WO2023097522A1

WO2023097522A1 - Model updating method and apparatus for wireless channel processing, device, and medium

Info

Publication number: WO2023097522A1
Application number: PCT/CN2021/134661
Authority: WO
Inventors: 田文强
Original assignee: Oppo广东移动通信有限公司
Priority date: 2021-11-30
Filing date: 2021-11-30
Publication date: 2023-06-08
Also published as: CN117897917A

Abstract

The present application relates to the field of communications, and discloses a model updating method and apparatus for wireless channel processing, a device, and a medium. The method comprises: a first wireless communication device updates a local machine learning model on the basis of local data; and the first wireless communication device transmits an updated local machine learning model to a second wireless communication device, the updated local machine learning model being configured to update a global machine learning model. By updating the local machine learning model by the wireless communication device in a distributed mode and updating the global machine learning model, the machine learning model can be updated by using the local data of different wireless communication devices. The training data of training the machine learning model can be enriched, so that the machine learning model adapts to different application scenes. Therefore, the problem of scene adaptation when the machine learning model processes a wireless communication problem is solved, and the collection of a large amount of data is avoided. The efficiency and accuracy of updating the machine learning model are improved.

Description

Model updating method, device, equipment and medium for wireless channel processing

technical field

The present application relates to the communication field, and in particular to a model update method, device, equipment and medium for wireless channel processing.

Background technique

Solutions based on AI (Artificial Intelligence, artificial intelligence) are more and more applied in wireless communication systems. For example, machine learning models are used to solve channel state information (Channel State Information, CSI) feedback problems, channel estimation problems, positioning problems, and beam management problems.

In AI-based solutions, the training data of machine learning models is strongly related to wireless communication scenarios, environments, and channels. AI-based solutions need to meet the requirements of adapting to different scenarios and universally applicable to multiple scenarios. In related technologies, different machine learning models are constructed for different scenarios, environments, and channels, so that corresponding machine learning models can be used in specific situations to solve corresponding wireless communication problems. For example, different CSI feedback models, different channel estimation models, different positioning models, and different beam management models are constructed in different cells, indoors and outdoors, and in cities and suburbs.

In the process of training the machine learning model based on the above method, it is necessary to obtain various data in different scenarios, environments, and channels in advance, but it is difficult to collect complete data for various scenarios, environments, and channels, and The collection process is also more cumbersome. Further discussion and research is needed on the training update of machine learning models.

Contents of the invention

Embodiments of the present application provide a model update method, device, device and medium for wireless channel processing. Described technical scheme is as follows:

According to an aspect of the present application, a method for updating a model for wireless channel processing is provided, the method is executed by a first wireless communication device, and the method includes:

updating a local machine learning model based on local data by the first wireless communication device;

The first wireless communication device sends the updated local machine learning model to the second wireless communication device, and the updated local machine learning model is used to update the global machine learning model.

According to another aspect of the present application, a method for updating a model for wireless channel processing is provided, the method is executed by a second wireless communication device, and the method includes:

The second wireless communication device receives an updated local machine learning model sent by the first wireless communication device, and the updated local machine learning model is obtained by updating the local machine learning model based on local data by the first wireless communication device of;

The second wireless communication device updates a global machine learning model according to the updated local machine learning model.

According to another aspect of the present application, a model updating device for wireless channel processing is provided, the device comprising:

An update module for updating a local machine learning model based on local data;

A sending module, configured to send the updated local machine learning model to the second wireless communication device, and the updated local machine learning model is used to update the global machine learning model.

A receiving module, configured to receive an updated local machine learning model sent by the first wireless communication device, where the updated local machine learning model is obtained by updating the local machine learning model based on local data by the first wireless communication device;

An update module, configured to update the global machine learning model according to the updated local machine learning model.

According to another aspect of the present application, a terminal is provided, and the terminal includes: a processor; a transceiver connected to the processor; a memory for storing executable instructions of the processor; wherein, the The processor is configured to load and execute the executable instructions to implement the model update method for wireless channel processing as described in the above aspect.

According to another aspect of the present application, a network device is provided, and the network device includes: a processor; a transceiver connected to the processor; a memory for storing executable instructions of the processor; wherein, The processor is configured to load and execute the executable instructions to implement the model update method for wireless channel processing as described in the above aspects.

According to another aspect of the present application, a computer-readable storage medium is provided, wherein executable instructions are stored in the readable storage medium, and the executable instructions are loaded and executed by the processor to implement the above-mentioned aspects. The described model update method for wireless channel processing.

According to another aspect of the present application, a chip is provided, the chip includes a programmable logic circuit and/or program instructions, and when the chip is run on a computer device, it is used to implement the wireless Model update method for channel processing.

According to another aspect of the present application, there is provided a computer program product or computer program, said computer program product or computer program comprising computer instructions stored in a computer-readable storage medium, a processor from said computer The readable storage medium reads and executes the computer instructions, so that the computer device executes the model update method for wireless channel processing described in the above aspects.

The technical solution provided by the application at least includes the following beneficial effects:

The first wireless communication device updates the local machine learning model in a distributed manner to update the global machine learning model, so that local data of different wireless communication devices can be used to update the machine learning model. The local data of different wireless communication devices are generated under different scenarios, environments and channels, so it can enrich the training data for training machine learning models and make machine learning models adapt to different application scenarios. Therefore, it is possible to avoid a large amount of data collection work while solving the scene adaptation problem when the machine learning model handles the wireless communication problem. Improve the efficiency and accuracy of updating machine learning models.

Description of drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings that need to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those skilled in the art, other drawings can also be obtained based on these drawings without creative effort.

FIG. 1 is a schematic diagram of a process for solving a CSI feedback problem provided by an exemplary embodiment of the present application;

FIG. 2 is a schematic diagram of a process for solving a channel estimation problem provided by an exemplary embodiment of the present application;

Fig. 3 is a schematic diagram of a process of solving a positioning problem provided by an exemplary embodiment of the present application;

FIG. 4 is a schematic diagram of a process for solving a beam management problem provided by an exemplary embodiment of the present application;

FIG. 5 is a schematic diagram of a federated learning process provided by an exemplary implementation of the present application;

FIG. 6 is a schematic diagram of a system architecture of a communication system provided by an exemplary embodiment of the present application;

FIG. 7 is a schematic diagram of a process of updating a model for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 8 is a schematic diagram of the process of uploading and downloading models provided by an exemplary embodiment of the present application;

FIG. 9 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

FIG. 10 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 11 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 12 is a schematic diagram of the second wireless communication device determining the duration of the first wireless communication device as a valid user according to an exemplary embodiment of the present application;

Fig. 13 is a schematic diagram of the second wireless communication device determining the duration of the first wireless communication device as a valid user provided by an exemplary embodiment of the present application;

Fig. 14 is a schematic diagram of the second wireless communication device determining the duration of the first wireless communication device as a valid user according to an exemplary embodiment of the present application;

Fig. 15 is a schematic diagram of the process of sending and updating a machine learning model provided by an exemplary embodiment of the present application;

Fig. 16 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 17 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application;

Fig. 18 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 19 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application;

Fig. 20 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 21 is a schematic diagram of the process of sending and updating a machine learning model provided by an exemplary embodiment of the present application;

Fig. 22 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 23 is a schematic diagram of the process of sending and updating a machine learning model provided by an exemplary embodiment of the present application;

Fig. 24 is a flowchart of a model updating method for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 25 is a schematic diagram of the process of sending and updating a machine learning model provided by an exemplary embodiment of the present application;

Fig. 26 is a block diagram of a model updating device for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 27 is a block diagram of a model updating device for wireless channel processing provided by an exemplary embodiment of the present application;

Fig. 28 is a schematic structural diagram of a communication device provided by an exemplary embodiment of the present application.

Detailed ways

In order to make the purpose, technical solution and advantages of the present application clearer, the implementation manners of the present application will be further described in detail below in conjunction with the accompanying drawings.

Reference will now be made in detail to the exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numerals in different drawings refer to the same or similar elements unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the present invention. Rather, they are merely examples of apparatuses and methods consistent with aspects of the invention as recited in the appended claims.

The terminology used in the present disclosure is for the purpose of describing particular embodiments only, and is not intended to limit the present disclosure. As used in this disclosure and the appended claims, the singular forms "a", "the", and "the" are intended to include the plural forms as well, unless the context clearly dictates otherwise. It should also be understood that the term "and/or" as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

It should be understood that although the terms first, second, third, etc. may be used in the present disclosure to describe various information, the information should not be limited to these terms. These terms are only used to distinguish information of the same type from one another. For example, without departing from the scope of the present disclosure, first information may also be called second information, and similarly, second information may also be called first information. Depending on the context, the word "if" as used herein may be interpreted as "at" or "when" or "in response to a determination."

First, the relevant technical background involved in the embodiment of the present application is introduced:

Introduction to AI-based wireless communication solutions:

Currently, more and more AI-based solutions are applied in wireless communication systems, such as through AI-based machine learning models, to solve the following problems:

(1) CSI feedback problem. Exemplarily, FIG. 1 is a schematic diagram of a process for solving a CSI feedback problem provided by an exemplary embodiment of the present application. As shown in Figure 1, the network device inputs the unprocessed CSI 101 into the AI encoder 102 of the CSI feedback model to obtain compressed CSI 103. Then input the compressed CSI 103 into the AI decoder 104 of the CSI feedback model to obtain the processed CSI 105. By introducing the AI encoder 102 and the AI decoder 104, compression and feedback of CSI information based on AI can be realized.

(2) Channel estimation problem. Exemplarily, FIG. 2 is a schematic diagram of a process for solving a channel estimation problem provided by an exemplary embodiment of the present application. As shown in FIG. 2 , the network device can estimate a given channel through a channel estimation model and according to data symbols and reference signal symbols, so as to restore a problematic channel. High performance estimation for a given channel can be achieved.

(3) Positioning problem. Exemplarily, FIG. 3 is a schematic diagram of a process for solving a positioning problem provided by an exemplary embodiment of the present application. As shown in FIG. 3 , the network device inputs positioning channel information 301 into a positioning model 302 , and the positioning model 302 processes the positioning channel information 301 through an AI-based positioning algorithm, thereby obtaining a high-precision positioning result 303 .

(4) Beam management problem. Exemplarily, FIG. 4 is a schematic diagram of a process for solving a beam management problem provided by an exemplary embodiment of the present application. As shown in Figure 4, the network device inputs the known beam information 401 into the beam management model 402, and the beam management model 402 processes the known beam information 401 through an AI-based beam management algorithm to obtain optimized beam information 403, thereby realizing Obtain preferred or more refined beam information, or obtain a prediction of beam information at a future moment.

The above-mentioned AI-based wireless communication solution shows a better performance gain than the current traditional solution using AI in the current research and application. But at the same time, the above-mentioned AI-based solutions are often strongly related to wireless communication scenarios, environments, and channels, and an important problem that AI-based solutions need to solve is how to deal with the problem of scene adaptation and multi-scenario universal application. That is, it is necessary to solve the problem of how to adapt the machine learning model to different scenarios, environments and channels, including the above-mentioned CSI feedback model, channel estimation model, positioning model, and beam management model for different scenarios, environments and channels. In addition, other AI-based solutions that are highly dependent on scenes, environments, and channels, such as codecs, noise cancellation, etc., also need to solve such problems.

