAN 100 may also include other RAN nodes, such as wireless relay devices and / or wireless backhaul devices (not shown in Figure 1). Terminal 120 is wirelessly connected to RAN node 110. RAN node 110 is wirelessly or wired connected to core network 200. The core network node in core network 200 and RAN node 110 in RAN 100 can be different physical devices, or they can be the same physical device integrating core network logical functions and radio access network logical functions.

[0077] In one possible implementation, a core network node can refer to a device in the core network 200 that provides service support to terminal 120. In this embodiment, the core network node in the core network 200 includes a sensing function (SF) network element. The SF network element is mainly used to implement sensing functions, which may include sensing control functions and / or sensing computation functions. Furthermore, the SF network element can also support sensing billing functions when terminal 120 and / or RAN node 110 perform sensing. For example, the sensing control function may include determining sensing devices, sensing nodes, etc. A sensing device can be understood as a device that sends and / or receives sensing signals. Further, the sensing device also performs corresponding signal processing based on the received echo signals to obtain sensing measurement data. For example, the sensing device can be RAN node 110 or terminal 120, etc. A sensing node can refer to a network node in the wireless network that participates in the sensing service process. The sensing computation function may include performing corresponding signal processing on the echo signals received by the sensing device to obtain sensing measurement data, and further processing based on the sensing measurement data and application information to obtain sensing results, etc.

[0078] For example, an SF network element can sometimes be called a communication device; for instance, an SF network element can be understood as a communication device with core network sensing capabilities. Furthermore, an SF network element can also be called a sensing server, etc., without limitation.

[0079] In one possible scenario, the function of the SF network element can be implemented by the network data analytics function (NWDAF) network element, or the SF network element and the NWDAF network element can be co-located.

[0080] Optionally, in addition to SF network elements, core network nodes in core network 200 may also include at least one of the following: access and mobility management function (AMF) network elements, session management function (SMF) network elements, user plane function (UPF) network elements, policy control function (PCF) network elements, unified data management (UDM) network elements, application function (AF) network elements, network exposure function (NEF) network elements, network slice selection function (NSSF) network elements, or location management function (LMF) network elements, etc. Of course, core network 200 may also include other core network nodes without restriction.

[0081] The AMF (Agency Flow Management) network element is deployed in the core network 200 to provide mobility management and connectivity management for the network, such as user location updates, user registration with the network, and user handover. The AMF network element can act as an intermediate route between the LMF, SMF, and RAN 100. The SMF network element is mainly responsible for session management in the mobile network, such as session establishment, modification, and release. The UPF (User Plane Function) network element is a user plane function element, mainly responsible for connecting to external networks and processing user packets, such as forwarding and charging. The PCF (Programmable Flow Function) network element is mainly responsible for providing policies to the AMF and SMF, such as Quality of Service (QoS) policies and slice selection policies. The UDM (User DM) network element is used to store user data, such as subscription information and authentication / authorization information. The AF (Agency Flow) network element is responsible for providing services to the 3GPP network. The NEF (Network Flow Function) network element is mainly used to open the capabilities of various network functions and is responsible for converting internal and external information. The LMF network element is a device or component deployed in the core network 200 to provide positioning functions for the terminal 120; for example, the LMF network element can initiate a positioning process to locate a specific terminal.

[0082] It should be noted that in this application, network elements can also be referred to as entities or functional entities. For example, an SF network element can also be referred to as an SF entity or an SF functional entity. In addition, the aforementioned AMF network elements, SMF network elements, UPF network elements, PCF network elements, UDM network elements, AF network elements, NEF network elements, and LMF network elements may have other names in future communication systems, and this application does not impose specific limitations on them.

[0083] In one possible implementation, RAN 100 can be a 3rd Generation Partnership Project (3GPP) related cellular system, such as a 4G, 5G mobile communication system, or a future-oriented evolution system. RAN 100 can also be an open RAN (O-RAN or ORAN), a cloud radio access network (CRAN), an NTN network (such as an NTN supporting pass-through mode and / or regenerative mode, or an NTN supporting eye-viewing mode (earth fixed cell) and / or non-eye-viewing mode (earth moving cell), or a wireless fidelity (WiFi) system. RAN 100 can also be a communication system that integrates two or more of the above systems.

[0084] RAN node 110, sometimes also referred to as access network equipment, RAN entity, or access node, constitutes part of the communication system and is used to help terminals achieve wireless access. Multiple RAN nodes 110 in RAN 100 can be of the same type or different types. In some scenarios, the roles of RAN node 110 and terminal 120 are relative. For example, network element 120i in Figure 1 can be a helicopter or drone, which can be configured as a mobile base station. For terminals 120j accessing RAN 100 through network element 120i, network element 120i is a base station; but for base station 110a, network element 120i is a terminal. RAN node 110 and terminal 120 are sometimes both referred to as communication devices. For example, network elements 110a and 110b in Figure 1 can be understood as communication devices with base station functions, and network elements 120a-120j can be understood as communication devices with terminal functions.

[0085] For RAN node 110, in one possible scenario, RAN node 110 can be a base station, an evolved NodeB (eNodeB, also known as eNB), an access point (AP), a transmission reception point (TRP), a next-generation NodeB (gNB), a next-generation base station in a future mobile communication system, or an access node in a WiFi system, etc. RAN node 110 can be a macro base station (as shown in Figure 1, 110a), a micro base station or indoor station (as shown in Figure 1, 110b), a relay node or donor node, or a radio controller in a CRAN scenario. Examples include: satellite base stations, radio network controllers (RNCs), base station controllers (BSCs), base transceiver stations (BTSs), home base stations (e.g., home evolved NodeBs, or home NodeBs, HNBs), relay stations, balloon stations, drone stations, radio backhaul nodes, or grant nodes (G nodes) in satellite flash, etc. It is understood that network equipment can be either ground-based or non-ground-based (such as satellites, drones, high-altitude communication equipment, etc.). Furthermore, the names of network equipment with base station functions may differ in communication systems employing different wireless access technologies; this application does not limit this. Optionally, RAN node 110 can also be a server, wearable device, vehicle, or in-vehicle equipment. For example, the access network equipment in vehicle-to-everything (V2X) technology can be a roadside unit (RSU). RAN node 110 is also referred to as a next-generation RAN (NG-RAN) node.

[0086] In another possible scenario, multiple RAN nodes 110 collaborate to assist the terminal in achieving wireless access, with each RAN node 110 implementing a portion of the base station's functions. For example, a RAN node 110 can be a central unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU), etc. CUs and DUs can be set up separately or included in the same network element, such as a baseband unit (BBU). RUs can be included in radio frequency equipment or radio frequency units, such as remote radio units (RRUs), active antenna units (AAUs), or remote radio heads (RRHs).

[0087] In different systems, CU (or CU-CP and CU-UP), DU, or RU may have different names, but those skilled in the art will understand their meaning. For example, in an ORAN system, CU can also be called O-CU (open CU), DU can also be called O-DU, CU-CP can also be called O-CU-CP, CU-UP can also be called O-CU-UP, and RU can also be called O-RU. For ease of description, this application uses CU, CU-CP, CU-UP, DU, and RU as examples. Any of the units among CU (or CU-CP, CU-UP), DU, and RU in this application can be implemented through software modules, hardware modules, or a combination of software and hardware modules.

[0088] In one possible scenario, terminal 120 can be a device used to implement wireless communication functions, such as a terminal, a chip or circuit that can be used in the terminal, or an entity associated with the terminal. Specifically, terminal 120 can be user equipment (UE), access terminal, terminal unit, terminal station, mobile station (MS), mobile station, remote station, remote terminal, mobile device, wireless communication equipment, terminal agent or terminal device, subscriber unit, smartphone, wireless data card, tablet computer, wireless modem, laptop computer, machine-type communication (MTC) terminal, tag, etc., in a 5G network or a future evolved public land mobile network (PLMN). The access terminal can be a cellular phone, cordless phone, session initiation protocol (SIP) phone, wireless local loop (WLL) station, personal digital assistant (PDA), handset with wireless communication capabilities, computing device or other processing device connected to a wireless modem, in-vehicle device or wearable device, virtual reality (VR) terminal, augmented reality (AR) terminal, wireless terminal in industrial control, wireless terminal in self-driving, wireless terminal in remote medical care, wireless terminal in smart grid, wireless terminal in transportation safety, wireless terminal in smart city, wireless terminal in smart home, or terminal node (T-node) in StarSpark, etc. In one possible implementation, terminal 120 can be mobile or fixed. It is understood that the terminal and the mobile user can be completely independent. All information related to a user can be stored in a subscriber identity module (SIM) card, which can be used on a terminal device.The terminal can send and / or receive signals via the air interface to complete the interaction with network-side devices.

[0089] The chip or circuit in the terminal includes components inside the terminal, such as at least one of a chip, a central processing unit (CPU), a network processing unit (NPU), and a terminal radio frequency module.

[0090] Entities associated with the terminal include terminal-side servers, computing / processing nodes, computing / processing entities, computing / processing units, and servers such as over-the-top (OTT) servers. OTT refers to various services provided to users by a third party other than the network operator via the operator's network. Examples of OTT services include OTT voice communication services, OTT multimedia services, and OTT data processing services. The terminal interacts with relevant information (e.g., data) through communication with this associated network entity. For example, this associated network entity and the terminal may belong to the same vendor. Since model training, model selection, etc., may not be executed on the terminal but rather on the terminal-side OTT server, the term "terminal" in this embodiment also includes the terminal-side OTT server.

[0091] It should be understood that the terminal in this embodiment may also be referred to as the "UE side" or the "UE part".

[0092] As exemplarily shown in Figure 2, this is one exemplary implementation of the system shown in Figure 1. The communication system may include an AI / ML node, a first terminal, and a first device. The first device can provide services to the terminal.

[0093] Optionally, the first device can be a server, which can be a single server or a server cluster consisting of multiple servers. In some embodiments, the server cluster can also be a distributed cluster. The server can provide services to the chip, and therefore can also be called a chip server. Alternatively, the first device can be a first network element in the core network. The first device can train models for the terminals it serves, or transmit models or model inference results.

[0094] Optionally, the communication system shown in Figure 2 may include network devices. Network devices can be any type of device deployed in the access network capable of wireless communication with terminals (e.g., a first terminal, a second terminal). They can also be chips or chip systems that can be configured in the aforementioned devices, logical nodes or logical modules, or functions implemented in software. They are primarily responsible for air interface-side wireless physical control functions, resource scheduling, wireless resource management, quality of service management, data compression and encryption, wireless access control, and mobility management. Specifically, network devices can be devices that support wired access or devices that support wireless access.

[0095] In some embodiments, when the model data is stored in a network device, the model can be directly transmitted to the first terminal by the network device. When the model data is stored in a first device, it can be transmitted to the first terminal by the first device via UP plane data transmission, or the server can transmit the model to the network device, and then the network device can forward it to the first terminal.

[0096] The AI / ML node in Figure 2 is used to support the use of AI / ML technology in AI / ML scenarios.

[0097] Optionally, the AI / ML node can be deployed in one or more of the following locations in the communication system shown in Figure 2: network device, first terminal, second terminal, first device, etc., or the AI / ML node can be deployed independently, for example, in a location other than any of the above devices.

[0098] For example, AI / ML nodes can be deployed on the host or cloud server of an OTT system. When a device deploying an AI / ML node communicates with a network device, that device can also act as a terminal in the communication system. When a device deploying an AI / ML node communicates with a terminal, that device can also act as a network device in the communication system.

[0099] It is understood that this application does not limit the number of AI / ML nodes. For example, when there are multiple AI / ML nodes, these nodes can be divided based on function, such as different AI / ML nodes being responsible for different functions.

[0100] It can also be understood that AI / ML nodes can be independent devices, or they can be integrated into the same device to implement different functions. Alternatively, they can be network elements in hardware devices, software functions running on dedicated hardware, or virtualization functions instantiated on a platform (e.g., a cloud platform). This application does not limit the specific form of the aforementioned AI / ML nodes.