An introduction to federated learning:

Traditional neural network training is centralized, such as model training after receiving a large amount of training data in a data center. However, after considering factors such as user privacy protection and computing power distribution, a special neural network training method "federated learning" was proposed. Federated learning means that in the process of training a neural network (machine learning model), the training set is distributed on each child node, and the local model is trained by the child nodes, and then the master node integrates the local models trained by each child node to obtain a global model. Exemplarily, FIG. 5 is a schematic diagram of a federated learning process provided by an exemplary implementation of the present application. As shown in FIG. 5 , firstly, the sub-node 502 , the sub-node 503 and the sub-node 504 generate a local local neural network based on the local training set, and then upload the local local neural network to the master node 501 . Afterwards, the master node 501 can synthesize the current global neural network according to the obtained local local neural networks, and transmit the global neural network to each sub-node. Then, the child nodes continue to use the new global neural network for the next training iteration. Finally, the training of the neural network is completed under the cooperation of multiple nodes.

AI-based solutions are highly dependent on scenarios, environments, and channels. A basic way to solve this problem is to build different machine learning models for different scenarios, environments, and channels. For example, different CSI feedback models, different channel estimation models, different positioning models, and different beam management models are constructed in different cells, indoors and outdoors, and in cities and suburbs. Then use the corresponding model to solve the corresponding wireless communication problem in specific cases.

However, there are still problems in the above solutions. Even if multiple models are prepared in advance to adapt to different scenarios, environments and channels, how to obtain various training data in different scenarios, environments and channels in advance (for example: channel data , beam data, location data) is still a problem, it is difficult to do complete data collection, data set construction required for machine learning, and algorithm and model construction for various scenarios, environments and channels in advance.

For the above-mentioned new problem, a potential solution is to re-update the algorithm and model for the new scene, environment and channel. For example, after the user (wireless communication device) arrives at a new scene, he can Do data collection in the scene, and then use the collected data as the data set required for algorithm and model construction, which is a method of online training and updating of the model. However, this method still has problems, mainly because it is difficult for a single user to complete the construction of a large amount of data sets in a short period of time. In addition, it will be a problem to build and train an independent solution (such as a neural network model) for each single user. Tasks with very high costs, including computing power, storage, and transmission, are difficult to support.

To sum up, through layer-by-layer analysis and feasibility analysis, the data limitation problem faced by scene adaptation is obvious, and it is an unresolved problem that needs to comprehensively consider the actual availability. The method provided in the embodiment of the present application can provide a model update method based on multi-user participation, and utilize the advantages of multi-user distributed data acquisition and data set construction to form a multi-user model for new scenarios, environments and channels. An update scheme of the model to solve the above problems.

Fig. 6 shows a schematic diagram of a system architecture of a communication system provided by an embodiment of the present application. The system architecture may include: a terminal device 10 , an access network device 20 and a core network device 30 .

The terminal device 10 may refer to a UE (User Equipment, user equipment), an access terminal, a subscriber unit, a subscriber station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a wireless communication device, a user agent or a user device. Optionally, the terminal equipment can also be a cellular phone, a cordless phone, a SIP (Session Initiation Protocol, session initiation protocol) phone, a WLL (Wireless Local Loop, wireless local loop) station, a PDA (Personal Digital Assistant, personal digital processing), Handheld devices with wireless communication functions, computing devices or other processing devices connected to wireless modems, vehicle-mounted devices, wearable devices, terminal devices in 5GS (5th Generation System, fifth-generation mobile communication system) or future evolved PLMN ( Public Land Mobile Network, terminal equipment in the public land mobile communication network), the embodiment of the present application is not limited to this. For convenience of description, the devices mentioned above are collectively referred to as terminal devices. The number of terminal devices 10 is generally multiple, and one or more terminal devices 10 may be distributed in a cell managed by each access network device 20 .

The access network device 20 is a device deployed in an access network to provide a wireless communication function for the terminal device 10 . The access network device 20 may include various forms of macro base stations, micro base stations, relay stations, access points, and so on. In systems using different wireless access technologies, the names of devices with access network device functions may be different. For example, in 5G NR systems, they are called gNodeB or gNB. With the evolution of communication technology, the name "access network equipment" may change. For the convenience of description, in the embodiment of the present application, the above-mentioned devices that provide the wireless communication function for the terminal device 10 are collectively referred to as access network devices. Optionally, a communication relationship may be established between the terminal device 10 and the core network device 30 through the access network device 20 . Exemplarily, in the LTE system, the access network device 20 may be EUTRAN (Evolved Universal Terrestrial Radio Access Network, Evolved Universal Terrestrial Radio Network) or one or more eNodeBs in EUTRAN; in the 5G NR system, the access The network device 20 may be the RAN or one or more gNBs in the RAN.

The functions of the core network device 30 are mainly to provide user connections, manage users, and carry out services, and provide an interface to external networks as a bearer network. For example, the core network equipment in the 5G NR system can include AMF (Access and Mobility Management Function, access and mobility management function) entity, UPF (User Plane Function, user plane function) entity and SMF (Session Management Function, session management function) entity and other equipment. The access network device 20 and the core network device 30 may be collectively referred to as network devices.

In an example, the access network device 20 and the core network device 30 communicate with each other through some air technology, such as the NG interface in the 5G NR system. The access network device 20 and the terminal device 10 communicate with each other through a certain air technology, such as a Uu interface.

Taking the method provided in the embodiment of the present application applied to the communication system shown in FIG. 6 as an example, the method provided in the embodiment of the present application is generally introduced. Fig. 7 is a schematic diagram of a process of updating a model for wireless channel processing provided by an exemplary embodiment of the present application. As shown in FIG. 7, the first wireless communication device (for example, including UE 702 and UE 703) updates the local machine learning model (for example, CSI feedback model, channel estimation model, positioning model, and beam management model) based on local data, so that Get the updated local machine learning model. Afterwards, the first wireless communication device uploads the updated local machine learning model to the second wireless communication device 702 . The second wireless communication device 702 updates the global machine learning model according to the local machine learning model uploaded by the first wireless communication device to obtain an updated global machine learning model. Then download the updated global machine learning model to the first wireless communication device, so that it can use the updated global machine learning model. In addition, the first wireless communication device and the second wireless communication device 702 can also perform the above steps multiple times to update the local machine learning model and the global machine learning model, so as to improve the accuracy of the machine learning model, and realize that the machine learning model can be used in various scenarios. , environment and channel adaptation.

In addition, the scope of the first wireless communication device participating in the model update needs to be specified, and only when the first wireless communication device is a valid user, it will participate in the model update. In addition, the first wireless communication device needs to be configured with transmission resources used for transmitting the updated local machine learning model. Exemplarily, FIG. 8 is a schematic diagram of a process of uploading and downloading a model provided by an exemplary embodiment of the present application. As shown in FIG. 8, in the case that UE 802 and UE 803 are valid users, the updated local machine learning model can be sent to base station 801. When the UE 804 is not a valid user, it cannot send the updated local machine learning model to the base station 801. Determining that the UE is a valid user can be determined in a variety of ways, for example, through the instruction of the base station 801, through the UE's self-determined or without a process of determining that the UE is a valid user. When the UE 802 and the UE 803 transmit the updated local machine learning model to the base station 801, they will transmit it through designated uplink transmission resources. The designated uplink transmission resource is configured by the base station 801, and the base station 801 can configure the designated uplink transmission resource transmission in various ways. For example, it is configured according to the request of the UE or directly configured for the UE. The base station 801 can update the global machine learning model according to the received updated local machine learning model of each UE. Afterwards, the updated global machine learning model is sent to the UE for use, and the base station 801 and the UE can continue to update the machine learning model in the above-mentioned manner.

The update of the machine learning model in the wireless communication system is performed in a manner of distributed updating of the wireless communication devices, so that local data of different wireless communication devices can be used to update the machine learning model. The local data of different wireless communication devices are generated under different scenarios, environments and channels, so it can enrich the training data for training machine learning models and make machine learning models adapt to different application scenarios. Therefore, it is possible to avoid a large amount of data collection work while solving the scene adaptation problem when the machine learning model handles the wireless communication problem. Improve the efficiency and accuracy of updating machine learning models.

Fig. 9 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 9 exemplifies that the method is applied to a terminal device in the communication system shown in FIG. 6 . The method includes:

Step 902: The first wireless communication device updates a local machine learning model based on local data.

The local machine learning model is a machine learning model deployed on the first wireless communication device. Optionally, the local machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The local machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The local data is related data generated by the first wireless communication device during wireless communication. The first wireless communication device includes one wireless communication device or multiple wireless communication devices.

The first wireless communication device updates the local machine learning model based on the local data, that is, the first wireless communication device uses the local data to train the local machine learning model. Optionally, updating the local machine learning model by the first wireless communication device means that the first wireless communication device updates at least one of coefficients and gradient information of the local machine learning model based on local data.

Optionally, the first wireless communication device periodically updates the local machine learning model, or updates the local machine learning model when local data changes, or updates the local machine learning model according to instructions.

Step 904: the first wireless communication device sends the updated local machine learning model to the second wireless communication device.

The updated local machine learning model is used by the second wireless communication device to update the global machine learning model, where the global machine learning model is a machine learning model deployed on the second wireless communication device. Optionally, the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The second wireless communication device is a base station. The global machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The global machine learning model has the same structure as the local machine learning model. Optionally, the structure of the global machine learning model is different from that of the local machine learning model. Illustratively, the local machine learning model is a subset of the global machine learning model. For example, the global machine learning model includes 10 network layers, and the local machine learning model includes 3 network layers, and the 3 network layers are a subset of the 10 network layers, that is, the 3 network layers and the 10 network layers 3 structures are the same. Exemplarily, the global machine learning model is a model integrating multiple sub-models, and the local machine learning model includes one or more of the multiple sub-models used to integrate the global machine learning model. For example, the structure of the global machine learning model is the cascade of model 1, model 2, and model 3, and the local machine learning model is model 2. In the case that the local machine learning model includes the foregoing multiple network layers, updating the local machine learning model by the first wireless communication device refers to updating at least one network layer. In the case that the local machine learning model includes the foregoing multiple sub-models, updating the local machine learning model by the first wireless communication device refers to updating at least one sub-model.

Sending the updated local machine learning model by the first wireless communication device includes sending at least one of coefficients and gradient information of the updated local machine learning model. The first wireless communication device periodically sends the updated local machine learning model, or sends the updated local machine learning model when the local machine learning model is updated, or sends the updated local machine learning model according to an instruction.

Optionally, if the first wireless communication device is a valid user, the first wireless communication device will send the updated local machine learning model to the second wireless communication device. The first wireless communication device is a valid user, which can be regarded as an authentication that the first wireless communication device can participate in updating the global machine learning model. The first wireless communication device is a valid user, which can be determined in various ways. For example, the first wireless communication device is determined to be a valid user through the indication of the second wireless communication device. The first wireless communication device is determined by the first wireless communication device as a valid user. Alternatively, the first wireless communication device is a valid user without a process of determining.