[0101] AI / ML nodes can be AI / ML network elements or AI / ML modules.

[0102] It is understood that Figure 2 above is merely a schematic diagram and does not constitute a limitation on the applicable scenarios of the technical solutions provided in this application. Those skilled in the art should understand that in specific implementation processes, the communication system shown in Figure 2 may include fewer devices than those shown in Figure 2, or the communication system shown in Figure 2 may also include other devices. At the same time, the number of devices in the communication system shown in Figure 2 can be determined according to specific needs and is not limited.

[0103] Optionally, the devices in Figure 2, such as the first device, the first terminal, the network device, and the second terminal, can also be referred to as communication devices. They can be general-purpose devices or special-purpose devices. This application embodiment does not specifically limit them in this regard.

[0104] Optionally, the functions of each device in Figure 2 of this application can be implemented by one device, multiple devices working together, or one or more functional modules within a single device. This application does not specifically limit these functions. It is understood that the aforementioned functions can be network elements in hardware devices, software functions running on dedicated hardware, a combination of hardware and software, or virtualization functions instantiated on a platform (e.g., a cloud platform).

[0105] In one possible implementation, the network device (e.g., access node or core network node) and terminal 120 in this embodiment can also be referred to as communication devices. These devices can be general-purpose or dedicated devices. The network device may include an access node (RAN node), an operation administration and maintenance (OAM) device, or a core network node. For the OAM device, it may include devices in the element management system (EMS) or the network management system (NMS). It should be understood that the network device in this embodiment can also be referred to as a "network side" or a "network part." This embodiment does not specifically limit its use in this regard.

[0106] In one possible implementation, the relevant functions of the terminal 120 or network device in this application embodiment can be implemented by one device, multiple devices working together, or one or more functional modules within a single device. This application embodiment does not specifically limit this. It is understood that the above functions can be network elements in hardware devices, software functions running on dedicated hardware, a combination of hardware and software, or virtualization functions instantiated on a platform (e.g., a cloud platform).

[0107] It should be noted that a RAN node can be a device or a component within a device in the aforementioned NG-RAN, such as an ng-eNB node, a gNB node, or a transmission point (TP), transmission and reception point (TRP) within an ng-eNB node and a gNB node, or a central unit (CU) integrated on the NG-RAN. A RAN node can also be a network element with transmission capabilities, such as a transmission measurement function (TMF) network element. In some embodiments, a RAN node can also be an access node in an O-RAN system. A RAN typically consists of a series of modules, such as antennas, RRUs, and BBUs. Traditional RAN architectures define the overall reception and output of a RAN node but do not restrict the transmission and communication between internal modules. O-RAN architectures define the architectural connections and standardized interfaces between various modules within the RAN, allowing the RAN to be decoupled into multiple standard modules, thereby enabling the combination and replacement of modules.

[0108] As exemplified, Figure 3 illustrates a possible, non-limiting structural diagram of an O-RAN system. The Service Management and Orchestration Framework (SMO), as the network management device in the O-RAN, is used for the operation and management of devices within the O-RAN. The Non-Real-Time RAN Intelligent Controller (Non-RT RIC), located within the SMO module, implements non-real-time intelligent management of RAN functions, such as enabling AI / ML workflows including model training and updates, and guiding applications / functions within the Near-RT RIC based on policies. The Near-Real-Time RAN Intelligent Controller (Near-RT RIC) implements near-real-time intelligent management of the RAN. Through data collection and related operations on the E2 interface, it achieves near-real-time control and optimization of O-RAN modules and resources.

[0109] The O-RAN central unit (O-CU) comprises the O-RAN central unit control plane (O-CU-CP) and the O-RAN central unit user plane (O-CU-UP). The O-CU implements the radio resource control (RRC) layer, the packet data convergence protocol (PDCP) layer, the service data adaptation protocol (SDAP) layer, and other control functions. Specifically, the O-CU-CP implements the RRC layer functions and the PDCP control plane functions. The O-CU-UP implements the SDAP layer functions and the PDCP user plane functions.

[0110] The O-RAN distributed unit (O-DU) is used to implement the radio link control (RLC) layer, media access control (MAC) layer, and higher physical layer (Higher PHY). The higher physical layer functions include one or more of the following: forward error correction (FEC) encoding / decoding, scrambling / descrambling, or modulation / demodulation.

[0111] The O-RAN radio unit (O-RU) is used to implement lower physical layer (PHY) functions and radio frequency (RF) functions. These PHY functions include one or more of the following: fast Fourier transform (FFT) / inverse fast Fourier transform (iFFT), digital beamforming, or extraction and filtering of the physical random access channel (PRACH). In other words, the O-RU possesses functions similar to TRP and RRH RF devices, as well as PHY processing capabilities. Furthermore, the O-RU, O-CU, and O-DU can also be used as a single unit, i.e., the O-eNB / gNB, to implement the aforementioned functions.

[0112] O-RAN cloud (O-Cloud) is a cloud computing platform that includes physical infrastructure nodes for hosting O-RAN functions such as RIC and O-DU. O-Cloud supports software components (such as operating systems, virtual machine monitoring, and container runtimes), management, and orchestration functions.

[0113] In one possible scenario, the O-RAN system also includes a sensing unit (SU). The SU is mainly used to implement sensing-related functions, such as sending sensing signals and / or receiving echo signals of sensing signals, performing corresponding signal processing based on the received echo signals to obtain sensing measurement data, and performing sensing-related processing, etc.

[0114] As one possible implementation, a RAN node may include at least one of CU, DU, SU, and RU. A communication interface exists between CU and SU. A communication interface may or may not exist between SU ​​and DU. If no communication interface exists between SU ​​and DU, SU and DU can communicate through CU.

[0115] In the O-RAN architecture, the module that receives the report of the difference between the twin channel and the measurement channel can be CU, RT RIC, Non-RT RIC, etc. DU is responsible for receiving signals, signal processing, multipath measurement, and channel difference calculation.

[0116] For example, an O-RAN system includes communication interfaces between newly added internal components and other communication interfaces. For instance, the A1 interface serves as the interface between Non-RT RICs and Near-RT RICs, used for intelligent and dynamic control of radio resources within the O-RAN. Non-RT RICs can provide policies, enriched information, and ML model updates to Near-RT RICs via the A1 interface, while Near-RT RICs can provide policy feedback to Non-RT RICs via the A1 interface.

[0117] The E2 interface is an open interface between two endpoints used to connect the Near-RT RIC and the RAN node. The RAN node includes the CU and DU in 5G, the O-RAN compatible eNB in ​​4G, and the O-CU (O-CU-CP and / or O-CU-UP) and / or O-DU in O-RAN. The Near-RT RIC can obtain data collection and feedback from the RAN node through the E2 node, and the RAN node can obtain control feedback from the Near-RT RIC through the E2 node.

[0118] The O1 interface is the interface between the management entity in the SMO and the O-RAN module, used for operation management. This interface enables network management (such as fault management, configuration management, billing management, performance management, and security management, also known as FCAPS management), software management, and file management. The O2 interface is the interface between the SMO and the infrastructure management framework that supports O-RAN virtual network functions.

[0119] The Open Fronthaul (FH) CUS-Plane interface includes a control plane (C-Plane), a user plane (U-Plane), and a synchronization plane (S-Plane). The control plane is used for real-time control between the O-DU and O-RU, such as transmitting beamforming weights from the O-DU to the O-RU or performing power control from the O-DU to the O-RU. The user plane is used to transmit communication data between the DU and RU for access network devices and terminals. The synchronization plane is used by the O-DU to provide clock synchronization to the O-RU. The Open FH M-Plane interface is the management plane interface, used for connection between the O-RU and O-DU, as well as the SMO, enabling management, monitoring, and configuration functions.

[0120] In addition, the NG interface is the interface between RAN nodes (e.g., base stations, CUs, CU-CPs, CU-UPs) and the core network; NG-u is the user plane NG interface; and NG-c is the control plane NG interface. The Xn interface is the interface between NR RAN nodes; Xn-u is the user plane Xn interface; and Xn-c is the control plane Xn interface. The X2 interface is the interface between LTE RAN nodes; X2-u is the user plane X2 interface; and X2-c is the control plane X2 interface. In NR systems, the X2 interface is mainly used in E-UTRA-NR dual connectivity (EN-DC) scenarios, where the primary base station is an LTE RAN node connected to the LTE core network via the X2 interface. The E1 interface is the interface between CU-CPs and CU-UPs; the F1-c interface is the interface between CU-CPs and DUs; and the F1-u interface is the interface between CU-UPs and DUs.

[0121] In some embodiments, the communication device in this application can implement AI / ML workflow (also known as model operation), including data collection, model training, model transfer, model update, model inference, model monitoring, and model management.

[0122] Data collection refers to the process by which a communication device gathers necessary data before performing other model operations (model training, model updating, data analysis, model inference, and model monitoring). During data collection, the communication device can divide data corresponding to different characteristics into different datasets and train different models to better adapt the model to specific features during the inference phase, achieving better performance. Model training refers to training the model using the collected training data to obtain a trained model. Model transfer, also known as model transmission or model sending, refers to transmitting the model's structure and / or parameters. Model updating refers to updating the trained model using newly collected data, similar to retraining the model. Model inference refers to inputting the collected data into the trained model and outputting the corresponding result data. Model monitoring refers to monitoring the relevant states during model inference to detect the model's inference performance. The model's inference performance may be affected by environmental changes. When the channel environment changes, if the model can no longer adapt well to the channel environment, the model's reliability will decrease. To ensure model reliability, network devices and / or terminals can employ model monitoring to detect model inference performance. When performance degrades, network devices and / or terminals need to promptly switch or deactivate the model. Model management refers to the management operations at each stage of the model, including model selection, model switching, model activation, and model deactivation.

[0123] It should be understood that the models involved in the embodiments of this application can be described as functions (such as AI functions or ML functions), characteristics, or algorithms, etc., and "model operation" can also be called "functional operation". The above-mentioned model training, model transfer, model update, model inference, model monitoring, and model management can be replaced by functional training, functional update, functional inference, functional monitoring (or performance monitoring), and functional management, respectively. Here, "function" can be understood as a function corresponding to artificial intelligence. A model can implement one or more functions, and one or more models can also work together to implement a function.

[0124] The node executing model inference can be a terminal, a network device, or both. Therefore, model deployment methods are divided into one-sided model deployment and two-sided model deployment.

[0125] One-sided model deployment refers to the process where, for a single air interface feature, use case, or function, the entire inference process can be completed by deploying an AI / ML model solely on the network device side (referred to as the network device-side AI / ML model) or on the terminal side (referred to as the terminal-side AI / ML model). For example, BM (Balanced Machine) or positioning technology use cases can complete the entire inference process with only one-sided AI / ML model deployment. One-sided models include network-side models and UE-side models. For network-side models, the terminal can report measurement information to the network device as input data for model inference / training / monitoring / management. For terminal-side models, the terminal can perform model inference / training / monitoring based on measurement information and report inference or monitoring results to the network device. Regardless of whether it's a network-side or terminal-side model, for some use cases, the network device needs to instruct or configure corresponding resources (e.g., reference signal resources) for the terminal's measurements to generate the reported measurement information or for terminal-side model inference, monitoring, or management.

[0126] Bilateral model deployment refers to deploying AM / ML models on both the terminal and network device for a single air interface feature / use case / function. In this case, the network device-side model and the terminal-side model need to be paired to complete the entire inference process for that air interface feature. Taking CSI compression as an example, the terminal-side model infers the original CSI to achieve compression and feeds back the compressed CSI data to the network device via the air interface. Upon receiving the compressed CSI data, the network device uses its own model to infer the compressed CSI data to achieve decompression and obtain the restored CSI. For well-matched network device-side and terminal-side models, the restored CSI is closer to the original CSI. For poorly matched models, the restored CSI differs significantly from the original CSI.

[0127] For example, by combining different deployment methods of the model and the relationship between the model and the function, the terminal and network device can be applied to the following 6 scenarios when deploying the model or function.

[0128] Scenario 1: A model is jointly deployed on the terminal side and the network side, and this model can implement at least one function. For example, the model can implement one or more of function 1, function 2, and function 3; wherein, the terminal is used to execute and call function 1 in the model, and the network device is used to execute and call function 2 and / or function 3 in the model.

[0129] Scenario 2: A model is deployed on the network side, which can perform at least one function. For example, the model can perform function 1 and / or function 2; wherein, the terminal can report the measurement information required by the model to the network device, and the network device can invoke and execute function 1 and / or function 2 in the model.

[0130] Scenario 3: A model is deployed on the terminal side, which can implement at least one function. For example, the model can implement function 1 and / or function 2; wherein, the terminal can obtain measurement information and invoke and execute function 1 and / or function 2 in the model.

[0131] Scenario 4: A function is jointly deployed on the terminal side and the network side, and this function is implemented through at least one model. For example, the function is implemented through one or more of Model 1, Model 2, and Model 3; wherein the terminal is used to execute and call Model 1, and the network device is used to execute and call Model 2 and / or Model 3.

[0132] Scenario 5: A function is deployed on the network side, which is implemented through at least one model. For example, the function is implemented through model 1 and / or model 2; wherein, the terminal can report the measurement information required by the model to the network device. The network device is used to execute the call to model 1 and / or model 2.

[0133] Scenario 6: A function is deployed on the terminal side, which is implemented through at least one model. For example, the function is implemented through model 1 and / or model 2; wherein, the terminal can obtain measurement information and invoke model 1 and / or model 2 to implement the function.

[0134] In some embodiments, model delivery can be triggered by a network device or by a terminal.

[0135] For example, as shown in Figure 4, the model transfer method can be implemented through the following steps for a terminal-side model in a one-sided model deployment or a terminal-side model in a two-sided model deployment.

[0136] Step 401: The terminal sends model instruction information to the network device. Correspondingly, the network device receives the model instruction information from the terminal.

[0137] The model indication information is used to indicate which models the terminal can support, is available, or is applicable. The model indication information can be a model support indication, a model availability indication, or a model applicability indication.

[0138] For example, the supported model indication information can be supported model indication information.

[0139] Step 402: The network device selects a model based on the model indication information.

[0140] For example, a network device can select models that the terminal can support, use, or apply from the model metadata.

[0141] Step 403: The network device sends model configuration information to the terminal. Correspondingly, the terminal receives the model configuration information from the network device.

[0142] The model configuration information includes the configuration parameters of the selected AI / ML model.

[0143] For example, the model configuration information can be RRC configuration information or RRC reconfiguration information. Afterwards, the terminal and network devices can perform model transmission / transfer operations.

[0144] Step 404: The terminal sends a model transmission / delivery wait indication message to the network device. Correspondingly, the network device receives the model transmission / delivery wait indication message from the terminal.

[0145] For example, the model transmission / delivery waiting indication information can be an RRC configuration complete message or an RRC reconfiguration complete message. Step 404 is an optional step, meaning that step 404 can be omitted, and the terminal can choose not to send the model transmission / delivery waiting indication information.

[0146] Step 405: The terminal and network devices perform model transmission / transfer operations.

[0147] For example, if the model configured on the terminal is not available, the required model can be obtained through a model transfer / transfer operation.

[0148] Step 406: The terminal sends a model transmission / transmission completion indication message to the network device. Correspondingly, the network device receives the model transmission / transmission completion indication message from the terminal.

[0149] For example, the model transmission / transmission completion indication information can be an RRC configuration complete message or an RRC reconfiguration complete message. Step 406 is optional; that is, step 406 can be omitted, and the terminal may not send the model transmission / transmission completion indication information.

[0150] In some embodiments, the model transfer process involved in this application is applicable to various communication systems. Taking the O-RAN system as an example, as shown in Figure 5, when the network device is a node with separate CU and DU, the network device includes: CU-CP, CU-machine learning plane (MLP) and DU. Taking the terminal requesting the transmission of the model to the network device as an example, for the terminal-side model in a one-sided model deployment or the terminal-side model in a two-sided model deployment, the model transfer method can be implemented through the following steps.

[0151] Step 501: The terminal sends terminal capability information to the CU-CP. Correspondingly, the CU-CP receives the terminal capability information from the terminal.

[0152] As an example, terminal capability information is: UE capability information.

[0153] Step 502: CU-CP determines whether to use the ML function.

[0154] As an example, ML functionality can also be referred to as ML functionality.

[0155] Step 503: If the CU-CP determines that the ML function is to be used, the CU-CP sends a Terminal Context Setting Request message to the CU-MLP. Correspondingly, the CU-MLP receives the Terminal Context Setting Request message from the CU-CP.

[0156] As an example, the terminal context setup request message is: UE context setup request.

[0157] Step 504: CU-MLP sends a model setup request message to DU. Correspondingly, DU receives the model setup request message from CU-MLP.

[0158] As an example, the model setup request message is: ML model setup request.

[0159] Step 505: DU sends a model setup response message to CU-MLP. Correspondingly, CU-MLP receives the model setup response message from DU.

[0160] As an example, the model setup response message is: UE context setup response.

[0161] Step 506: CU-MLP sends a terminal context setting response message to CU-CP. Correspondingly, CU-CP receives the terminal context setting response message from CU-MLP.

[0162] As an example, the terminal context setup response message is: UE context setup response.

[0163] Step 507: The CU-CP sends an RRC reconfiguration request message to the terminal. Correspondingly, the terminal receives the RRC reconfiguration request message from the CU-CP.

[0164] Optionally, the RRC reconfiguration request may include an ML configuration message.

[0165] As an example, the RRC reconfiguration request message is: RRC reconfiguration request.

[0166] The ML configuration message is: ML configuration information.

[0167] Step 508: The terminal sends an RRC reconfiguration complete message to the CU-CP. Correspondingly, the CU-CP receives the reconfiguration complete message from the terminal.

[0168] As an example, the RRC reconfiguration complete message is: RRC reconfiguration complete.

[0169] Step 509: The terminal downloads the ML model. In some embodiments, if the access node stores the ML model data, the access node can transmit the model data to the terminal; if the cloud server stores the ML model data, the cloud server can transmit the model data to the terminal through the access node.

[0170] Step 510: The terminal sends an ML model usage preparation message to the CU-CP. Correspondingly, the CU-CP receives the ML model usage preparation message from the terminal.

[0171] As an example, the ML model uses the preparation message: ML model ready for use.

[0172] Step 511: CU-CP sends an ML model usage preparation message to CU-MLP. Correspondingly, CU-MLP receives the ML model usage preparation message from CU-CP.

[0173] Step 512: CU-CP sends an ML model activation message to the terminal. Correspondingly, the terminal receives the ML model activation message from CU-CP.

[0174] As an example, the ML model activation message is: Activate ML model.

[0175] It should be noted that the system described in the embodiments of this application is for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and does not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0176] The communication method provided in this application will be described below with reference to the communication system shown in Figure 1, taking the interaction between the first node and the second node as an example. It should be noted that in the following embodiments of this application, the message names, parameter names, or information names between the first node and the second node are just examples, and other names may be used in other embodiments. The method provided in this application does not specifically limit these names.

[0177] It is understood that in the embodiments of this application, each communication device (including the first node or the second node) may execute some or all of the steps in the embodiments of this application. These steps or operations are merely examples, and the embodiments of this application may also execute other operations or variations thereof. Furthermore, the steps may be executed in different orders as presented in the embodiments of this application, and it is not necessary to execute all the operations in the embodiments of this application.

[0178] It is understood that this application uses the first node and the second node as examples to illustrate the execution of the interaction, but this application does not limit the execution entities of the interaction. For example, the method executed by the first node in this application can also be executed by a module applied to the first node (e.g., a chip, chip system, or processor), or by a logic node, logic module, or software capable of implementing all or part of the functions of the first node. Or, for example, the method executed by the second node in this application can also be executed by a module applied to the second node (e.g., a chip, chip system, or processor), or by a logic node, logic module, or software capable of implementing all or part of the functions of the second node.

[0179] The communication method provided in the embodiments of this application will be described below. As shown in FIG6, the communication method may include the following steps:

[0180] Step 601: The first node obtains the first information.

[0181] For example, the first node can be a terminal in the aforementioned communication system, or a chip or circuit that can be used in the terminal, or an entity associated with the terminal, etc. Alternatively, the first node can also be a network device in the aforementioned communication system, such as an access node or a core network node, and the second node can also be a chip or circuit in the network device, or an entity associated with the network device, etc.

[0182] For example, the first information can be generated by the first node based on data obtained from its own monitoring, or it can be obtained by the first node from a third party. The first model can be a new model generated through model training, or it can be an enhancement of the current model; this application embodiment does not specifically limit this.

[0183] The first model is used to enable AI / ML functions, and the model delivery performance metrics are used to characterize the model delivery performance of the first model. The model delivery performance of the first model is also called the model delivery cost-effectiveness of the first model, which can be used to evaluate the relationship between the communication resources used for model delivery and the expected benefits of the model.

[0184] In some embodiments, the model transmission performance index information can also be determined separately for each cell. That is, when the same first node is in different cells, the corresponding model transmission performance index information is calculated and statistically analyzed separately.

[0185] In this embodiment, when the model transmission performance of the model deployed on the terminal side in a unilateral deployment is poor, the second node needs to consume a large amount of overhead during model transmission, and it is difficult to ensure that the model can generate high returns. When the model transmission performance of the model deployed on both sides is poor, in addition to the above problems, the network device will also waste resources such as AI computing power and power consumption because the expected returns of the model cannot be guaranteed.

[0186] In this embodiment, model transmission incurs overhead. Frequent transmission of models with poor transmission performance wastes air interface resources and increases the complexity of terminal receiving and storing models. The transmission performance of the first model is affected by both the communication resources occupied by model transmission and the expected benefits of the model. The communication resources occupied by model transmission are determined by the data size of the model itself and the number of times the model is transmitted in actual use. The expected benefits of the model are determined by the model's usage effect and the duration of model usage.

[0187] When the first node has good storage conditions (e.g., ample available storage space), it can store the transmitted model data for a longer period, enabling data reuse and reducing the number of model transmissions. Different models are suitable for different scenarios, and the adaptability between the scenario and the model determines the model's effectiveness and actual performance. Therefore, the higher the scene recognition accuracy, the more suitable the selected first model is for the actual scenario, and the better its actual performance. Furthermore, the generalization ability of the first model determines the duration of its adaptability; better generalization means a longer duration of continuous use. In summary, the transmission performance of the first model is related to the storage conditions of the first node, the scene recognition accuracy, and the generalization ability of the first model.

[0188] In some embodiments, the model transmission performance index information of the first model can be determined by an entity that determines model performance. For example, when the first model is deployed on a terminal, the model performance of the first model is determined by the terminal or by a network device. In this case, the model transmission performance index information of the first model is determined by the terminal.

[0189] When the first model is deployed on a network device, its performance is determined by the network device or the terminal. In this case, the performance metrics of the first model are determined by the network device.

[0190] When the first model is deployed on both the terminal and network devices, its performance can be determined by the terminal, the network devices, or both. In this case, the performance metrics of the first model can also be determined by the terminal, the network devices, or both.

[0191] Furthermore, the statistical entity for model performance metrics information can be either a terminal or a network device. For example, terminal statistics can be tailored to various models from specific terminal manufacturers, making the statistical results more robust; network devices can combine model performance metrics information from terminals from different manufacturers, making them suitable for more general scenarios. When the model performance determination entity and the statistical entity for model performance metrics information are different entities, the model performance determination entity needs to pass the model performance to the statistical entity for model performance metrics information in order to generate the model performance metrics information.

[0192] It should be understood that the above descriptions of the model can also be applied to functions, characteristics, or algorithms. The model can be expressed in different ways in different scenarios.

[0193] Step 602: The first node sends the first information to the second node. Correspondingly, the second node receives the first information from the first node.