Optionally, when the first wireless communication device is a valid user, the first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource. The designated uplink transmission resource is configured by the second wireless communication device for the first wireless communication device, and the second wireless communication device can configure the designated uplink transmission resource for the first wireless communication device in various ways. For example, the second wireless communication device configures specified uplink transmission resources for the first wireless communication device according to the uplink resource application of the first wireless communication device. Or, the second wireless communication device directly configures the designated uplink transmission resource for the first wireless communication device.

Optionally, the downlink transmission resource used by the second wireless communication device to send a message to the first wireless communication device (for example, indicating that the first wireless communication device is a valid user, and configuring a specified uplink transmission resource) belongs to at least one of the following:

·Broadcast messages;

paging;

· Radio Resource Control (RRC) message;

· Media Access Control Control Element (MAC CE);

· Downlink Control Information (DCI);

·Transmission resources for carrying downlink control transmissions;

·Transmission resources for carrying downlink data transmission;

Carry the transmission resources of downlink artificial intelligence control transmission;

·Transmission resources for carrying downlink artificial intelligence data transmission.

Optionally, the uplink transmission resource (for example, the designated uplink transmission resource) used by the first wireless communication device to send the message to the second wireless communication device belongs to at least one of the following:

· Uplink Control Information (UCI);

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Carry the transmission resources of uplink artificial intelligence control transmission;

Carry the transmission resources for uplink artificial intelligence data transmission.

After the second wireless communication device receives the updated local machine learning model of the first wireless communication device, it can update the global machine learning model. Afterwards, the updated global machine learning model is sent to the first wireless communication device, and the first wireless communication device continues to perform the steps of updating the machine learning model and sending the updated machine learning model based on the updated global machine learning model.

Optionally, the second wireless communication device can also be a terminal. When sending the updated local machine learning model to the second wireless communication device, the first wireless communication device will send the updated local machine learning model through transmission resources of a sidelink (sidelink). At this time, the second wireless communication device forwards the updated local machine learning model of the first wireless communication device to the base station. If the second wireless communication device is used to merge and update the machine learning model at this time, and the second machine learning model is deployed with a global machine learning model, the second wireless communication device can Make updates to the global machine learning model. The transmission resources of the above-mentioned sidelink include at least one of the control channel and the data channel of the sidelink, such as a Physical Sidelink Control Channel (PSCCH) and a Physical Sidelink Shared Channel (Physical Sidelink Shared Channel). , PSSCH).

To sum up, the method provided in this embodiment updates the global machine learning model by means of distributed updating of the local machine learning model by the first wireless communication device, and can realize updating machine learning using local data of different wireless communication devices. Model. The local data of different wireless communication devices are generated under different scenarios, environments and channels, so it can enrich the training data for training machine learning models and make machine learning models adapt to different application scenarios. Therefore, it is possible to avoid a large amount of data collection work while solving the scene adaptation problem when the machine learning model handles the wireless communication problem. Improve the efficiency and accuracy of updating machine learning models.

Fig. 10 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 10 uses an example in which the method is applied to an access network device in the communication system shown in FIG. 6 . The method includes:

Step 1002: The second wireless communication device receives the updated local machine learning model sent by the first wireless communication device.

The local machine learning model is a machine learning model deployed on the first wireless communication device. Optionally, the local machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The local machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The first wireless communication device updates the local machine learning model according to the local data. Optionally, updating the local machine learning model by the first wireless communication device means that the first wireless communication device updates at least one of coefficients and gradient information of the local machine learning model based on local data. The local data is related data generated by the first wireless communication device during wireless communication. The first wireless communication device includes one wireless communication device or multiple wireless communication devices. Sending the updated local machine learning model by the first wireless communication device refers to sending at least one of coefficients and gradient information of the updated local machine learning model. The second wireless communication device is a base station.

Optionally, when the first wireless communication device is a valid user, the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device. The first wireless communication device is a valid user, which can be regarded as an authentication that the first wireless communication device can participate in updating the global machine learning model. The first wireless communication device is a valid user, which can be determined in various ways. For example, the first wireless communication device is determined to be a valid user through the indication of the second wireless communication device. The first wireless communication device is determined by the first wireless communication device as a valid user. Alternatively, the first wireless communication device is a valid user without a process of determining.

Optionally, when the first wireless communication device is a valid user, the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device through the designated uplink transmission resource. The designated uplink transmission resource is configured by the second wireless communication device for the first wireless communication device, and the second wireless communication device can configure the designated uplink transmission resource for the first wireless communication device in various ways. For example, the second wireless communication device configures specified uplink transmission resources for the first wireless communication device according to the uplink resource application of the first wireless communication device. Or, the second wireless communication device directly configures the designated uplink transmission resource for the first wireless communication device.

Step 1004: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

The global machine learning model is a machine learning model deployed on the second wireless communication device. Optionally, the second wireless communication device updates the global machine learning model according to the updated local machine learning model, which refers to updating at least one of coefficients and gradient information of the global machine learning model. The global machine learning model has the same structure as the local machine learning model. Optionally, the structure of the global machine learning model is different from that of the local machine learning model. For example, a local machine learning model is a subset of a global machine learning model. Alternatively, the global machine learning model is a model that integrates multiple sub-models, and the local machine learning model includes one or more of the multiple sub-models used to integrate the global machine learning model. The global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme, and beam management. The global machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model.

After the second wireless communication device receives the updated local machine learning model of the first wireless communication device, it can update the global machine learning model. Afterwards, the updated global machine learning model is sent to the first wireless communication device, and the first wireless communication device continues to perform the steps of updating the machine learning model and sending the updated machine learning model based on the updated global machine learning model. Therefore, the second wireless communication device will receive the updated global machine learning model sent by the first wireless communication device.

Optionally, the second wireless communication device can also be a terminal. At this time, when the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device, it will receive the updated local machine learning model through the transmission resource of the sidelink link.

When the first wireless communication device is a valid user, it can update the local machine learning model and send the updated local machine learning model to the second wireless communication device, so that the second wireless communication device can update the global machine learning model. The first wireless communication device is a valid user, which can be divided into three cases: (1) indicated by the second wireless communication device as a valid user. (2) The first wireless communication device determines that it is a valid user. (3) There is no process of determining that the first wireless communication device is a valid user, and the first wireless communication device is a valid user.

In addition, when the first wireless communication device sends the updated local machine learning model to the second wireless communication device, it can use the designated uplink transmission resource for transmission. Configuring designated uplink transmission resources for the first wireless communication device can be divided into two cases: (1) After the first wireless communication device applies for uplink transmission resources to the second wireless communication device, the second wireless communication device is the first wireless communication device Configure the specified uplink transmission resources. (2) The second wireless communication device directly configures designated uplink transmission resources for the first wireless communication device.

Based on the combination of the above three situations in which the first wireless communication device is a valid user and the two situations in which designated uplink transmission resources are configured for the first wireless communication device, the first wireless communication device and the second wireless communication device are introduced through the following six embodiments. The process by which a communication device updates a machine learning model.

The first type: for the situation where the second wireless communication device indicates that the first wireless communication device is a valid user, the first wireless communication device applies for uplink transmission resources to the second wireless communication device.

Fig. 11 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 11 illustrates an example in which the method is applied to the communication system shown in FIG. 6 . The method includes:

Step 1102: The first wireless communication device updates a local machine learning model based on local data.

The local machine learning model is a machine learning model deployed on the first wireless communication device. Optionally, the local machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The local machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The local data is related data generated by the first wireless communication device during wireless communication. The first wireless communication device includes one wireless communication device or multiple wireless communication devices. Optionally, the first wireless communication device updates at least one of coefficients and gradient information of the local machine learning model based on the local data.

Step 1104: the second wireless communication device indicates to the first wireless communication device that the first wireless communication device is a valid user.

The second wireless communication device uses one of multiple downlink transmission resources to indicate to the first wireless communication device that the first wireless communication device is a valid user. For example, the second wireless communication device activates the first wireless communication device to participate in the update of the CSI feedback model (as a valid user) through 1 bit in the DCI, or activates the first wireless communication device to participate in the update of the CSI feedback model through paging. Optionally, the second wireless communication device determines a valid user in one of the following manners, so as to indicate a valid user:

(1) The second wireless communication device randomly selects a wireless communication device from candidate wireless communication devices as a valid user. Candidate wireless communication devices include wireless communication devices capable of communicating messages with a second wireless communication device.

(2) The second wireless communication device selects the wireless communication device as a valid user according to the grouping of candidate wireless communication devices. This group corresponds to an Identity Document (ID). Optionally, the grouping is obtained by grouping the candidate wireless communication devices based on a local machine learning model on the candidate wireless communication devices, or by performing grouping for other purposes. And, for the same candidate wireless communication device. It can belong to one or more groups. For example, a certain candidate wireless communication device belongs to the first CSI update group, belongs to the second channel estimation group, belongs to the third positioning update group, and belongs to the fourth beam management group. The second wireless communication device selects one or more candidate wireless communication devices in a certain group as valid users, or selects one or more candidate wireless communication devices in multiple groups as valid users. For example, the candidate wireless communication devices are divided into 10 groups, and the second wireless communication device activates one group each time as a valid user, and activates each group in turn.

(3) The second wireless communication device selects the wireless communication device as a valid user according to the transmission complexity of the candidate wireless communication device. For example, the second wireless communication device selects a wireless transmission device with a smaller transmission load among candidate wireless transmission devices as an effective user for updating the CSI feedback model.

(4) The second wireless communication device selects the wireless communication device as a valid user according to the information processing performance of the candidate wireless communication device. The information processing performance is used to indicate the accuracy of the information output by the local machine learning model on the candidate wireless communication device. For example, the second wireless communication device selects a group of candidate wireless communication devices with poor CSI feedback performance as effective users, and updates the CSI feedback model for the group of candidate wireless communication devices.

Step 1106: In the case that the first wireless communication device is a valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device.

After the first wireless communication device is indicated as a valid user, the first wireless communication device will apply for an uplink transmission resource from the second wireless communication device. Optionally, in the case that the first wireless communication device is a valid user, if a trigger event is detected, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device. Optionally, the first wireless communication device detects the trigger event according to information processing performance. Alternatively, the first wireless communication device detects the trigger event according to the information-based transmission status. The information processing performance is used to indicate the accuracy of the information output by the local machine learning model, and the information-based transmission status is a communication status of the first wireless communication device based on the information output by the local machine learning model.

Exemplarily, the first wireless communication device determines whether to update the CSI feedback model based on the result of the current CSI feedback. For example, when the CSI feedback performance obtained by the first wireless communication device through the CSI feedback model does not meet a requirement of a specific threshold, the first wireless communication device detects a trigger event. Exemplarily, the first wireless communication device judges whether the current CSI feedback model needs to be updated based on the success rate or failure rate of current data or data packet transmission, so as to determine whether a trigger event is detected.