[0194] For example, the second node can be a network device in the aforementioned communication system, such as an access node or a core network node. The second node can also be a chip or circuit within the network device, or an entity associated with the network device. Alternatively, the first node can be a terminal in the aforementioned communication system, or a chip or circuit that can be used in the terminal, or an entity associated with the terminal.

[0195] In some embodiments, the first node may report first information using capability information, or the first node model may report first information when the second node registers, or the first node may report first information when reporting supported models. The first information may be carried in proprietary information or in an RRC message, for example, in an RRC message carrying the capability information or UE assistance information (UAI) of the carrying terminal device.

[0196] The relevant information regarding the first piece of information can be found in step 601 above, and will not be repeated here.

[0197] Step 603: The second node determines the model transfer strategy based on the first information.

[0198] In some embodiments, the second node can determine a management strategy based on the first information and current network conditions. For example, network conditions include information such as channel environment, communication quality, or operational load. Exemplarily, the second node can combine the model's expected or historical returns with the cost of model propagation to determine whether and how to trigger model propagation (e.g., determining the quantization method of the propagated model). For instance, models expected to be used more frequently after propagation have a higher priority for propagation, while models expected to be used less frequently after propagation have a lower priority.

[0199] For example, when the second node is a network device in the aforementioned communication system, the model transfer strategy can be for the second node to directly send model data, or it can be to request the peer to send model data and wait for the peer's response. When the second node is a terminal in the aforementioned communication system, the model transfer strategy can be to request the terminal to send model data.

[0200] Currently, it is difficult for the network side to assess the expected performance of the terminal side after model transmission. This can lead to network devices' model transmission strategies failing to guarantee high expected model returns. For example, network devices may be limited by communication resources and only be able to transmit a portion of the models. However, if the transmitted models do not achieve good model returns, the model transmission strategy developed by the network devices will fail to achieve good model returns. Furthermore, the excessive communication resources consumed can affect the normal transmission of other models. Therefore, in this application, the first node can send relevant information representing the model transmission performance indicators of the first model to the second node. This allows the second node to fully consider the model transmission performance of the first model (also known as the cost-effectiveness of model transmission, the return value of model transmission, the target performance of model transmission, the stability of model transmission, or the robustness of model transmission) when formulating its model transmission strategy. Thus, the second node can rationally plan the communication resources required for model transmission based on the relevant information of the model transmission performance indicators, avoiding waste of communication resources while achieving higher model returns.

[0201] As one possible implementation, the model can convey performance metric information in a variety of ways.

[0202] In some embodiments, the model transfer performance metric information includes at least one model transfer performance metric, and the at least one model transfer performance metric includes at least one model transfer performance metric value and / or at least one first value range.

[0203] In this embodiment of the application, the model transfer performance index value refers to the specific numerical value corresponding to the model transfer performance index. For example, the model transfer performance of the first model is greater than 0.9.

[0204] Model transfer performance metrics are typically used to characterize the minimum acceptable value for model transfer performance. Similarly, the first value range refers to the range of values ​​corresponding to the model transfer performance metric. The model transfer performance metric and model transfer performance can be positively or negatively correlated. For example, a higher value indicates better model transfer performance, or vice versa. Since model transfer performance is affected by changes in the channel environment, adjustments to communication configuration, or external physical interference, the first node can choose to report model transfer performance information in the form of a single value or a value range, allowing the second node to consider the impact of various factors when determining the model transfer strategy and ensuring the normal use of the model.

[0205] For example, at least one model transfer performance metric value includes at least two model transfer performance metric values. At least one first value range includes at least two first value ranges. In this way, the first information can include multiple values / value ranges of the model transfer performance metric, so that the second node has more choices when determining the model transfer strategy.

[0206] In some embodiments, the model transfer performance metric information includes the value of the model transfer performance metric and the metric type.

[0207] The value can be a specific numerical value or a range of values. Indicator types can be represented by fields, identifiers, or mapping tables. For example, the first node can directly report the field information of the indicator type, or report an identifier that has a mapping relationship with the indicator type. The second node can determine the indicator type indicated by the identifier through a mapping table.

[0208] For example, when the first node needs to send multiple model transfer performance metrics of the first model to the second node, the model transfer performance metric information may include the values ​​of multiple model transfer performance metrics and the metric types corresponding to the multiple model transfer performance metrics, so that the second node can determine each model transfer performance metric and its corresponding value.

[0209] In some embodiments, model transfer performance metric information includes models that meet the model transfer performance metrics or the identifier of a model.

[0210] For example, the first node can report the identifier of the model that meets the model transfer performance index according to the instructions of the second node, and the second node can directly determine the corresponding model transfer strategy based on the reported model identifier.

[0211] As one possible embodiment, the model transfer performance metrics information includes mathematical statistics used to characterize the model transfer performance of the first model.

[0212] In some embodiments, model transfer performance information includes mathematical statistics for characterizing the model transfer performance of the first model.

[0213] This mathematical statistics can also be called probability statistics, such as proportion, mean / variance / standard deviation, or the cumulative distribution function (CDF) / probability mass function (PMF) / probability density function (PDF) of model transfer performance, etc.

[0214] For example, the mathematical statistics of model transfer performance can be:

[0215] 1. The probability that the model transferability is <0.7 is 10%;

[0216] 2. The probability that the model transferability is between 0.7 and 0.8 is 20%.

[0217] 3. The probability that the model transferability is between 0.8 and 0.9 is 60%.

[0218] 4. The probability that the model transferability is >0.9 is 10%.

[0219] In other words, model transfer performance can satisfy a specific probability distribution, and the model transfer performance corresponding to different intervals has corresponding mathematical statistics. Furthermore, besides probabilistic representation, mathematical statistics can also be expressed as mean / variance / standard deviation, or x%CDF / PMF / PDF, where x is a non-negative number, which will not be elaborated further. Thus, this application can determine the corresponding model transfer performance index information by performing statistical analysis on the model transfer performance index, making the model transfer performance index information more accurately reflect the model transfer performance of the first model.

[0220] In some embodiments, the first information includes at least one model transfer performance index, and one or more of the at least one model transfer performance index have corresponding data features or data feature identifiers.

[0221] In some embodiments, at least one model transmits performance metric information that corresponds one-to-one with at least one data feature or data feature identifier.

[0222] In some embodiments, at least one model delivery performance metric corresponds one-to-one with at least one data feature or data feature identifier.

[0223] For example, the first model corresponds to at least one model that transmits performance index information, and the performance index information transmitted by the at least one model includes the performance index information transmitted by the first model, which corresponds to the first data feature or the first data feature identifier.

[0224] For example, at least one model transfer performance index information also includes second model transfer performance index information, which corresponds to a second data feature or a second data feature identifier.

[0225] For example, at least one model transfer performance index information includes model transfer performance index information 1 and model transfer performance index information 2, and at least one data feature includes data feature A and data feature B, wherein model transfer performance index information 1 corresponds to data feature A, and model transfer performance index information 2 corresponds to data feature B.

[0226] For example, at least one model transfer performance index information includes model transfer performance index information 1, model transfer performance index information 2, and model transfer performance index information 3, and at least one data feature identifier includes data feature identifier A, data feature identifier B, and data feature identifier C, wherein model transfer performance index information 1 corresponds to data feature identifier A, model transfer performance index information 2 corresponds to data feature C, and model transfer performance index information 3 corresponds to data feature B.

[0227] In other words, the data features or data feature identifiers corresponding to the performance metrics information transmitted by different models can be the same or different. At least one model's performance metrics correspond to different data features or data feature identifiers. Alternatively, at least one model's performance metrics information may contain some model-transmitted performance metrics that do not have corresponding data features or data feature identifiers.

[0228] In this application's embodiments, "feature" can be replaced with: condition, situation, environment, type, status, data, or data distribution. Similarly, "feature" in other terms such as "xx feature" (e.g., "reference feature," "physical feature," or "feature of the xth data") in this application's embodiments can be replaced similarly. Optionally, "feature" can also be called "data feature" or "data classification feature."

[0229] Optionally, in this embodiment, the aforementioned data feature identifier can be referred to as associated ID, data ID, dataset ID, data categorization ID, or property ID. The data feature identifier can be used by the terminal to classify measurements, where the measurements are used to determine model inputs or training data used for model training (e.g., model output, labels). It can also be used for inference-related operations; for example, the terminal can assume that downlink transmission beam sets / lists have the same or similar data features under the same data feature identifier.

[0230] In this embodiment, features can be used to represent the characteristics of the signals, channels, or associated data (e.g., data that can be used to train the first model or data that can be used for inference / processing of the first model) that are received / transmitted due to one or more of the following physical factors: device configuration or environmental factors, such as node deployment, antenna configuration, and transmit / receive waveforms. Different nodes may possess different features, which may affect model training and / or inference. Therefore, in this embodiment, model transmission performance indicators can be bound to data features or data feature identifiers so that the first and second nodes can consider the impact of device configuration or environmental factors when performing related operations.

[0231] In some embodiments, the first node may also send data features or data feature identifiers to the second node. The first node may add the data features or data feature identifiers to first information and send them to the second node through the first information. For example, the first information may also include data features or data feature identifiers. The first node may also send data features or data feature identifiers to the second node independently.

[0232] In some embodiments, data features are used to characterize the features possessed by the first data, which is the data corresponding to the first model.

[0233] The data corresponding to the first model refers to the data used to train the first model, or the data that can be used for inference / processing of the first model.

[0234] The data in this application embodiment includes data acquired from a communication network, such as signal processing information, channel information, and radio frequency (RF) information. Signal processing information includes information generated during baseband signal processing, such as information generated during signal sampling, modulation, demodulation, encoding, decoding, precoding, resource mapping, and / or digital filtering. Channel information includes information corresponding to the channel environment, such as at least one or a combination of at least two of the following: power information, amplitude information, phase information, delay information, multipath information, signal propagation time information, distance information, velocity information, large-scale channel information, small-scale channel information, channel scattering information, and LOS / NLOS information. For example, data in CSI and BM use cases is represented as channel information (CSI), while data in positioning use cases is represented as channel information and / or location information. Radio frequency (RF) information includes information generated during analog processing, such as digital-to-analog conversion, analog-to-digital conversion, digital predistortion, frequency conversion, RF modulation, RF demodulation, power amplification, low-noise amplification, analog filtering, and / or duplex processing.

[0235] Taking CSI feedback or beam management as an example, the first data includes CSI information. Optionally, the CSI information includes at least one of the following: channel response information, channel quality information, and beam information.

[0236] The channel response information can be at least one of the following: precoding matrix indicator (PMI), precoding matrix, precoding vector, eigenvector, eigenma, channel vector, channel matrix, layer information (LI), and rank information (RI). The channel quality information can be at least one of RSRP, SINR, and channel quality indicator (CQI). The beam information can be at least one of CRI and synchronization signal block resource indicator (SSBRI). Beam information can also be referred to as reference signal identification information.

[0237] As one possible implementation, the model transfer performance index information is determined by state information and / or model transfer information, whereby the state information includes first valid state information and / or first failed state information.

[0238] In some embodiments, the model transfer information includes first model transfer quantization information, which includes the number of model transfers and / or the model transfer time.

[0239] Here, "model transfer count" refers to the number of times the first model is transferred, and "model transfer time" refers to the length of time the first model is transferred. When the first model is transferred multiple times, the model transfer time can be the sum of the time lengths of each transfer. Thus, the state information can characterize the expected return of the first model, and the model transfer information can characterize the cost of the first model's transfer. Therefore, the model transfer performance index information can be determined through the state information and / or the model transfer information.

[0240] For example, when model transfer performance metrics are determined by state information, a higher expected return of the first model as represented by the state information indicates better model transfer performance; conversely, a lower expected return indicates worse model transfer performance. When model transfer performance metrics are determined by model transfer information, a lower transfer cost (e.g., fewer transfers) indicates better model transfer performance; conversely, a higher transfer cost indicates worse model transfer performance. When model transfer performance metrics are determined by both state information and model transfer information, a comprehensive evaluation of the expected return and transfer cost of the first model can be conducted, thus more accurately reflecting the model transfer performance.

[0241] In some embodiments, the first valid state information corresponds to the valid state of the first model, and the valid state includes at least one of the following:

[0242] The state in which the performance of the first model reaches or exceeds the first target performance index;

[0243] The activation state of the first model; or,

[0244] The availability / applicability / usability of the first model;

[0245] And / or, the first failure state information corresponds to the failure state of the first model, and the failure state includes at least one of the following:

[0246] The performance of the first model did not reach or exceed the second target performance index;

[0247] The inactive state of the first model; or,

[0248] The first model is in a state of being unsupportable / unapplicable / unavailable.

[0249] In other words, the first valid state information is used to characterize the information of the first model in the valid state, and the first valid state information is used to characterize the information of the first model in the failed state.

[0250] Optionally, in this embodiment, the target performance may correspond to the inference performance, prediction performance, or monitoring performance of the first model. Inference performance or prediction performance refers to the operational performance of the first model when it is invoked (or activated / configured / enabled / triggered / executed / runs), and monitoring performance refers to the monitoring performance of the first model when it is invoked.

[0251] In some embodiments, when the target performance corresponds to the inference performance or prediction performance of the first model, the target performance can be the system performance of the terminal when calling the first model, such as parameters used to characterize system performance, such as throughput, RSRP, signal-to-noise ratio (SNR) / signal-to-interference plus noise ratio (SINR), or block error rate (BLER).

[0252] Optionally, in this embodiment, the target performance can also be the performance of the first model during runtime, such as runtime accuracy. For example, for a model used for CSI compression feedback, the target performance can be the accuracy of restoring (decompressing) the compressed CSI. For a model used for CSI prediction, the target performance can be the accuracy of CSI prediction. For a model used for beam prediction, the target performance can be the accuracy of beam prediction. For a model used for location inference, the target performance can be the accuracy of location inference. Furthermore, the first model can also be one or a combination of multiple models selected from modulation, demodulation, encoding, decoding, channel equalization, channel estimation, pilot generation, precoding, resource mapping, resource demapping, interference suppression, interference estimation, interference prediction, receiver, or transmitter, etc., and will not be listed exhaustively.

[0253] It should be understood that accuracy here is also called intermediate key performance indicator (KPI), or prediction accuracy information. Accuracy is used to characterize the similarity between the model's inference results and the corresponding true values, such as generalized cosine similarity (GCS), squared generalized cosine similarity (SGCS), normalized mean square error (NMSE), beam information prediction accuracy, reference signal label prediction accuracy, RSRP prediction accuracy, or location information prediction accuracy, etc. Here, the true value corresponds to the measurement result obtained from the measurement. For example, the CSI restoration accuracy or CSI prediction accuracy of the aforementioned CSI compressed feedback can be represented by SGCS or NMSE, and the beam prediction accuracy can be represented by the comparison information (e.g., confusion matrix) between the prediction result and the true value or label of the RSRP / channel state information-reference signal (CSI-RS) resource indicator (CRI). The accuracy of location inference can be represented by comparing the predicted results with the actual location coordinates or location labels.

[0254] Optionally, in this embodiment, the target performance can also be a relative performance compared to the performance of a baseline model. For example, in this application, the performance of an algorithm for a particular model can be used as the baseline performance, which can be represented as described above, and the target performance can be a relative performance compared to this baseline performance.

[0255] In some embodiments, when the target performance corresponds to the monitoring performance of the first model, the target performance can be the monitoring accuracy of the first model. For example, for a model used for CSI compressed feedback, the target performance can be the monitoring accuracy when monitoring the performance of the CSI compressed feedback model. For a model used for CSI prediction, the target performance can be the monitoring accuracy when monitoring the performance of the CSI prediction model. For instance, the monitoring accuracy can be obtained by comparing the SGCS value corresponding to the performance monitoring result of the CSI prediction model with the SGCS value corresponding to the model's actual performance result (e.g., obtaining the difference between the two SGCS values).

[0256] Optionally, in this embodiment, the first target performance index and the second target performance index may be the same or different. The performance of the first model reaching or exceeding the first target performance index includes the performance of the first model reaching or exceeding the value of the first target performance index, or falling within the value range of the first target performance index. The performance of the first model not reaching or exceeding the second target performance index includes the performance of the first model not reaching or exceeding the value of the second target performance index, or falling outside the value range of the second target performance index.

[0257] Since the performance of the first model is often affected by various factors during actual use, the embodiments of this application can be divided into multiple intervals according to the value or value range. For example, the performance interval where the performance of the first model reaches or exceeds the value of the first target performance index is the interval corresponding to the effective state, or the performance interval within the value range of the first target performance index is the interval corresponding to the effective state; the performance interval where the performance of the first model does not reach or exceed the value of the second target performance index is the interval corresponding to the failed state, or the performance interval within the value range of the second target performance index is the interval corresponding to the failed state. In this way, this application can determine the current state of the first model based on the interval in which the performance of the first model falls.

[0258] Optionally, in this embodiment, the activated state refers to the model being called a configured model, an enabled model, or a model that triggers execution / running. The model being in an activated state means that the device, based on the model input, can obtain the corresponding output results by running the model within a specific time period.

[0259] Optionally, in this embodiment, a "supported" state refers to the device's ability to run a specific model. For example, if a terminal reports a supported model, it means that when the terminal obtains the model, it can run the model; that is, based on the model input, it can obtain the corresponding output results by running the model within a specific time period.

[0260] Optionally, in this embodiment, the available state means that the device can obtain and run the model. For example, the terminal locally stores model-A and can run model-A; or, for another example, the terminal can obtain model-A from a network device or a cloud server and can run model-A.

[0261] Optionally, in this embodiment, the applicable state refers to the model being ready for activation / inference / prediction. For example, if model-A is applicable to scenario-X, then model-A is in the applicable state when the current scenario is scenario-X. The inactive state and the unsupportable / unapplicable / unavailable state are the opposite states of the above-mentioned active state and supportable / applicable / available state, and will not be described further.

[0262] In some embodiments, the first valid state information includes the first valid quantization information of the first model, which includes valid time information and / or valid count information, wherein the valid time information includes the length of time the first model is in a valid state, and the valid count information includes the number of times the first model is in a valid state;

[0263] And / or, the first failure state information includes the first failure quantification information of the first model, the first failure quantification information includes failure time information and / or failure count information, wherein the failure time information includes the length of time the first model is in a failure state, and the failure count information includes the number of times the first model is in a failure state.

[0264] Thus, this application can quantify valid state information and failed state information according to different dimensions, such as time dimension and statistical frequency dimension, so as to evaluate the model transmission performance index information of the first model.

[0265] Optionally, in this embodiment, the duration of the first model being in a valid state and the duration of its invalid state can be determined based on one or more configured monitoring resources. For example, the first node can monitor one or more monitoring resources configured within time period T1 to determine whether the first model is in a valid state. If it is in a valid state, the node records the duration of the first model being in a valid state plus T1. Conversely, if it is in a invalid state, the first node records the duration of the first model being in a invalid state plus T1.

[0266] Optionally, in this embodiment, the number of times the first model is in a valid state and the number of times it is in a invalid state can be determined based on one or more configured monitoring resources. For example, the first node can monitor one or more configured monitoring resources to determine whether the first model is in a valid state. If it is in a valid state, the number of times the first model is in a valid state is incremented by one. Correspondingly, if it is in a invalid state, the first node increments the number of times the first model is in a invalid state by one.

[0267] In some embodiments, the model transmits performance metric information including at least one of the following: first proportional information, first difference information, and first probability statistics. The model transmits performance metric information through proportional information, difference information, and probability statistics.

[0268] For example, the first ratio information includes: the ratio of the first effective quantification information to the model transfer information, or the ratio of the first failure quantification information to the model transfer information, or the ratio of the model transfer information to the first effective quantification information, or the ratio of the model transfer information to the first failure quantification information. For instance, the first ratio information can be expressed as the ratio of effective time information to model transfer time, the ratio of effective time information to the number of model transfers, the ratio of effective number of transfers information to model transfer time, the ratio of effective number of transfers information to the number of model transfers, the ratio of failure time information to model transfer time, the ratio of failure time information to the number of model transfers, the ratio of failure number of transfers information to model transfer time, or the ratio of failure number of transfers information to the number of model transfers.

[0269] In other words, the model transfer performance index information in this application embodiment can be calculated based on various combinations of the first effective quantization information and / or the first failure quantization information, and the model transfer information. The model transfer performance index information and the model transfer performance of the first model can be positively correlated or negatively correlated. For example, the larger the model transfer performance index information, the better the model transfer performance of the first model; the smaller the model transfer performance index information, the worse the model transfer performance of the first model. Alternatively, the larger the model transfer performance index information, the worse the model transfer performance of the first model; the smaller the model transfer performance index information, the better the model transfer performance of the first model.

[0270] In some embodiments, when the target performance metric is omitted, the effective time information can be the duration for which the model is in an active state, the duration for which the model is in a supportable state, the duration for which the model is in a usable state, or the duration for which the model is in a usable state. When the target performance metric exists, the effective time information can be the duration for which the performance of the first model reaches or exceeds the first target performance metric.

[0271] Taking the ratio of effective time information to model delivery counts as an example, effective time information can be one or more of the following: the duration the model is in an active state, the duration the model is in a supportable state, the duration the model is in a usable state, or the duration the model is in a usable state. Effective time information can be represented as T0. The number of model delivery counts can be represented as N0. The above durations can be a continuous time period or a combination of multiple continuous time periods. Model delivery performance metrics information. At this point, the model transfer performance index information is negatively correlated with the model transfer performance of the first model.

[0272] Taking the ratio of effective time information to model transmission counts as an example, effective time information can be represented as T0, and model transmission counts can be represented as N0. Model transmission performance metrics information... At this point, the model transfer performance index is positively correlated with the model transfer performance of the first model.

[0273] As shown in Figure 7, the second node transmits the first model to the first node A and the first node B, and instructs the first node A and the first node B to activate the first model, respectively. The effective time information of the first model at the first node A is... The effective time information of the first model of the first node B is: The model propagation count for the first node A is 2, and the model propagation count for the first node B is 1. Therefore, the model propagation performance index information for the first node A is... The model of the first node B transmits performance index information. As shown in Figure 7, the model transfer performance index of the first node A is greater than that of the first node B, and the model transfer performance of the first model in the first node A is worse than that of the first model in the first node B.

[0274] In some embodiments, the effective time information can be the duration for which the performance of the first model reaches or exceeds the first target performance indicator, and the effective count information can be the number of times the performance of the first model reaches or exceeds the first target performance indicator. The failure time information can be the duration for which the performance of the first model does not reach or exceed the second target performance indicator, and the failure count information can be the number of times the performance of the first model does not reach or exceed the second target performance indicator. Compared to the scheme in the above embodiments that uses activation time as the effective time, the method of using the target performance indicator as the effective time has finer granularity, thus determining more accurate model performance indicator information.

[0275] Optionally, when special values ​​exist in the aforementioned model-transmitted performance index information, model-transmitted information, first effective state information, or first failure state information, such as 0 or ∞, these special values ​​can be represented by predefined methods such as special fields. Alternatively, when the values ​​of the aforementioned model-transmitted performance index information, model-transmitted information, first effective state information, or first failure state information exceed a preset threshold, such as when the value of the model-transmitted performance index information is too small, it can be represented by a default field. The model-transmitted performance index information can also be represented by multiple value ranges, such as 0 to 0.1, 0.1 to 0.01, 0.01 to 0.001, and less than 0.001. The model-transmitted performance index information can also be represented explicitly or implicitly, such as directly displaying calculated values ​​or implicitly representing them through the relationship between calculated model-transmitted performance index information and preset model-transmitted performance index information.

[0276] In some embodiments, when the first model corresponds to multiple target performance indicators, each target performance indicator can correspond to a separate model transmission performance indicator, or a third target performance indicator can be determined based on the multiple target performance indicators, and the model transmission performance indicator information can be determined based on the third target performance indicator. The method for determining the model transmission performance indicator information, where each corresponds to a separate model transmission performance indicator, can refer to the method described above.