Optionally, the trigger event includes at least one of the following:

When the local machine learning model is used for CSI feedback (the local machine learning model is a CSI feedback model), the CSI to be compressed and the restored CSI do not satisfy the first condition;

In the case where the local machine learning model is used for CSI feedback, the transmission state of the result of the first wireless communication based on the CSI feedback does not satisfy the second condition;

In the case where the local machine learning model is used for channel estimation (the local machine learning model is a channel estimation model), the channel estimation performance of the first wireless communication device does not meet the third condition;

In the case where a local machine learning model is used for channel estimation, the transmission state of the first wireless communication device based on the result of channel estimation does not satisfy the fourth condition;

When the local machine learning model is used for positioning (the local machine learning model is a positioning model), the positioning accuracy of the first wireless communication device does not meet the fifth condition;

· In the case where the local machine learning model is used for beam management (the local machine learning model is a beam management model), the beam management accuracy of the first wireless communication device does not meet the sixth condition;

• In the case where the local machine learning model is used for beam management, the transmission state of the first wireless communication device based on the result of beam management does not satisfy the seventh condition.

Exemplarily, when the local machine learning model is used for CSI feedback, the trigger event includes: the degree of deviation between the CSI to be compressed and the restored CSI is greater than or equal to the first threshold, and the degree of similarity between the CSI to be compressed and the restored CSI is less than or is equal to the second threshold, and the success rate or failure rate of the first wireless communication device performing data transmission (or performing data packet transmission) does not meet at least one of the specific thresholds. For example, the block error rate (Block Error Rate, BLER) of the first wireless communication device for data transmission is higher than the third threshold, or, the bit error probability (Bit Error Ratio, BER) of the first wireless communication device for data transmission is higher than the third threshold Four thresholds. In the case where a local machine learning model is used for channel estimation, the trigger events include: the channel estimation error of the first wireless communication device is higher than the fifth threshold, the channel similarity of the first wireless communication device is lower than the sixth threshold, and the first The success rate or failure rate of data transmission (or data packet transmission) performed by the wireless communication device does not meet at least one of specific thresholds. In the case where the local machine learning model is used for positioning, the trigger event includes: the positioning accuracy of the first wireless communication device is less than the seventh threshold. In the case where the local machine learning model is used for beam management, the triggering event includes: the beam management accuracy of the first wireless communication device is less than the eighth threshold, and the success rate of the first wireless communication device for data transmission (or for data packet transmission) Or the failure rate does not meet at least one of certain thresholds.

Optionally, the foregoing conditions are stipulated by a protocol, or configured by the second wireless communication device for the first wireless communication device. For example, the foregoing threshold is stipulated in an agreement, or the second wireless communication device configures the foregoing threshold for the first wireless communication device.

Step 1108: the second wireless communication device configures a designated uplink transmission resource to the first wireless communication device.

After the second wireless communication device configures the designated uplink transmission resource to the first wireless communication device, the first wireless communication device can receive the designated uplink transmission resource. The second wireless communication device configures designated uplink transmission resources for the first wireless communication device, including configuring time domain and frequency domain resources and corresponding transmission configuration information.

Optionally, the second wireless communication device can configure designated uplink transmission resources for it in response to the application of the first wireless communication device. The second wireless communication device can also not respond to the application of the first wireless communication device. In this case, the second wireless communication device can also send a message to notify the first wireless communication device to reject its application, or notify the first wireless communication device that it does not need Perform local machine learning model updates. The second wireless communication device can notify the first wireless communication device through one or more types of downlink transmission resources.

Optionally, the specified uplink transmission resource belongs to at least one of the following:

UCI;

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Step 1110: the first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.

The updated local machine learning model is used by the second wireless communication device to update the global machine learning model, where the global machine learning model is a machine learning model deployed on the second wireless communication device. Optionally, the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The global machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The global machine learning model has the same structure as the local machine learning model. Optionally, sending the updated local machine learning model by the first wireless communication device refers to sending at least one of coefficients and gradient information of the updated local machine learning model.

It should be noted that, when the second wireless communication device indicates that the first wireless communication device is a valid user, the duration of the first wireless communication device as a valid user includes at least one of the following:

(1) The time period between when the first wireless communication device is indicated as a valid user by the second wireless communication device and when the first wireless communication device sends the updated local machine learning model to the second wireless communication device.

(2) The time period between the first wireless communication device being indicated as a valid user by the second wireless communication device and the end of the preset time period. Optionally, the preset duration can be predefined through a protocol, or the preset duration is configured by the second wireless communication device for the first wireless communication device. Exemplarily, the preset duration is N time slots (slots), or M milliseconds, or K seconds, minutes, hours, etc.

(3) The time period between the first wireless communication device being indicated as a valid user by the second wireless communication device and the end of the preset number of times. The preset number of times is set for the number of times the first wireless communication device sends the updated local machine learning model to the second wireless communication device. The first wireless communication device can update the local machine learning model multiple times and send it to the second wireless communication device. If the number of times the first wireless communication device sends the updated local machine learning model to the second wireless communication device reaches a preset number of times, it is necessary to re-determine whether the first wireless communication device is a valid user.

(4) The time period between when the first wireless communication device is indicated as a valid user by the second wireless communication device and when the first wireless communication device is not indicated as a valid user by the second wireless communication device.

If the first wireless communication device is not a valid user, it cannot send the updated local machine learning model to the second wireless communication device.

Exemplarily, FIG. 12 is a schematic diagram of the second wireless communication device determining the duration of the first wireless communication device as a valid user provided by an exemplary embodiment of the present application. As shown in Figure 12, in the case that the base station 1201 indicates that the user 1202 (wireless communication device) is a valid user, the base station 1201 receives the updated local machine sent by the user 1202 according to the base station 1201 indicating that the user 1202 is a valid user. The duration between learning models determines the duration of user 1202 as a valid user. That is, before each updated local machine learning model is transmitted, it is necessary to determine whether the user 1202 is a valid user.

Exemplarily, FIG. 13 is a schematic diagram of the second wireless communication device determining the duration of the first wireless communication device as a valid user provided by an exemplary embodiment of the present application. As shown in Figure 13, when the base station 1301 indicates that the user 1302 is a valid user, the base station 1301 determines the duration of the user 1302 as a valid user according to the time period between the base station 1301 indicating that the user 1302 is a valid user and the end of the preset time period. duration. Alternatively, the base station 1301 determines the duration of the user 1302 as a valid user according to the time period between the base station 1301 indicating that the user 1302 is a valid user and the end of the preset times.

Exemplarily, FIG. 14 is a schematic diagram of the second wireless communication device determining the duration of the first wireless communication device as a valid user provided by an exemplary embodiment of the present application. As shown in FIG. 14 , after the base station 1401 determines the valid user qualification of the user 1402, the user 1402 is always valid as a valid user before the base station 1402 notifies that the valid user is invalid. The base station 1401 can notify the user 1402 that he is no longer a valid user, or no longer participate in updating the machine learning model, through one or more types of downlink transmission resources.

Step 1112: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

The global machine learning model has the same structure as the local machine learning model. Optionally, the second wireless communication device updating the global machine learning model according to the updated local machine learning model refers to updating at least one of coefficients and gradient information of the global machine learning model.

Exemplarily, each UE participating in machine learning model update uploads an updated local CSI feedback model to the base station, and the base station receives multiple local CSI feedback models transmitted by multiple UEs. Afterwards, the global CSI feedback model is updated based on the received local CSI feedback model, the base station transmits the global CSI feedback model to each UE, and the UE receives the updated global CSI feedback model. The UE directly uses the received CSI feedback model to perform CSI compression and feedback, and the UE can also use the received CSI feedback model to continue to perform local model update and continue the above steps.

Exemplarily, FIG. 15 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application. As shown in Figure 15, after determining that UE 1502 is a valid user, base station 1501 will indicate to UE 1502 that it is a valid user. UE 1502 applies for uplink transmission resources to base station 1501, and base station 1501 configures and specifies uplink transmission resources for UE 1502. Afterwards, the UE 1502 sends the updated local CSI feedback model to the base station 1501 by specifying uplink transmission resources.

In addition, valid users are determined among candidate wireless communication devices in various ways, so that a suitable wireless communication device can be flexibly selected to participate in updating the machine learning model. Configuring designated uplink transmission resources for the first wireless communication device to transmit the updated local machine learning model can ensure transmission efficiency and avoid impact on transmission of other messages.

The second type: for the situation where the second wireless communication device indicates that the first wireless communication device is a valid user, the second wireless communication device directly configures a designated uplink transmission resource for the first wireless communication device.

FIG. 16 shows a flow chart of model update for wireless channel processing provided by an embodiment of the present application. FIG. 16 illustrates the method applied to the communication system shown in FIG. 6 . The method includes:

Step 1602: The first wireless communication device updates a local machine learning model based on local data.

Step 1604: The second wireless communication device indicates to the first wireless communication device that the first wireless communication device is a valid user.

Optionally, the second wireless communication device determines a valid user in one of the following ways:

The second wireless communication device randomly selects a wireless communication device from candidate wireless communication devices as an effective user. The second wireless communication device selects the wireless communication device as a valid user according to the grouping of candidate wireless communication devices. The second wireless communication device selects the wireless communication device as a valid user according to the transmission complexity of the candidate wireless communication device. The second wireless communication device selects the wireless communication device as an effective user according to the information processing performance of the candidate wireless communication device, and the information processing performance is used to indicate the accuracy of the information output by the local machine learning model on the candidate wireless communication device.

Step 1606: In the case that the first wireless communication device is a valid user, the second wireless communication device configures a designated uplink transmission resource to the first wireless communication device.

When the first wireless communication device is a valid user, the second wireless communication device can directly configure the designated uplink transmission resource to the first wireless communication device, so that the first wireless communication device receives the designated uplink transmission resource. For example, after the second wireless communication device indicates that the first wireless communication device is a valid user, it configures the designated uplink transmission resource to the first wireless communication device immediately.

UCI;

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Step 1608: The first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.

The updated local machine learning model is used by the second wireless communication device to update the global machine learning model, and the global machine learning model is a machine learning model deployed on the second wireless communication device. Optionally, the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The global machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The global machine learning model has the same structure as the local machine learning model. Optionally, sending the updated local machine learning model by the first wireless communication device refers to sending at least one of coefficients and gradient information of the updated local machine learning model.

The time period between when the first wireless communication device is indicated as a valid user by the second wireless communication device and when the first wireless communication device sends the updated local machine learning model to the second wireless communication device;

The time period between the first wireless communication device being indicated as a valid user by the second wireless communication device and the end of the preset time period;

The length of time between the first wireless communication device being indicated as a valid user by the second wireless communication device and the end of the preset number of times, the preset number of times is for the first wireless communication device to send the updated local machine learning to the second wireless communication device The number of times of the model is set;

The time period between when the first wireless communication device is indicated as a valid user by the second wireless communication device and when the first wireless communication device is not indicated as a valid user by the second wireless communication device.

Step 1610: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

Exemplarily, FIG. 17 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application. As shown in Figure 17, after determining that UE 1702 is a valid user, base station 1701 will indicate to UE 1702 that it is a valid user. Afterwards, the base station 1701 configures and specifies uplink transmission resources for the UE 1702. Afterwards, the UE 1702 sends the updated local CSI feedback model to the base station 1701 by specifying uplink transmission resources.