[0277] In some embodiments, when the first model corresponds to one or more quantization methods, each quantization method corresponds to a model that transmits performance index information.

[0278] Model quantization refers to converting the character type of parameters in the model data into the target character type. Different character types correspond to different numerical precisions and different bit occupancy. In other words, the quantization method of a model can affect its performance and data size.

[0279] For example, character types include single-precision floating-point numbers (FP32), half-precision floating-point numbers (FP16), 8-bit integers (INT8), and 4-bit integers (INT4). FP32 represents a model parameter using 32 bits, including 1 sign bit, 8 exponent bits, and 23 decimal bits. FP16 represents a model parameter using 16 bits, including 1 sign bit, 5 exponent bits, and 10 decimal bits. The higher the numerical precision of the character type, the smaller the performance loss of the model, and the higher the corresponding model performance. However, the more bits the model parameters occupy, and the larger the model data becomes.

[0280] Besides classifying by character type, quantization methods can also be classified in other ways, such as by different algorithms or by the execution stage of the quantization operation. This application does not limit this classification.

[0281] Optionally, in this embodiment, the first node may report model transmission performance index information corresponding to each quantization method. The method for determining the model transmission performance index information can be found in the aforementioned related content and will not be elaborated upon further.

[0282] Optionally, in this embodiment of the application, the model transfer performance index information of the first model is determined by the model transfer performance index information corresponding to one or more quantization methods respectively.

[0283] Taking the ratio of model transfer counts to effective time information as an example, with quantization method Q-1 (e.g., INT8), the effective time information is T1, and the number of model transfer counts is N1. With quantization method Q-1 (e.g., FP32), the effective time information is T2, and the number of model transfer counts is N2. Model transfer performance metrics information. Among them, p and q are predefined by the protocol, or determined by negotiation between the terminal and the network device.

[0284] Taking the ratio of model transmission time to effective time information as an example, as shown in Figure 8, the second node transmits the first model to the first node A according to quantization method Q-1 and instructs the first node A to activate the first model. Then, the second node transmits the first model to the first node A according to quantization method Q-2 and instructs the first node A to activate the first model. The model transmission time for the second node to transmit the first model to the first node A according to quantization method Q-1 is... The second node transmits the first model to the first node A according to the quantization method Q-2. The model transmission time is... The effective time information for the first model of the first node A in quantization method Q-1 is: The effective time information for the first model at node A in quantization method Q-2 is: At this point, the model transmits performance metric information. At this point, when quantization method Q-2 is assigned a larger weight, it has a greater impact on the model's performance metrics corresponding to quantization method Q-2. For quantization methods that require special attention, a higher weight can be assigned, meaning that the quantization method will contribute more to the model's performance.

[0285] In some embodiments, the first valid status information is first valid status information determined within at least one first time period; and / or, the first failure status information is first failure status information determined within at least one second time period.

[0286] The first time period and the second time period can be the same time period or different time periods. For different time periods, the lengths of the first time period and the second time period can be the same or different. Since the model transmission performance of the first model is affected by changes in the channel environment, adjustments to communication configuration, or external physical interference, its model transmission performance changes over time. Therefore, in this embodiment, the first valid state information and the first failed state information can be statistically analyzed in time segments to further determine the model transmission performance indicators of the first model at different times.

[0287] In some embodiments, at least one first time period includes at least one of the following:

[0288] For at least one time period, performance monitoring is performed on the first model;

[0289] The first model is in an active state for at least one time period;

[0290] The first model was in a supportable / applicable / available state for at least one period of time;

[0291] The second node indicates at least one time period;

[0292] Alternatively, at least one time period determined by the first node;

[0293] And / or, at least one second time period includes at least one of the following:

[0294] For at least one time period, performance monitoring is performed on the first model;

[0295] The first model is in an active state for at least one time period;

[0296] The first model was in a supportable / applicable / available state for at least one period of time;

[0297] The second node indicates at least one time period;

[0298] Alternatively, at least one time period determined by the first node.

[0299] The aforementioned time period can be predefined or determined through negotiation between the first and second nodes, such as through RRC configuration. For example, when the time period is defaulted, it can be represented by -∞ and the time period can be counted accordingly.

[0300] In some embodiments, at least one first time period is included within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, and at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the first time unit for transmitting performance indicator information by the model; and / or, at least one second time period is included within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, and at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the second time unit for transmitting performance indicator information by the model.

[0301] The activation time window can be a period of time during which the first model is in an active state. Activation can also be referred to as being configured or enabled. The applicable time window can be a period of time during which the first model is in an applicable state, and the available time window can be a period of time during which the first model is in an available state.

[0302] In this embodiment of the application, the first time period may not include an activation time window, a monitoring time window, an applicable time window, or an available time window. For example, the first time unit may not be located within a monitoring time window, or the first time unit may not be located within an activation time window, or the first time unit may not be located within an applicable time window, or the first time unit may not be located within an available time window. And / or, the second time unit may not be located within a monitoring time window, or the second time unit may not be located within an activation time window, or the second time unit may not be located within an applicable time window, or the second time unit may not be located within an available time window.

[0303] In this embodiment, the first time period can be earlier than the first time unit for sending model transmission performance indicator information, and the second time period can be earlier than the second time unit for sending model transmission performance indicator information. For example, if the first model has never been activated / monitored / is in an applicable state / is in an available state, the model transmission performance indicator information can be reported according to the default value.

[0304] As shown in Figure 9, the first time period includes a monitoring time window n and an activation time window M. The first valid state information is the first valid state information determined from the monitoring time window n and the activation time window M included in this first time period. The first node reports the model transmission performance index information after the activation time window M. The first failure state information is not described in detail.

[0305] For example, the first time unit is located within the monitoring time window, or the first time unit is located within the activation time window. The second time unit is located within the monitoring time window, or the second time unit is located within the activation time window.

[0306] As shown in Figure 10, the first time period includes a monitoring time window n, a monitoring time window n+1, and an activation time window M. The first valid state information is the first valid state information determined from the monitoring time window n and the activation time window M included in this first time period. The first node reports the model transmission performance index information within the monitoring time window n+1. The first failure state information is not described in detail.

[0307] The overall process of the communication method provided in this application has been described above. The following section describes the calculation, reporting, and statistics of the first information in the case of a terminal, access node, and core network node / OAM.

[0308] As shown in Figure 11, the communication method includes the following steps:

[0309] Step 1101: The terminal and access node determine the calculation method for the model transmission performance index information.

[0310] In some embodiments, step 1101 is an optional step. For example, the calculation method for model transfer performance index information between the terminal and the access node can be implemented in a predefined manner. For instance, the calculation method for model transfer performance index information with multiple quantization methods can be determined based on the weight corresponding to each quantization method.

[0311] Step 1102: The access node sends monitoring instruction information to the terminal. Correspondingly, the terminal receives the monitoring instruction information from the access node.

[0312] The monitoring instruction information is used to instruct the terminal to report model transmission performance metrics. For example, the monitoring instruction information can also instruct the monitoring resource; for instance, the monitoring instruction information may include a time period T during which the access node instructs the terminal to monitor the model transmission performance. This monitoring resource can also be indicated in other ways, and this application does not limit its scope.

[0313] In some embodiments, step 1102 is an optional step. For example, the terminal reports model transmission performance index information using capability information; or, the terminal reports model transmission performance index information when registering its model with the network device; or, the terminal reports model transmission performance index information when reporting supported models. The model transmission performance index information may be carried in proprietary information or in an RRC message, such as in the capability information of the carrying terminal device or the RRC message of the UAI.

[0314] Step 1103: The terminal and / or access node perform performance monitoring and calculate the performance index information transmitted by the model.

[0315] Optionally, if the model transmission performance metrics are determined by the terminal, the terminal may also send the model transmission performance metrics to the access node. If the model transmission performance metrics are determined by the access node, the access node may also send the model transmission performance metrics to the terminal.

[0316] For example, performance indicator information transmitted by the model can be reported proactively by the terminal, or requested to be reported by the terminal by the network device, or proactively sent to the terminal by the network device, or sent to the terminal by the network device in response to the terminal's request.

[0317] In some embodiments, the interaction of model transfer performance metrics information can occur either when requesting model transfer or before requesting model transfer. For example, a terminal may request model transfer from a network device while simultaneously reporting the model transfer performance metrics information of the requested model; or, the network device may instruct the terminal to report the model transfer performance metrics information of the requested model. As another example, the network device may request model transfer to the terminal while simultaneously instructing the terminal to provide the model transfer performance metrics information for that model; or the terminal may request the network device to instruct the terminal to provide the model transfer performance metrics information for that model.

[0318] In some embodiments, the interaction of model transmission performance metrics information can occur after a request to transmit the model has been made. For example, after a terminal requests the transmission of a model from a network device, the terminal reports the model transmission performance metrics information for that model; or, the network device instructs the terminal to report the model transmission performance metrics information for that model. As another example, after a network device requests the transmission of a model to a terminal, the network device instructs the terminal to transmit the model transmission performance metrics information for that model; or the terminal requests the network device to instruct the terminal to transmit the model transmission performance metrics information for that model.

[0319] Step 1104: The access node and / or terminal instruct the core network node / OAM to update the model transmission performance index information of the first model.

[0320] In other words, model transmission performance metrics can be generated by the terminal or by the access node; these metrics can be statistically analyzed by the terminal or mathematically analyzed by the access node. For example, when the first model is in a valid state, if the access node triggers the deactivation of the first model, the network device can independently analyze the model transmission performance metrics at this time and report them to the core network node or OAM. Similarly, when the terminal triggers the deactivation of the first model while it is in a valid state, the terminal can analyze the model transmission performance metrics at this time, characterizing the model's generalization ability to the terminal's internal conditions.

[0321] In some embodiments, model delivery performance metrics can be model attributes on the terminal side, which is suitable for scenarios where the terminal's model delivery performance is decoupled from the network-side vendor. Alternatively, model delivery performance can be statistically analyzed separately for each cell global identifier (CGI). The model delivery performance metrics reported by the terminal can be the model delivery performance of a single model, or the model delivery performance of models provided by the terminal vendor to which the terminal belongs.

[0322] In some embodiments, when the time period T for monitoring model delivery performance is periodic monitoring, the terminal can retain the most recent model delivery performance (starting from T or multiple T), and update the historical model delivery performance (starting from -∞).

[0323] The method provided in this application has been described above. In addition, this application also provides a communication device for implementing the functions described in the above method embodiments.

[0324] It is understood that, in order to achieve the aforementioned functions, the communication device includes hardware structures and / or software modules corresponding to the execution of each function. Those skilled in the art should readily recognize that, based on the units and algorithm steps of the examples described in conjunction with the embodiments disclosed herein, this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed in hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0325] This application embodiment can divide the communication device into functional modules according to the above method embodiment. For example, each function can be divided into a separate functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. It should be noted that the module division in this application embodiment is illustrative and only represents one logical functional division. In actual implementation, there may be other division methods.

[0326] Figure 12 shows a schematic diagram of a communication device 120. The communication device 120 includes a processing module 1201 and a transceiver module 1202. The communication device 120 can be used to implement the functions of the first node or the second node described above.

[0327] In some embodiments, the communication device 120 may further include a storage module (not shown in FIG12) for storing program instructions and data.

[0328] In some embodiments, the transceiver module 1202, also referred to as a transceiver unit, is used to implement sending and / or receiving functions. The transceiver module 1202 may consist of a transceiver circuit, a transceiver, a transceiver unit, or a communication interface.

[0329] In some embodiments, the transceiver module 1202 may include a receiving module and a sending module, respectively configured to perform the receiving and sending steps performed by the first node or the second node in the above method embodiments, and / or other processes to support the technology described herein; the processing module 1201 may be configured to perform the processing steps performed by the first node or the second node in the above method embodiments, and / or other processes to support the technology described herein.

[0330] When the communication device 120 is used to implement the function of the first node:

[0331] The processing module 1201 is used to obtain first information through the transceiver module 1202; the first information includes model transmission performance index information of the first model; wherein, the first model is used to enable artificial intelligence (AI) / machine learning (ML) functions; the model transmission performance index information is used to characterize the model transmission performance of the first model; the transceiver module 1202 is used to send the first information to the second node.

[0332] In one possible design, the model transfer performance information includes at least one model transfer performance index, which includes at least one model transfer performance index value and / or at least one first value range.

[0333] In one possible design, at least one model transfer performance metric value includes at least two model transfer performance metric values. At least one first value range includes at least two first value ranges.

[0334] In one possible design, the model transfer performance metrics include mathematical statistics used to characterize the model transfer performance of the first model.

[0335] In one possible design, the model transfer performance information includes mathematical statistics of the model transfer performance metrics used to characterize the first model.

[0336] In one possible design, the first information includes at least one model transfer performance index; one or more of the model transfer performance indexes have corresponding data features or data feature identifiers.

[0337] In one possible design, at least one model conveys performance metric information corresponding to at least one data feature or data feature identifier.

[0338] In one possible design, at least one model delivery performance metric corresponds one-to-one with at least one data feature or data feature identifier.

[0339] In one possible design, the first model corresponds to at least one model that transmits performance index information, and the at least one model that transmits performance index information includes the first model that transmits performance index information, which corresponds to a first data feature or a first data feature identifier.

[0340] In one possible design, at least one model transmits performance index information, which also includes a second model transmits performance index information, the second model transmits performance index information corresponding to a second data feature or a second data feature identifier.

[0341] In one possible design, the model transfer performance information is determined by state information and / or model transfer information; the state information includes first valid state information and / or first failure state information.

[0342] In one possible design, the model transfer information includes first model transfer quantization information; the first model transfer quantization information includes the number of model transfers and / or the model transfer time.

[0343] In one possible design, the first valid state information corresponds to the valid state of the first model, which includes at least one of the following: the performance of the first model reaches or exceeds the first target performance index; the first model is active; or the first model is supportable / applicable / usable; and / or, the first failure state information corresponds to the failure state of the first model, which includes at least one of the following: the performance of the first model does not reach or exceed the second target performance index; the first model is inactive; or the first model is not supportable / applicable / usable.

[0344] In one possible design, the performance of the first model reaching or exceeding the first target performance index includes the performance of the first model reaching or exceeding the value of the first target performance index, or being within the range of the value of the first target performance index; the performance of the first model not reaching or exceeding the second target performance index includes the performance of the first model not reaching or exceeding the value of the second target performance index, or not being within the range of the value of the second target performance index.

[0345] In one possible design, the first valid state information includes the first valid quantization information of the first model, which includes valid time information and / or valid count information, wherein the valid time information includes the length of time the first model is in a valid state, and the valid count information includes the number of times the first model is in a valid state; and / or, the first failure state information includes the first failure quantization information of the first model, which includes failure time information and / or failure count information, wherein the failure time information includes the length of time the first model is in a failed state, and the failure count information includes the number of times the first model is in a failed state.

[0346] In one possible design, the model transmits performance metric information including at least one of the following: first proportional information, first difference information, and first probability statistics.

[0347] In one possible design, the first ratio information includes: the ratio of the first effective quantization information to the model transfer information, or the ratio of the first failure quantization information to the model transfer information, or the ratio of the model transfer information to the first effective quantization information, or the ratio of the model transfer information to the first failure quantization information.

[0348] In one possible design, the first valid state information is first valid state information determined within at least one first time period; and / or, the first failure state information is first failure state information determined within at least one second time period.

[0349] In one possible design, at least one first time period includes at least one of the following: at least one time period for performance monitoring of the first model; at least one time period for the first model to be in an active state; at least one time period for the first model to be in a supportable / applicable / available state; at least one time period indicated by a second node; or at least one time period determined by a first node; and / or, at least one second time period includes at least one of the following: at least one time period for performance monitoring of the first model; at least one time period for the first model to be in an active state; at least one time period for the first model to be in a supportable / applicable / available state; at least one time period indicated by a second node; or at least one time period determined by a first node.

[0350] In one possible design, at least one first time period is contained within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, wherein at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the first time unit for transmitting performance metrics of the sending model; and / or, at least one second time period is contained within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, wherein at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the second time unit for transmitting performance metrics of the sending model.

[0351] In one possible design, the first time unit is not located within the monitoring time window, or the first time unit is not located within the activation time window, or the first time unit is not located within the applicable time window, or the first time unit is not located within the available time window; and / or, the second time unit is not located within the monitoring time window, or the second time unit is not located within the activation time window, or the second time unit is not located within the applicable time window, or the second time unit is not located within the available time window.

[0352] When the communication device 120 is used to implement the function of the second node:

[0353] The transceiver module 1202 is used to receive first information from the first node; the first information includes model transmission performance index information of the first model; wherein, the first model is used to enable artificial intelligence (AI) / machine learning (ML) functions; the model transmission performance index information is used to characterize the model transmission performance of the first model; the processing module 1201 is used to determine the model transmission strategy based on the first information.

[0354] In one possible design, the model transfer performance information includes at least one model transfer performance index, which includes at least one model transfer performance index value and / or at least one first value range.

[0355] In one possible design, at least one model transfer performance metric value includes at least two model transfer performance metric values. At least one first value range includes at least two first value ranges.

[0356] In one possible design, the model transfer performance metrics include mathematical statistics used to characterize the model transfer performance of the first model.

[0357] In one possible design, the model transfer performance information includes mathematical statistics of the model transfer performance metrics used to characterize the first model.

[0358] In one possible design, the first information includes at least one model transfer performance index; one or more of the model transfer performance indexes have corresponding data features or data feature identifiers.

[0359] In one possible design, at least one model conveys performance metric information corresponding to at least one data feature or data feature identifier.

[0360] In one possible design, at least one model delivery performance metric corresponds one-to-one with at least one data feature or data feature identifier.

[0361] In one possible design, the first model corresponds to at least one model that transmits performance index information, and the at least one model that transmits performance index information includes the first model that transmits performance index information, which corresponds to a first data feature or a first data feature identifier.

[0362] In one possible design, at least one model transmits performance index information, which also includes a second model transmits performance index information, the second model transmits performance index information corresponding to a second data feature or a second data feature identifier.

[0363] In one possible design, the model transfer performance information is determined by state information and / or model transfer information; the state information includes first valid state information and / or first failure state information.

[0364] In one possible design, the model transfer information includes first model transfer quantization information; the first model transfer quantization information includes the number of model transfers and / or the model transfer time.

[0365] In one possible design, the first valid state information corresponds to the valid state of the first model, which includes at least one of the following: the performance of the first model reaches or exceeds the first target performance index; the first model is active; or the first model is supportable / applicable / usable; and / or, the first failure state information corresponds to the failure state of the first model, which includes at least one of the following: the performance of the first model does not reach or exceed the second target performance index; the first model is inactive; or the first model is not supportable / applicable / usable.

[0366] In one possible design, the performance of the first model reaching or exceeding the first target performance index includes the performance of the first model reaching or exceeding the value of the first target performance index, or being within the range of the value of the first target performance index; the performance of the first model not reaching or exceeding the second target performance index includes the performance of the first model not reaching or exceeding the value of the second target performance index, or not being within the range of the value of the second target performance index.

[0367] In one possible design, the first valid state information includes the first valid quantization information of the first model, which includes valid time information and / or valid count information, wherein the valid time information includes the length of time the first model is in a valid state, and the valid count information includes the number of times the first model is in a valid state; and / or, the first failure state information includes the first failure quantization information of the first model, which includes failure time information and / or failure count information, wherein the failure time information includes the length of time the first model is in a failed state, and the failure count information includes the number of times the first model is in a failed state.

[0368] In one possible design, the model transmits performance metric information including at least one of the following: first proportional information, first difference information, and first probability statistics.

[0369] In one possible design, the first ratio information includes: the ratio of the first effective quantization information to the model transfer information, or the ratio of the first failure quantization information to the model transfer information, or the ratio of the model transfer information to the first effective quantization information, or the ratio of the model transfer information to the first failure quantization information.

[0370] In one possible design, the first valid state information is first valid state information determined within at least one first time period; and / or, the first failure state information is first failure state information determined within at least one second time period.

[0371] In one possible design, at least one first time period includes at least one of the following: at least one time period for performance monitoring of the first model; at least one time period for the first model to be in an active state; at least one time period for the first model to be in a supportable / applicable / available state; at least one time period indicated by a second node; or at least one time period determined by a first node; and / or, at least one second time period includes at least one of the following: at least one time period for performance monitoring of the first model; at least one time period for the first model to be in an active state; at least one time period for the first model to be in a supportable / applicable / available state; at least one time period indicated by a second node; or at least one time period determined by a first node.

[0372] In one possible design, at least one first time period is contained within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, wherein at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the first time unit for transmitting performance metrics of the sending model; and / or, at least one second time period is contained within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, wherein at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the second time unit for transmitting performance metrics of the sending model.

[0373] In one possible design, the first time unit is not located within the monitoring time window, or the first time unit is not located within the activation time window, or the first time unit is not located within the applicable time window, or the first time unit is not located within the available time window; and / or, the second time unit is not located within the monitoring time window, or the second time unit is not located within the activation time window, or the second time unit is not located within the applicable time window, or the second time unit is not located within the available time window.

[0374] All relevant content of each step involved in the above method embodiments can be referenced from the functional description of the corresponding functional module, and will not be repeated here.

[0375] In this application, the communication device 120 can be presented in an integrated manner by dividing it into various functional modules. Here, "module" can refer to an application-specific integrated circuit (ASIC), a circuit, a processor and memory that executes one or more software or firmware programs, integrated logic circuits, and / or other devices that can provide the above functions.

[0376] In some embodiments, when the communication device 120 in FIG12 is a chip or chip system, the function / implementation process of the transceiver module 1202 can be implemented through the input / output interface (or communication interface) of the chip or chip system, and the function / implementation process of the processing module 1201 can be implemented through the processor (or processing circuit) of the chip or chip system.

[0377] Since the communication device 120 provided in this embodiment can execute the above method, the technical effects it can achieve can be referred to the above method embodiment, and will not be repeated here.

[0378] As a possible product form, the first node or the second node described in the embodiments of this application can be implemented using the following: one or more field programmable gate arrays (FPGAs), programmable logic devices (PLDs), controllers, state machines, gate logic, discrete hardware components, any other suitable circuits, or any combination of circuits capable of performing the various functions described throughout this application.

[0379] As another possible product form, the first node or second node described in this application embodiment can be implemented using a general bus architecture. For ease of explanation, refer to FIG13, which is a schematic diagram of the structure of a communication device 1300 provided in an embodiment of this application. The communication device 1300 includes a processor 1301 and a transceiver 1302. The communication device 1300 can be a first node, or a chip or chip system therein; or, the communication device 1300 can be a second node, or a chip or module therein. FIG13 only shows the main components of the communication device 1300. In addition to the processor 1301 and transceiver 1302, the communication device may further include a memory 1303 and input / output devices (not shown in FIG13).

[0380] Optionally, the processor 1301 is mainly used to process communication protocols and communication data, control the entire communication device, execute software programs, and process the data of the software programs, thereby implementing the methods provided in the above-described method embodiments. The memory 1303 is mainly used to store software programs and data. The transceiver 1302 may include a radio frequency (RF) circuit and an antenna. The RF circuit is mainly used for converting baseband signals to RF signals and processing RF signals. The antenna is mainly used for transmitting and receiving RF signals in the form of electromagnetic waves. Input / output devices, such as touch screens, displays, and keyboards, are mainly used to receive user input data and output data to the user.

[0381] Optionally, the processor 1301, transceiver 1302, and memory 1303 can be connected via a communication bus.