In addition, valid users are determined among candidate wireless communication devices in various ways, so that a suitable wireless communication device can be flexibly selected to participate in updating the machine learning model. Configuring designated uplink transmission resources for the first wireless communication device to transmit the updated local machine learning model can ensure transmission efficiency and avoid impact on transmission of other messages. There is no need for the first wireless communication device to apply for uplink resources, and the process of configuring designated uplink transmission resources can be simplified.

The third type: for the case where the first wireless communication device applies for an uplink transmission resource from the second wireless communication device without confirming that the first wireless communication device is a valid user.

Fig. 18 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 18 exemplifies that the method is applied to the communication system shown in FIG. 6 . The method includes:

Step 1802: The first wireless communication device updates the local machine learning model based on the local data.

Step 1804: In the case that the first wireless communication device is a valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device.

For the first wireless communication device, it can become a valid user to participate in updating the machine learning model without a definite process. Exemplarily, when the first wireless communication device satisfies the valid user condition, the first wireless communication device is a valid user. Wherein, the first wireless communication device meets valid user conditions including at least one of the following:

The first wireless communication device is a default valid user;

• The device capabilities of the first wireless communication device include that the first wireless communication device is a valid user.

If the first wireless communication device is a valid user, the first wireless communication device will apply for uplink transmission resources from the second wireless communication device. Optionally, in the case that the first wireless communication device is a valid user, if a trigger event is detected, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device.

Optionally, the first wireless communication device detects the trigger event according to information processing performance. Alternatively, the first wireless communication device detects the trigger event according to the information-based transmission status. Wherein, the information processing performance is used to indicate the accuracy of the information output by the local machine learning model, and the information-based transmission status is a communication status of the first wireless communication device based on the information output by the local machine learning model.

Step 1806: The second wireless communication device configures a designated uplink transmission resource to the first wireless communication device.

UCI;

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Step 1808: The first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.

Step 1810: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

Exemplarily, FIG. 19 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application. As shown in Figure 19, when UE 1902 is a valid user, it will apply for uplink transmission resources to base station 1901. The base station 1901 configures and specifies uplink transmission resources for the UE 1902. Afterwards, the UE 1902 sends the updated local CSI feedback model to the base station 1901 by specifying uplink transmission resources.

In addition, configuring designated uplink transmission resources for the first wireless communication device to transmit the updated local machine learning model can ensure transmission efficiency and avoid impact on transmission of other messages.

The fourth type: for the case where the first wireless communication device determines that it is a valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device.

Fig. 20 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 20 exemplifies that the method is applied to the communication system shown in FIG. 6 . The method includes:

Step 2002: The first wireless communication device updates a local machine learning model based on local data.

Step 2004: the first wireless communication device determines that the first wireless communication device is a valid user.

The first wireless communication device can determine whether it is a valid user by itself. Optionally, the first wireless communication device determines that the first wireless communication device is a valid user according to a magnitude relationship between the random number and the probability threshold. Wherein, the random number is generated by the first wireless communication device. For example, when the random number is higher than or lower than the probability threshold, the first wireless communication device determines that it is a valid user. The probability threshold can be pre-agreed through a protocol, or configured by the second wireless communication device to the first wireless communication device, specifically, the second wireless communication device configures one or more types of downlink transmission resources. In addition, the probability threshold can be changed. For example, when more users are expected to participate in the update of the machine learning model, the probability threshold can be adjusted to a value that makes the wireless communication device more likely to become a valid user. Wireless communication devices that meet the probability threshold requirements Will be more. The above update can also be notified by the second wireless communication device to the first wireless communication device.

Step 2006: In the case that the first wireless communication device is a valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device.

Step 2008: the second wireless communication device configures a designated uplink transmission resource to the first wireless communication device.

UCI;

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Step 2010: The first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.

Step 2012: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

Exemplarily, FIG. 21 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application. As shown in Figure 21, when UE 2102 determines that it is a valid user, it will apply for uplink transmission resources to base station 2101. The base station 2101 configures and specifies uplink transmission resources for the UE 2102. Afterwards, the UE 2102 sends the updated local CSI feedback model to the base station 2101 by specifying uplink transmission resources.

In addition, the first wireless communication device determines that it is a valid user, which can simplify the process of determining a valid user. Configuring designated uplink transmission resources for the first wireless communication device to transmit the updated local machine learning model can ensure transmission efficiency and avoid impact on transmission of other messages.

The fifth type: the first wireless communication device determines that it is a valid user, the first wireless communication device notifies the second wireless communication device that it is a valid user, and the first wireless communication device applies for uplink transmission resources to the second wireless communication device.

Fig. 22 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 22 exemplifies that the method is applied to the communication system shown in FIG. 6 . The method includes:

Step 2202: The first wireless communication device updates a local machine learning model based on local data.

Step 2204: the first wireless communication device determines that the first wireless communication device is a valid user.

The first wireless communication device can determine whether it is a valid user by itself. Optionally, the first wireless communication device determines that the first wireless communication device is a valid user according to a magnitude relationship between the random number and the probability threshold. Wherein, the random number is generated by the first wireless communication device.

Step 2206: When the first wireless communication device determines that the first wireless communication device is a valid user, the first wireless communication device notifies the second wireless communication device that the first wireless communication device is a valid user.

When the first wireless communication device determines that it is a valid user, the first wireless communication device will notify the second wireless communication device that the first wireless communication device is a valid user, so that the second wireless communication device can learn about the first wireless communication Device is a valid user. In the case where the first wireless communication device is determined as a valid user by the first wireless communication device, the second wireless communication device will receive a notification sent by the first wireless communication device that the first wireless communication device is a valid user.

Step 2208: In the case that the first wireless communication device is a valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device.

Step 2210: The second wireless communication device configures a designated uplink transmission resource to the first wireless communication device.

UCI;

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Step 2212: The first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying uplink transmission resources.

Step 2214: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

Exemplarily, FIG. 23 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application. As shown in Figure 23, when the UE 2302 determines that it is a valid user, it will notify the base station 2301 that the UE 2302 is a valid user. Then the UE 2302 applies to the base station 2301 for an uplink transmission resource. The base station 2301 configures and specifies uplink transmission resources for the UE 2302. Afterwards, the UE 2302 sends the updated local CSI feedback model to the base station 2301 by specifying uplink transmission resources.

The sixth type: the first wireless communication device determines that it is a valid user, the first wireless communication device notifies the second wireless communication device that it is a valid user, and the second wireless communication device directly configures the specified uplink transmission resource for the first wireless communication device Condition.

Fig. 24 shows a flowchart of a model updating method for wireless channel processing provided by an embodiment of the present application. FIG. 24 exemplifies that the method is applied to the communication system shown in FIG. 6 . The method includes:

Step 2402: The first wireless communication device updates a local machine learning model based on local data.

Step 2404: the first wireless communication device determines that the first wireless communication device is a valid user.

Step 2406: When the first wireless communication device determines that the first wireless communication device is a valid user, the first wireless communication device notifies the second wireless communication device that the first wireless communication device is a valid user.

Step 2408: When the first wireless communication device is a valid user, the second wireless communication device configures a designated uplink transmission resource to the first wireless communication device.

UCI;

· RRC message;

The transmission resources that carry the uplink control transmission;

·Transmission resources for carrying uplink data transmission;

Step 2410: The first wireless communication device sends the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.

The updated local machine learning model is used by the second wireless communication device to update the global machine learning model, where the global machine learning model is a machine learning model deployed on the second wireless communication device. Optionally, the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management. The global machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model. The global machine learning model has the same structure as the local machine learning model. Optionally, the first wireless communication device sending the updated local machine learning model refers to sending at least one of coefficients and gradient information of the updated local machine learning model.

Step 2412: The second wireless communication device updates the global machine learning model according to the updated local machine learning model.

Exemplarily, FIG. 25 is a schematic diagram of a process of sending and updating a machine learning model provided by an exemplary embodiment of the present application. As shown in Figure 25, when the UE 2502 determines that it is a valid user, it will notify the base station 2501 that the UE 2502 is a valid user. Afterwards, the base station 2501 configures and specifies uplink transmission resources for the UE 2502. Afterwards, the UE 2502 sends the updated local CSI feedback model to the base station 2501 by specifying uplink transmission resources.

It should be noted that, in the foregoing embodiment, the local machine learning model includes at least one of a CSI feedback model, a channel estimation model, a positioning model, and a beam management model as an example for description. The machine learning model for solving other problems in wireless communication can also be updated through the methods provided in the above embodiments. The embodiment of the present application does not limit this.

In addition, the first wireless communication device determines that it is a valid user, which can simplify the process of determining a valid user. Configuring designated uplink transmission resources for the first wireless communication device to transmit the updated local machine learning model can ensure transmission efficiency and avoid impact on transmission of other messages. There is no need for the first wireless communication device to apply for uplink resources, and the process of configuring designated uplink transmission resources can be simplified.

The above-mentioned embodiments can be implemented independently or in combination freely.

Fig. 26 shows a block diagram of an apparatus for updating a model for wireless channel processing provided by an embodiment of the present application. As shown in Figure 26, the device includes:

An updating module 2601, configured to update a local machine learning model based on local data.

A sending module 2602, configured to send the updated local machine learning model to the second wireless communication device, where the updated local machine learning model is used to update the global machine learning model.

In an optional design, the sending module 2602 is used for:

If the device is a valid user, the updated local machine learning model is sent to the second wireless communication device.

In an optional design, the sending module 2602 is used for:

When the device is a valid user, the updated local machine learning model is sent to the second wireless communication device by specifying an uplink transmission resource.

In an optional design, the device also includes:

A determining module 2603, configured to determine that the apparatus is a valid user according to an instruction of the second wireless communication device;

or,

A determining module 2603, configured to determine that the device is a valid user.

In an optional design, valid users include at least one of the following:

a wireless communication device randomly selected by the second wireless communication device from candidate wireless communication devices;

A wireless communication device selected by the second wireless communication device according to the grouping of candidate wireless communication devices;

The wireless communication device selected by the second wireless communication device according to the transmission complexity of the candidate wireless communication device;

The second wireless communication device is a wireless communication device selected according to the information processing performance of the candidate wireless communication device, and the information processing performance is used to indicate the accuracy of the information output by the local machine learning model on the candidate wireless communication device.

In an optional design, a determination module 2603 is used to:

According to the size relationship between the random number and the probability threshold, it is determined that the device is a valid user. Wherein, the random number is generated by the device.

In an optional design, if the device satisfies the valid user condition, the device is a valid user. Wherein, the device meeting valid user conditions includes at least one of the following:

The device is the default active user;

The device capabilities of the device include that the device is a valid user.

In an optional design, the sending module 2602 is used for:

In a case where the device determines that the device is a valid user, the second wireless communication device is notified that the device is a valid user.

In an optional design, the device also includes:

The sending module 2602 is configured to apply for an uplink transmission resource to the second wireless communication device when the device is a valid user. A receiving module 2604, configured to receive a designated uplink transmission resource configured by the second wireless communication device;

or,

The receiving module 2604 is configured to receive the specified uplink transmission resource configured by the second wireless communication device when the device is a valid user.

In an optional design, the sending module 2602 is used for:

When the device is a valid user, if a trigger event is detected, apply for uplink transmission resources to the second wireless communication device.

In an optional design, the trigger event includes at least one of the following:

In the case where the local machine learning model is used for CSI feedback, the CSI to be compressed and the restored CSI do not satisfy the first condition;

In the case where the local machine learning model is used for the CSI feedback, the transmission state of the device based on the result of the CSI feedback does not meet the second condition;

In the case where a local machine learning model is used for channel estimation, the channel estimation performance of the device does not meet the third condition;

In the case where the local machine learning model is used for channel estimation, the transmission state of the device based on the result of channel estimation does not meet the fourth condition;

In the case where a local machine learning model is used for positioning, the positioning accuracy of the device does not meet the fifth condition;

In the case where a local machine learning model is used for beam management, the beam management accuracy of the device does not meet the sixth condition;

In the case where the local machine learning model is used for beam management, the transmission state of the device based on the result of beam management does not satisfy the seventh condition.

In an optional design, when the device is indicated by the second wireless communication device as a valid user, the duration of the device as a valid user includes at least one of the following:

The time period between when the device is indicated as a valid user by the second wireless communication device and when the device sends the updated local machine learning model to the second wireless communication device.

A time period between when the device is indicated as a valid user by the second wireless communication device and when the preset time period ends.

The time period between when the device is indicated as a valid user by the second wireless communication device and when the preset number of times is set for the number of times the device sends the updated local machine learning model to the second wireless communication device.

The time period between when the device is indicated as a valid user by the second wireless communication device and when the device is not indicated as a valid user by the second wireless communication device.

In an optional design, update module 2601 for:

At least one of coefficients and gradient information of the local machine learning model is updated based on the local data.

In an optional design, the second wireless communication device is a terminal. Sending module 2602, used for:

The updated local machine learning model is sent to the second wireless communication device through the transmission resource of the sidelink.

In an optional design, the device also includes:

The receiving module 2604 is configured to receive the updated global machine learning model sent by the second wireless communication device.

The update module 2601 is configured to continue to execute the step of updating the machine learning model based on the updated global machine learning model, and the sending module 2602 is configured to execute the step of sending the updated machine learning model.

In an optional design, the local machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme, and beam management.

In an optional design, the downlink transmission resource used by the second wireless communication device to send a message to the device belongs to at least one of the following:

broadcast message;

paging;

RRC message;

MAC CE;

DCI;

transmission resources carrying downlink control transmission;

Transmission resources for carrying downlink data transmission;

Transmission resources for carrying downlink artificial intelligence control transmissions;

The transmission resource that carries the downlink artificial intelligence data transmission.

In an optional design, the specified uplink transmission resource belongs to at least one of the following:

UCI;

RRC message;

transmission resources carrying uplink control transmission;

Transmission resources for carrying uplink data transmission;

Transmission resources carrying uplink artificial intelligence control transmission;

Transmission resources that carry uplink artificial intelligence data transmission.

Fig. 27 shows a block diagram of a model updating device for wireless channel processing provided by an embodiment of the present application. As shown in Figure 27, the device includes:

The receiving module 2701 is configured to receive an updated local machine learning model sent by the first wireless communication device, where the updated local machine learning model is obtained by updating the local machine learning model based on local data by the first wireless communication device.

An update module 2702, configured to update the global machine learning model according to the updated local machine learning model.

In an optional design, the receiving module 2701 is used for:

If the first wireless communication device is a valid user, the updated local machine learning model sent by the first wireless communication device is received.

In an optional design, the receiving module 2701 is used for:

When the first wireless communication device is a valid user, the updated local machine learning model sent by the first wireless communication device through the specified uplink transmission resource is received.

In an optional design, the device also includes:

The sending module 2703 is configured to indicate to the first wireless communication device that the first wireless communication device is a valid user.

In an optional design, the device also includes:

A determining module 2704, configured to randomly select a wireless communication device from candidate wireless communication devices as a valid user.

In an optional design, the device also includes:

A determining module 2704, configured to select a wireless communication device as a valid user according to the grouping of candidate wireless communication devices.

In an optional design, the device also includes:

A determining module 2704, configured to select a wireless communication device as a valid user according to the transmission complexity of the candidate wireless communication device.

In an optional design, the device also includes:

The determination module 2704 is configured to select a wireless communication device as an effective user according to the information processing performance of the candidate wireless communication device, and the information processing performance is used to indicate the accuracy of the information output by the local machine learning model on the candidate wireless communication device.

In an optional design, it is determined by the first wireless communication device that the first wireless communication device is a valid user.

In an optional design, when the first wireless communication device satisfies a valid user condition, the first wireless communication device is a valid user. Wherein, the first wireless communication device meets valid user conditions including at least one of the following:

The first wireless communication device is a default valid user;

The device capability of the first wireless communication device includes that the first wireless communication device is a valid user.

In an optional design, the receiving module 2701 is used for:

In a case where the first wireless communication device determines that the valid user is the first wireless communication device, the apparatus receives a notification sent by the first wireless communication device that the first wireless communication device is a valid user.

In an optional design, the device also includes:

The sending module 2703 is configured to configure specified uplink transmission resources to the first wireless communication device according to the uplink transmission resource application of the first wireless communication device when the first wireless communication device is a valid user;

or,

The sending module 2703 is configured to configure a designated uplink transmission resource to the first wireless communication device when the first wireless communication device is a valid user.

In an optional design, the device also includes:

The determining module 2704 is configured to, when the device indicates that the first wireless communication device is a valid user, according to the device indicating that the first wireless communication device is a valid user, so that the device receives the updated local machine learning sent by the first wireless communication device The duration between models determines the duration for which the first wireless communication device is an active user.

In an optional design, the device also includes:

A determining module 2704, configured to determine the first wireless communication device according to the time period between the device indicating that the first wireless communication device is a valid user and the end of the preset time period when the device indicates that the first wireless communication device is a valid user Duration of being an active user.

In an optional design, the device also includes:

A determining module 2704, configured to determine the first wireless communication device according to the time period between the device indicating that the first wireless communication device is a valid user and the end of the preset number of times when the device indicates that the first wireless communication device is a valid user Duration of being an active user. Wherein, the preset number of times is set for the number of times that the first wireless communication device sends the updated local machine learning model to the device.

In an optional design, the device also includes:

A determining module 2704, configured to, in the case that the device indicates that the first wireless communication device is a valid user, according to the time period between when the device indicates that the first wireless communication device is a valid user and when the device indicates that the first wireless communication device is not a valid user , to determine the duration of the first wireless communication device as a valid user.

In an optional design, module 2702 is updated for:

At least one of coefficients and gradient information of the global machine learning model is updated according to the updated local machine learning model.

In an optional design, the device is a terminal. The receiving module 2701 is used for:

The updated local machine learning model sent by the first wireless communication device through the transmission resource of the sidelink is received.

In an optional design, the device also includes:

A sending module 2703, configured to send the updated global machine learning model to the first wireless communication device.

The receiving module 2701 is configured to receive the re-updated global machine learning model sent by the first wireless communication device.

In an optional design, the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme, and beam management.

In an optional design, the downlink transmission resource used by the device to send a message to the first wireless communication device belongs to at least one of the following:

broadcast message;

paging;

RRC message;

MAC CE;

DCI;

transmission resources carrying downlink control transmission;

Transmission resources for carrying downlink data transmission;

UCI;

RRC message;

transmission resources carrying uplink control transmission;

Transmission resources for carrying uplink data transmission;

It should be noted that when the device provided by the above embodiment realizes its functions, it only uses the division of the above-mentioned functional modules as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional modules according to actual needs. That is, the content structure of the device is divided into different functional modules to complete all or part of the functions described above.

Regarding the apparatus in the foregoing embodiments, the specific manner in which each module executes operations has been described in detail in the embodiments related to the method, and will not be described in detail here.

FIG. 28 shows a schematic structural diagram of a communication device (terminal device or network device) provided by an exemplary embodiment of the present application. The communication device 280 includes: a processor 2801, a receiver 2802, a transmitter 2803, a memory 2804 and a bus 2805 .

The processor 2801 includes one or more processing cores, and the processor 2801 executes various functional applications and information processing by running software programs and modules.

The receiver 2802 and the transmitter 2803 can be realized as a communication component, and the communication component can be a communication chip.

The memory 2804 is connected to the processor 2801 through the bus 2805 .

The memory 2804 may be used to store at least one instruction, and the processor 2801 is used to execute the at least one instruction, so as to implement various steps in the foregoing method embodiments.

In addition, the memory 2804 can be implemented by any type of volatile or non-volatile storage device or their combination, volatile or non-volatile storage devices include but not limited to: magnetic disk or optical disk, electrically erasable and programmable Read Only Memory (Erasable Programmable Read Only Memory, EEPROM), Erasable Programmable Read Only Memory (Erasable Programmable Read Only Memory, EPROM), Static Random Access Memory (SRAM), Read Only Memory (Read -Only Memory, ROM), magnetic memory, flash memory, programmable read-only memory (Programmable Read-Only Memory, PROM).

Wherein, when the communication device is implemented as a terminal device, the processor and the transceiver in the communication device involved in the embodiment of the present application may perform the steps performed by the terminal device in the method shown in any of the above method embodiments, where No longer.

Wherein, when the communication device is implemented as an access network device, the processor and the transceiver in the communication device involved in the embodiment of the present application may perform the steps performed by the access network device in any of the methods shown above, where I won't repeat them here.

In an exemplary embodiment, a computer-readable storage medium is also provided, the computer-readable storage medium stores at least one instruction, at least one program, a code set or an instruction set, the at least one instruction, the At least one program, the code set or the instruction set is loaded and executed by the processor to implement the model update method for wireless channel processing provided by the above method embodiments.

In an exemplary embodiment, a chip is also provided, the chip includes a programmable logic circuit and/or program instructions, and when the chip is run on a communication device, it is used to implement the functions provided by the above method embodiments. Model update method for wireless channel processing.

In an exemplary embodiment, a computer program product is also provided, which, when run on a processor of a computer device, causes the computer device to execute the above-mentioned model updating method for wireless channel processing.

Those skilled in the art should be aware that, in the foregoing one or more examples, the functions described in the embodiments of the present application may be implemented by hardware, software, firmware or any combination thereof. When implemented in software, the functions may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.

The above are only exemplary embodiments of the application, and are not intended to limit the application. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the application shall be included in the protection of the application. within range.