[0382] When the communication device is powered on, the processor 1301 can read the software program in the memory 1303, execute the instructions of the software program, and process the data of the software program. When data needs to be transmitted wirelessly, the processor 1301 performs baseband processing on the data to be transmitted and outputs the baseband signal to the radio frequency (RF) circuit. The RF circuit then performs RF processing on the baseband signal and transmits the RF signal outward in the form of electromagnetic waves through the antenna. When data is sent to the communication device, the RF circuit receives the RF signal through the antenna, converts the RF signal into a baseband signal, and outputs the baseband signal to the processor 1301. The processor 1301 converts the baseband signal into data and processes the data.

[0383] In another implementation, the radio frequency circuitry and antenna can be set up independently of the processor performing baseband processing. For example, in a distributed scenario, the radio frequency circuitry and antenna can be arranged remotely, independent of the communication device.

[0384] In some embodiments, those skilled in the art will recognize that the above-described communication device 120 can take the form of the communication device 1300 shown in FIG13 in terms of hardware implementation.

[0385] As an example, the function / implementation of the processing module 1201 in Figure 12 can be achieved by the processor 1301 in the communication device 1300 shown in Figure 13 calling computer execution instructions stored in the memory 1303. The function / implementation of the transceiver module 1202 in Figure 12 can be achieved by the transceiver 1302 in the communication device 1300 shown in Figure 13.

[0386] As another possible product form, the first node or the second node in this application may adopt the composition structure shown in FIG14, or include the components shown in FIG14. FIG14 is a schematic diagram of the composition of a communication device 1400 provided in this application. The communication device 1400 may be a first node or a chip or system-on-a-chip in the first node; or, it may be a second node or a chip or system-on-a-chip in the second node.

[0387] As shown in FIG14, the communication device 1400 includes at least one processor 1401 and at least one communication interface (FIG14 is merely an example illustrating the inclusion of a communication interface 1404 and a processor 1401). Optionally, the communication device 1400 may further include a communication bus 1402 and a memory 1403.

[0388] Processor 1401 can be a general-purpose central processing unit (CPU), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a PLD, or any combination thereof. Processor 1401 can also be other devices with processing functions, such as circuits, devices, or software modules, without limitation.

[0389] Communication bus 1402 is used to connect different components in communication device 1400, enabling communication between them. Communication bus 1402 can be a peripheral component interconnect (PCI) bus or an extended industry standard architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of illustration, only one thick line is used in Figure 14, but this does not indicate that there is only one bus or one type of bus.

[0390] Communication interface 1404 is used for communicating with other devices or communication networks. Exemplarily, communication interface 1404 can be a module, circuit, transceiver, or any device capable of communication. Optionally, the communication interface 1404 can also be an input / output interface located within processor 1401, used to implement signal input and signal output for the processor.

[0391] The memory 1403 may be a device with storage function, used to store instructions and / or data. The instructions may be computer programs.

[0392] For example, the memory 1403 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and / or instructions; it may also be a random access memory (RAM) or other type of dynamic storage device capable of storing information and / or instructions; it may also be an electrically erasable programmable read-only memory (EEPROM), a compact disc read-only memory (CD-ROM) or other optical disc storage, optical disc storage (including compressed optical discs, laser discs, optical discs, digital universal optical discs, Blu-ray discs, etc.), magnetic disk storage media or other magnetic storage devices, etc., without limitation.

[0393] It should be noted that the memory 1403 can exist independently of the processor 1401, or it can be integrated with the processor 1401. The memory 1403 can be located inside or outside the communication device 1400, without limitation. The processor 1401 can be used to execute the instructions stored in the memory 1403 to implement the methods provided in the following embodiments of this application.

[0394] Optionally, the processor 1401 and / or memory 1403 may include an artificial intelligence (AI) module, which is used to implement AI-related functions. The AI ​​module can be implemented through software, hardware, or a combination of both. For example, the AI ​​module may include a radio network intelligent controller (RIC) module. For example, the AI ​​module can be a near real-time RIC or a non-real-time RIC.

[0395] As an optional implementation, the communication device 1400 may also include an output device 1405 and an input device 1406. The output device 1405 communicates with the processor 1401 and can display information in various ways. For example, the output device 1405 may be a liquid crystal display (LCD), a light-emitting diode (LED) display device, a cathode ray tube (CRT) display device, or a projector, etc. The input device 1406 communicates with the processor 1401 and can receive user input in various ways. For example, the input device 1406 may be a mouse, keyboard, touchscreen device, or sensing device, etc.

[0396] In some embodiments, those skilled in the art will recognize that the communication device 120 shown in FIG12 can take the form of the communication device 1400 shown in FIG14 in terms of hardware implementation.

[0397] As an example, the function / implementation process of the processing module 1201 in Figure 12 can be implemented by the processor 1401 in the communication device 1400 shown in Figure 14 calling computer execution instructions stored in the memory 1403. The function / implementation process of the transceiver module 1202 in Figure 12 can be implemented by the communication interface 1404 in the communication device 1400 shown in Figure 14.

[0398] It should be noted that the structure shown in Figure 14 does not constitute a specific limitation on the first node or the second node. For example, in other embodiments of this application, the first node or the second node may include more or fewer components than shown in the figure, or combine some components, or split some components, or have different component arrangements. The components shown in the figure may be implemented in hardware, software, or a combination of software and hardware.

[0399] In some embodiments, this application also provides a communication device, which includes a processor for implementing the methods in any of the above method embodiments.

[0400] As one possible implementation, the communication device also includes a memory. This memory stores necessary computer programs and data. The computer program may include instructions, which a processor can invoke to instruct the communication device to execute the methods described in any of the above method embodiments. Alternatively, the memory may not be present in the communication device.

[0401] As another possible implementation, the communication device also includes an interface circuit, which is a code / data read / write interface circuit, used to receive computer execution instructions (which are stored in memory and may be read directly from memory or may be transmitted through other devices) and transmit them to the processor.

[0402] As another possible implementation, the communication device also includes a communication interface for communicating with modules outside the communication device.

[0403] It is understood that the communication device can be a chip or a chip system. When the communication device is a chip system, it can be composed of chips or may include chips and other discrete devices. This application does not specifically limit this.

[0404] This application also provides a computer-readable storage medium having a computer program or instructions stored thereon, which, when executed by a computer, implements the functions of any of the above-described method embodiments.

[0405] This application also provides a computer program product that, when executed by a computer, implements the functions of any of the above method embodiments.

[0406] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0407] It is understood that the systems, apparatuses, and methods described in this application can also be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative. For instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the couplings or direct couplings or communication connections shown or discussed may be through some interfaces; indirect couplings or communication connections between devices or units may be electrical, mechanical, or other forms.

[0408] The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. The components shown as units may or may not be physical units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0409] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0410] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented using software programs, implementation can be, in whole or in part, in the form of a computer program product. This computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another via wired (e.g., coaxial cable, fiber optic, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium can be any available medium accessible to a computer or a data storage device containing one or more servers, data centers, etc., that can be integrated with the medium. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid-state drive (SSD)). In embodiments of this application, the computer may include the aforementioned apparatus.

[0411] Although this application has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the accompanying drawings, disclosure, and appended claims, will understand and implement other variations of the disclosed embodiments in carrying out the claimed application. In the claims, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. A single processor or other unit can implement several functions listed in the claims. While different dependent claims may recite certain measures, this does not mean that these measures cannot be combined to produce good results.

[0412] Although this application has been described in conjunction with specific features and embodiments, it is obvious that various modifications and combinations can be made thereto without departing from the scope of this application. Accordingly, this specification and drawings are merely illustrative descriptions of the application as defined by the appended claims, and are considered to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Thus, if such modifications and modifications fall within the scope of the claims and their equivalents, this application is also intended to include such modifications and modifications.

Claims

1. A communication method, characterized in that, Applied to the first node, the method includes: Obtain first information; the first information includes model transfer performance index information of a first model; wherein, the first model is used to enable artificial intelligence (AI) / machine learning (ML) functions; the model transfer performance index information is used to characterize the model transfer performance of the first model; Send the first information to the second node.

2. A communication method, characterized in that, Applied to the second node, the method includes: Receive first information from a first node; the first information includes model delivery performance index information of a first model; wherein, the first model is used to enable artificial intelligence (AI) / machine learning (ML) functions; the model delivery performance index information is used to characterize the model delivery performance of the first model; The model transfer strategy is determined based on the first information.

3. The method according to claim 1 or 2, characterized in that, The model transfer performance index information includes at least one model transfer performance index, and the at least one model transfer performance index includes at least one model transfer performance index value and / or at least one first value range.

4. The method according to any one of claims 1-3, characterized in that, The model transfer performance metrics include mathematical statistics used to characterize the model transfer performance of the first model.

5. The method according to any one of claims 1-4, characterized in that, The first information includes at least one model transfer performance index; one or more of the at least one model transfer performance index have corresponding data features or data feature identifiers.

6. The method according to any one of claims 1-5, characterized in that, The model transmits performance index information, which is determined by state information and / or model transmit information; the state information includes first valid state information and / or first failure state information.

7. The method according to claim 6, characterized in that, The model transmission information includes first model transmission quantization information; the first model transmission quantization information includes the number of model transmissions and / or the model transmission time.

8. The method according to claim 6 or 7, characterized in that, The first valid state information corresponds to the valid state of the first model, and the valid state includes at least one of the following: The state in which the performance of the first model reaches or exceeds the first target performance index; The activation state of the first model; or, The supportability / applicability / availability status of the first model; And / or, the first failure state information corresponds to the failure state of the first model, and the failure state includes at least one of the following: The first model's performance has not reached or exceeded the second target performance index; The inactive state of the first model; or, The first model is in a state of being unsupported / unapplicable / unavailable.

9. The method according to any one of claims 6-8, characterized in that, The first valid state information includes the first valid quantization information of the first model, which includes valid time information and / or valid count information, wherein the valid time information includes the length of time the first model is in a valid state, and the valid count information includes the number of times the first model is in a valid state; And / or, the first failure state information includes the first failure quantification information of the first model, the first failure quantification information includes failure time information and / or failure count information, wherein the failure time information includes the length of time the first model is in a failure state, and the failure count information includes the number of times the first model is in a failure state.

10. The method according to any one of claims 6-9, characterized in that, The first valid state information is at least one first valid state information determined within a first time period; And / or, the first failure status information is a first failure status information determined within at least one second time period.

11. The method according to claim 10, characterized in that, The at least one first time period includes at least one of the following: At least one time period for which the performance of the first model is monitored; The first model is in an active state for at least one time period; The first model is in a supportable / applicable / available state for at least one time period; The second node indicates at least one time period; Alternatively, at least one time period determined by the first node; And / or, the at least one second time period includes at least one of the following: At least one time period for which the performance of the first model is monitored; The first model is in an active state for at least one time period; The first model is in a supportable / applicable / available state for at least one time period; The second node indicates at least one time period; Alternatively, at least one time period determined by the first node.

12. The method according to claim 10 or 11, characterized in that, The at least one first time period is contained within at least one monitoring time window, at least one activation time window, at least one applicable time window, or at least one available time window, and the at least one monitoring time window, at least one activation time window, at least one applicable time window, or at least one available time window is earlier than the first time unit for sending the model to transmit performance index information; And / or, the at least one second time period is contained within at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window, wherein the at least one monitoring time window, at least one active time window, at least one applicable time window, or at least one available time window is earlier than the second time unit in which the model transmits performance indicator information.

13. A communication device, characterized in that, include: A functional unit for performing the method as described in any one of claims 1-12; wherein the action performed by the functional unit is implemented by hardware or by hardware executing corresponding software.

14. A communication device, characterized in that, The communication device includes a processor; the processor is configured to run a computer program or instructions to cause the communication device to perform the method as described in any one of claims 1-12.

15. A computer-readable storage medium, characterized in that, A computer-readable storage medium stores computer instructions or programs that, when executed on a computer, cause the method described in any one of claims 1-12 to be performed.

16. A computer program product, characterized in that, The computer program product includes computer instructions; when some or all of the computer instructions are run on a computer, the method described in any one of claims 1-12 is performed.

Images