Claims

A method for updating a model for wireless channel processing, characterized in that the method is executed by a first wireless communication device, and the method includes:

updating a local machine learning model based on local data by the first wireless communication device;

The first wireless communication device sends the updated local machine learning model to the second wireless communication device, and the updated local machine learning model is used to update the global machine learning model.
The method according to claim 1, wherein the first wireless communication device sends the updated local machine learning model to the second wireless communication device, comprising:

If the first wireless communication device is a valid user, sending the updated local machine learning model to the second wireless communication device.
The method according to claim 2, wherein when the first wireless communication device is a valid user, sending the updated local machine learning model to the second wireless communication device includes :

When the first wireless communication device is the valid user, sending the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.
The method according to claim 3, characterized in that the method further comprises:

The first wireless communication device determines that the first wireless communication device is the valid user according to the instruction of the second wireless communication device;

or,

The first wireless communication device determines that the first wireless communication device is the valid user.
The method according to claim 4, wherein the valid users include at least one of the following:

A wireless communication device randomly selected by the second wireless communication device from candidate wireless communication devices;

a wireless communication device selected by the second wireless communication device according to the grouping of candidate wireless communication devices;

The wireless communication device selected by the second wireless communication device according to the transmission complexity of the candidate wireless communication device;

The wireless communication device selected by the second wireless communication device according to the information processing performance of the candidate wireless communication device, where the information processing performance is used to indicate the accuracy of the information output by the local machine learning model on the candidate wireless communication device .
The method according to claim 4, wherein the determining by the first wireless communication device that the first wireless communication device is the valid user comprises:

The first wireless communication device determines that the first wireless communication device is the valid user according to the magnitude relationship between the random number and the probability threshold;

Wherein, the random number is generated by the first wireless communication device.
The method according to claim 3, wherein when the first wireless communication device satisfies the valid user condition, the first wireless communication device is the valid user;

Wherein, the first wireless communication device satisfying valid user conditions includes at least one of the following:

The first wireless communication device is a default active user;

The device capability of the first wireless communication device includes that the first wireless communication device is the valid user.
The method according to claim 4, characterized in that the method further comprises:

When the first wireless communication device determines that the first wireless communication device is the valid user, the first wireless communication device notifies the second wireless communication device that the first wireless communication device is the valid user. valid user.
The method according to claim 3, characterized in that the method further comprises:

When the first wireless communication device is the valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device; the first wireless communication device receives the second wireless The designated uplink transmission resource configured by the communication device;

or,

If the first wireless communication device is the valid user, the first wireless communication device receives the designated uplink transmission resource configured by the second wireless communication device.
The method according to claim 9, wherein when the first wireless communication device is the valid user, the first wireless communication device applies for an uplink transmission resource to the second wireless communication device ,include:

In a case where the first wireless communication device is the valid user, if a trigger event is detected, the first wireless communication device applies to the second wireless communication device for the uplink transmission resource.
The method according to claim 10, wherein the trigger event includes at least one of the following:

In the case where the local machine learning model is used for channel state information CSI feedback, the CSI to be compressed and the restored CSI do not satisfy the first condition;

In the case where the local machine learning model is used for the CSI feedback, the transmission state of the result of the first wireless communication based on the CSI feedback does not meet the second condition;

In the case where the local machine learning model is used for channel estimation, the channel estimation performance of the first wireless communication device does not meet the third condition;

In the case where the local machine learning model is used for the channel estimation, the transmission state of the first wireless communication device based on the result of the channel estimation does not meet the fourth condition;

In the case where the local machine learning model is used for positioning, the positioning accuracy of the first wireless communication device does not meet the fifth condition;

In the case where the local machine learning model is used for beam management, the beam management accuracy of the first wireless communication device does not meet the sixth condition;

In a case where the local machine learning model is used for the beam management, the transmission state of the first wireless communication device based on the result of the beam management does not satisfy the seventh condition.
The method according to claim 4, wherein when the first wireless communication device is indicated as the valid user by the second wireless communication device, the first wireless communication device acts as the valid user. The duration of the user includes at least one of the following:

From when the first wireless communication device is indicated as the valid user by the second wireless communication device to when the first wireless communication device sends the updated local machine learning model to the second wireless communication device duration of time;

The time period between when the first wireless communication device is indicated as the valid user by the second wireless communication device to the end of the preset time period;

the time period between when the first wireless communication device is indicated as the valid user by the second wireless communication device to the end of a preset number of times, the preset number of times is for the first wireless communication device to the The number of times the second wireless communication device sends the updated local machine learning model is set;

The time period between when the first wireless communication device is indicated as the valid user by the second wireless communication device and when the first wireless communication device is not indicated as the valid user by the second wireless communication device .
The method according to any one of claims 1 to 12, wherein the first wireless communication device updates the local machine learning model based on local data, comprising:

The first wireless communication device updates at least one of coefficients and gradient information of the local machine learning model based on local data.
The method according to any one of claims 1 to 12, wherein the second wireless communication device is a terminal; the first wireless communication device sends an updated local machine learning model to the second wireless communication device, including :

The first wireless communication device sends the updated local machine learning model to the second wireless communication device through a transmission resource of a sidelink.
The method according to any one of claims 1 to 12, further comprising:

The first wireless communication device receives the updated global machine learning model sent by the second wireless communication device;

The first wireless communication device continues to perform the steps of updating the machine learning model and sending the updated machine learning model based on the updated global machine learning model.
The method according to any one of claims 1 to 12, wherein the local machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management.
The method according to any one of claims 1 to 12, wherein the downlink transmission resource used by the second wireless communication device to send a message to the first wireless communication device belongs to at least one of the following:

broadcast message;

paging;

Radio Resource Control RRC message;

Media Access Control Control Element MAC CE;

Downlink Control Information DCI;

transmission resources carrying downlink control transmission;

Transmission resources for carrying downlink data transmission;

Transmission resources for carrying downlink artificial intelligence control transmissions;

The transmission resource that carries the downlink artificial intelligence data transmission.
The method according to any one of claims 3 to 12, wherein the specified uplink transmission resource belongs to at least one of the following:

Uplink Control Information UCI;

RRC message;

transmission resources carrying uplink control transmission;

Transmission resources for carrying uplink data transmission;

Transmission resources carrying uplink artificial intelligence control transmission;

Transmission resources that carry uplink artificial intelligence data transmission.
A method for updating a model for wireless channel processing, characterized in that the method is executed by a second wireless communication device, and the method includes:

The second wireless communication device receives an updated local machine learning model sent by the first wireless communication device, and the updated local machine learning model is obtained by updating the local machine learning model based on local data by the first wireless communication device of;

The second wireless communication device updates a global machine learning model according to the updated local machine learning model.
The method according to claim 19, wherein the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device, comprising:

If the first wireless communication device is a valid user, the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device.
The method according to claim 20, wherein when the first wireless communication device is a valid user, the second wireless communication device receives the update sent by the first wireless communication device After the local machine learning model, including:

If the first wireless communication device is the valid user, the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device through a specified uplink transmission resource.
The method according to claim 21, further comprising:

The second wireless communication device indicates to the first wireless communication device that the first wireless communication device is the valid user.
The method according to claim 22, further comprising:

The second wireless communication device randomly selects a wireless communication device from candidate wireless communication devices as the effective user.
The method according to claim 22, further comprising:

The second wireless communication device selects a wireless communication device as the effective user according to the grouping of candidate wireless communication devices.
The method according to claim 22, further comprising:

The second wireless communication device selects a wireless communication device as the effective user according to the transmission complexity of the candidate wireless communication device.
The method according to claim 22, further comprising:

The second wireless communication device selects the wireless communication device as the effective user according to the information processing performance of the candidate wireless communication device, and the information processing performance is used to indicate the information output by the local machine learning model on the candidate wireless communication device accuracy.
The method according to claim 21, wherein the first wireless communication device is determined by the first wireless communication device as the valid user.
The method according to claim 21, wherein when the first wireless communication device satisfies the valid user condition, the first wireless communication device is the valid user;

Wherein, the first wireless communication device satisfying valid user conditions includes at least one of the following:

The first wireless communication device is a default active user;

The device capability of the first wireless communication device includes that the first wireless communication device is the valid user.
The method according to claim 27, further comprising:

When the first wireless communication device determines that the valid user is the first wireless communication device, the second wireless communication device receives the first wireless communication device sent by the first wireless communication device Notifications for the active user.
The method according to claim 21, further comprising:

When the first wireless communication device is the valid user, the second wireless communication device configures the specified assignment to the first wireless communication device according to the uplink transmission resource application of the first wireless communication device Uplink transmission resources;

or,

In a case where the first wireless communication device is the valid user, the second wireless communication device configures the designated uplink transmission resource to the first wireless communication device.
The method according to claim 22, further comprising:

In the case where the second wireless communication device indicates that the first wireless communication device is the valid user, the second wireless communication device indicates that the first wireless communication device is the user according to the second wireless communication device indicating that the first wireless communication device is From the effective user to the second wireless communication device receiving the updated local machine learning model sent by the first wireless communication device, determining the first wireless communication device as the valid user duration of .
The method according to claim 22, further comprising:

In the case where the second wireless communication device indicates that the first wireless communication device is the valid user, the second wireless communication device indicates that the first wireless communication device is the user according to the second wireless communication device indicating that the first wireless communication device is The duration between the valid user and the end of the preset duration determines the duration of the first wireless communication device as the valid user.
The method according to claim 22, further comprising:

In the case where the second wireless communication device indicates that the first wireless communication device is the valid user, the second wireless communication device indicates that the first wireless communication device is the user according to the second wireless communication device indicating that the first wireless communication device is The valid user, the duration between the end of the preset number of times, determines the duration of the first wireless communication device as the valid user;

Wherein, the preset number of times is set for the number of times that the first wireless communication device sends the updated local machine learning model to the second wireless communication device.
The method according to claim 22, further comprising:

In the case where the second wireless communication device indicates that the first wireless communication device is the valid user, the second wireless communication device indicates that the first wireless communication device is the user according to the second wireless communication device indicating that the first wireless communication device is From the valid user to the time period between the second wireless communication device indicating that the first wireless communication device is not the valid user, determine the duration of the first wireless communication device being the valid user.
The method according to any one of claims 19 to 34, wherein the second wireless communication device updates a global machine learning model according to the updated local machine learning model, comprising:

The second wireless communication device updates at least one of coefficients and gradient information of the global machine learning model according to the updated local machine learning model.
The method according to any one of claims 19 to 34, wherein the second wireless communication device is a terminal; the second wireless communication device receives the updated local machine learning model sent by the first wireless communication device, include:

The second wireless communication device receives the updated local machine learning model sent by the first wireless communication device through a sidelink transmission resource.
The method according to any one of claims 19 to 34, further comprising:

The second wireless communication device sends an updated global machine learning model to the first wireless communication device;

The second wireless communication device receives the re-updated global machine learning model sent by the first wireless communication device.
The method according to any one of claims 19 to 34, wherein the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management.
The method according to any one of claims 19 to 34, wherein the downlink transmission resource used by the second wireless communication device to send a message to the first wireless communication device belongs to at least one of the following:

broadcast message;

paging;

RRC message;

MAC CE;

DCI;

transmission resources carrying downlink control transmission;

Transmission resources for carrying downlink data transmission;

Transmission resources for carrying downlink artificial intelligence control transmissions;

The transmission resource that carries the downlink artificial intelligence data transmission.
The method according to any one of claims 21 to 34, wherein the specified uplink transmission resource belongs to at least one of the following:

UCI;

RRC message;

transmission resources carrying uplink control transmission;

Transmission resources for carrying uplink data transmission;

Transmission resources carrying uplink artificial intelligence control transmission;

Transmission resources that carry uplink artificial intelligence data transmission.
A model update device for wireless channel processing, characterized in that the device includes:

An update module for updating a local machine learning model based on local data;

A sending module, configured to send the updated local machine learning model to the second wireless communication device, and the updated local machine learning model is used to update the global machine learning model.
The device according to claim 41, wherein the sending module is configured to:

If the apparatus is a valid user, sending the updated local machine learning model to the second wireless communication device.
The device according to claim 42, wherein the sending module is configured to:

When the apparatus is the valid user, sending the updated local machine learning model to the second wireless communication device by specifying an uplink transmission resource.
The device according to claim 43, further comprising:

a determining module, configured to determine that the apparatus is the valid user according to an indication of the second wireless communication device;

or,

A determining module, configured to determine that the device is the valid user.
The device according to claim 44, wherein the valid user includes at least one of the following:

A wireless communication device randomly selected by the second wireless communication device from candidate wireless communication devices;

a wireless communication device selected by the second wireless communication device according to the grouping of candidate wireless communication devices;

The wireless communication device selected by the second wireless communication device according to the transmission complexity of the candidate wireless communication device;

The wireless communication device selected by the second wireless communication device according to the information processing performance of the candidate wireless communication device, where the information processing performance is used to indicate the accuracy of the information output by the local machine learning model on the candidate wireless communication device .
The device according to claim 44, wherein the determining module is configured to:

determining that the device is the valid user according to the size relationship between the random number and the probability threshold;

Wherein, the random number is generated by the device.
The device according to claim 43, wherein the device is the valid user when the device satisfies the valid user condition;

Wherein, the device satisfying valid user conditions includes at least one of the following:

said device is a default active user;

The device capabilities of the device include that the device is the active user.
The device according to claim 44, wherein the sending module is configured to:

In case the apparatus determines that the apparatus is the valid user, the second wireless communication device is notified that the apparatus is the valid user.
The device according to claim 43, further comprising:

a sending module, configured to apply for an uplink transmission resource from the second wireless communication device when the device is the valid user; a receiving module, configured to receive the specified uplink configured by the second wireless communication device transfer resources;

or,

A receiving module, configured to receive the designated uplink transmission resource configured by the second wireless communication device when the apparatus is the valid user.
The device according to claim 49, wherein the sending module is configured to:

In a case where the apparatus is the valid user, if a trigger event is detected, apply for the uplink transmission resource to the second wireless communication device.
The device according to claim 50, wherein the trigger event includes at least one of the following:

In the case where the local machine learning model is used for CSI feedback, the CSI to be compressed and the restored CSI do not satisfy the first condition;

In the case where the local machine learning model is used for the CSI feedback, the transmission status of the device based on the result of the CSI feedback does not meet the second condition;

In the case where the local machine learning model is used for channel estimation, the channel estimation performance of the device does not meet the third condition;

In the case where the local machine learning model is used for the channel estimation, the transmission state of the device based on the result of the channel estimation does not meet the fourth condition;

In the case where the local machine learning model is used for positioning, the positioning accuracy of the device does not meet the fifth condition;

In the case where the local machine learning model is used for beam management, the beam management accuracy of the device does not meet the sixth condition;

In a case where the local machine learning model is used for the beam management, the transmission status of the device based on the beam management result does not satisfy the seventh condition.
The device according to claim 44, wherein when the device is indicated by the second wireless communication device as the valid user, the duration of the device as the valid user includes at least one of the following kind:

a period of time between when the apparatus is indicated as the valid user by the second wireless communication device and when the apparatus sends the updated local machine learning model to the second wireless communication device;

a time period between when the device is indicated as the valid user by the second wireless communication device to the end of the preset time period;

The duration between when the device is indicated as the valid user by the second wireless communication device and the end of a preset number of times, the preset number of times is for the device to send the The number of times to update the local machine learning model is set;

The time period between when the apparatus is indicated as the valid user by the second wireless communication device and when the apparatus is not indicated as the valid user by the second wireless communication device.
The device according to any one of claims 41 to 52, wherein the update module is configured to:

Updating at least one of coefficients and gradient information of the local machine learning model based on local data.
The device according to any one of claims 41 to 52, wherein the second wireless communication device is a terminal; the sending module is configured to:

sending the updated local machine learning model to the second wireless communication device via a sidelink transmission resource.
The device according to any one of claims 41 to 52, wherein the device further comprises:

a receiving module, configured to receive the updated global machine learning model sent by the second wireless communication device;

The update module is configured to continue to execute the step of updating the machine learning model based on the updated global machine learning model, and the sending module is configured to execute the step of sending the updated machine learning model.
The device according to any one of claims 41 to 52, wherein the local machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management.
The device according to any one of claims 41 to 52, wherein the downlink transmission resource used by the second wireless communication device to send a message to the device belongs to at least one of the following:

broadcast message;

paging;

RRC message;

MAC CE;

DCI;

transmission resources carrying downlink control transmission;

Transmission resources for carrying downlink data transmission;

Transmission resources for carrying downlink artificial intelligence control transmissions;

The transmission resource that carries the downlink artificial intelligence data transmission.
The device according to any one of claims 43 to 52, wherein the specified uplink transmission resource belongs to at least one of the following:

UCI;

RRC message;

transmission resources carrying uplink control transmission;

Transmission resources for carrying uplink data transmission;

Transmission resources carrying uplink artificial intelligence control transmission;

Transmission resources that carry uplink artificial intelligence data transmission.
A model update device for wireless channel processing, characterized in that the device includes:

A receiving module, configured to receive an updated local machine learning model sent by the first wireless communication device, where the updated local machine learning model is obtained by updating the local machine learning model based on local data by the first wireless communication device;

An update module, configured to update the global machine learning model according to the updated local machine learning model.
The device according to claim 59, wherein the receiving module is configured to:

If the first wireless communication device is a valid user, receive the updated local machine learning model sent by the first wireless communication device.
The device according to claim 60, wherein the receiving module is configured to:

If the first wireless communication device is the valid user, receiving the updated local machine learning model sent by the first wireless communication device through a designated uplink transmission resource.
The device according to claim 61, further comprising:

A sending module, configured to indicate to the first wireless communication device that the first wireless communication device is the valid user.
The device according to claim 62, further comprising:

A determining module, configured to randomly select a wireless communication device from candidate wireless communication devices as the effective user.
The device according to claim 62, further comprising:

A determining module, configured to select a wireless communication device as the effective user according to the grouping of candidate wireless communication devices.
The device according to claim 62, further comprising:

A determining module, configured to select a wireless communication device as the effective user according to the transmission complexity of the candidate wireless communication device.
The device according to claim 62, further comprising:

A determining module, configured to select a wireless communication device as the effective user according to the information processing performance of the candidate wireless communication device, where the information processing performance is used to indicate the accuracy of information output by a local machine learning model on the candidate wireless communication device .
The apparatus according to claim 61, wherein the first wireless communication device is determined by the first wireless communication device as the valid user.
The apparatus according to claim 61, wherein, when the first wireless communication device satisfies the valid user condition, the first wireless communication device is the valid user;

Wherein, the first wireless communication device satisfying valid user conditions includes at least one of the following:

The first wireless communication device is a default valid user;

The device capability of the first wireless communication device includes that the first wireless communication device is the valid user.
The device according to claim 67, wherein the receiving module is configured to:

When the first wireless communication device determines that the valid user is the first wireless communication device, the apparatus receives the first wireless communication device being the valid user from the first wireless communication device User Notifications.
The device according to claim 61, further comprising:

A sending module, configured to configure the specified uplink transmission to the first wireless communication device according to the uplink transmission resource application of the first wireless communication device when the first wireless communication device is the valid user resource;

or,

A sending module, configured to configure the designated uplink transmission resource to the first wireless communication device when the first wireless communication device is the valid user.
The device according to claim 62, further comprising:

a determining module, configured to, when the device indicates that the first wireless communication device is the valid user, according to the device indicating that the first wireless communication device is the valid user, until the device receives The duration between the updated local machine learning models sent by the first wireless communication device determines the duration of the first wireless communication device as the valid user.
The device according to claim 62, further comprising:

A determining module, configured to, when the device indicates that the first wireless communication device is the valid user, according to the device indicating that the first wireless communication device is the valid user, until the end of the preset time period Determine the duration of the first wireless communication device as the valid user.
The device according to claim 62, further comprising:

A determining module, configured to, when the device indicates that the first wireless communication device is the valid user, according to the device indicating that the first wireless communication device is the valid user, until the preset number of times ends The length of time between, determine the duration of the first wireless communication device as the valid user;

Wherein, the preset number of times is set for the number of times that the first wireless communication device sends the updated local machine learning model to the apparatus.
The device according to claim 62, further comprising:

a determining module, configured to, when the device indicates that the first wireless communication device is the valid user, according to the device indicating that the first wireless communication device is the valid user, until the device indicates The duration between which the first wireless communication device is not the valid user is determined, and the duration of the first wireless communication device being the valid user is determined.
The device according to any one of claims 59 to 74, wherein the update module is configured to:

At least one of coefficients and gradient information of the global machine learning model is updated according to the updated local machine learning model.
The device according to any one of claims 59 to 74, wherein the device is a terminal; the receiving module is configured to:

Receive the updated local machine learning model sent by the first wireless communication device through the transmission resource of the sidelink.
The device according to any one of claims 59 to 74, wherein the device further comprises:

a sending module, configured to send the updated global machine learning model to the first wireless communication device;

The receiving module is configured to receive the re-updated global machine learning model sent by the first wireless communication device.
The device according to any one of claims 59 to 74, wherein the global machine learning model is used for at least one of CSI feedback, channel estimation, positioning scheme and beam management.
The device according to any one of claims 59 to 74, wherein the downlink transmission resource used by the device to send a message to the first wireless communication device belongs to at least one of the following:

broadcast message;

paging;

RRC message;

MAC CE;

DCI;

transmission resources carrying downlink control transmission;

Transmission resources for carrying downlink data transmission;

Transmission resources for carrying downlink artificial intelligence control transmissions;

The transmission resource that carries the downlink artificial intelligence data transmission.
The device according to claim 61-704, wherein the specified uplink transmission resource belongs to at least one of the following:

UCI;

RRC message;

transmission resources carrying uplink control transmission;

Transmission resources for carrying uplink data transmission;

Transmission resources carrying uplink artificial intelligence control transmission;

Transmission resources that carry uplink artificial intelligence data transmission.
A terminal, characterized in that the terminal includes:

processor;

a transceiver connected to the processor;

memory for storing executable instructions of the processor;

Wherein, the processor is configured to load and execute the executable instructions to implement the model update method for wireless channel processing according to any one of claims 1 to 18.
A network device, characterized in that the network device includes:

processor;

a transceiver connected to the processor;

memory for storing executable instructions of the processor;

Wherein, the processor is configured to load and execute the executable instructions to implement the model update method for wireless channel processing according to any one of claims 19 to 40.
A computer-readable storage medium, characterized in that executable instructions are stored in the readable storage medium, and the executable instructions are loaded and executed by a processor to implement the method described in any one of claims 1 to 40 A model update method for wireless channel processing.
A chip, characterized in that the chip includes a programmable logic circuit or program, and the chip is used to implement the model update method for wireless channel processing according to any one of claims 1 to 40.
A computer program product, characterized in that the computer program product includes computer instructions, the computer instructions are stored in a computer-readable storage medium, and a processor of a computer device reads the computer program from the computer-readable storage medium. Instructions, the processor executes the computer instructions, so that the computer device executes the model updating method for wireless channel processing according to any one of claims 1 to 40.