CN117813801A

CN117813801A - Communication methods, model training methods and equipment

Info

Publication number: CN117813801A
Application number: CN202180101261.0A
Authority: CN
Inventors: 田文强
Original assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd
Current assignee: Guangdong Oppo Mobile Telecommunications Corp Ltd
Priority date: 2021-11-23
Filing date: 2021-11-23
Publication date: 2024-04-02
Also published as: WO2023092307A1

Abstract

The embodiment of the application relates to a communication method, a model training method and a device, wherein the communication method comprises the steps that a terminal device receives a first signal, and the first signal is generated by a first model; the terminal equipment processes the first signal by adopting a second model to obtain first information, wherein the first information comprises channel information; the first model and the second model are obtained through combined training. The embodiment of the application can improve the overall performance of the network.

Description

Communication method, model training method and device

Technical Field

The present application relates to the field of communications, and more particularly, to a communication method, model training method, and apparatus.

Background

There are uplink and downlink reference signals in a wireless communication system, and these reference signals are used to achieve different purposes such as channel estimation. However, these reference signals are not considered to be applied to the wireless communication solutions based on the artificial intelligence (AI, artificial Intelligence) or neural network methods when they are designed, so that it is difficult to achieve the best matching result of the existing reference signals with the wireless communication solutions based on the AI or neural network methods. It can be seen how to implement an AI-based wireless communication solution and an adapted reference signal design as one overall solution to improve the overall advantages of the reference signal design and the wireless communication solution, which is a problem to be solved.

Disclosure of Invention

The embodiment of the application provides a communication method, a model training method and equipment, which can improve the overall advantages of reference signal design and a wireless communication solution.

The embodiment of the application provides a communication method, which comprises the following steps:

the terminal equipment receives a first signal, wherein the first signal is generated by a first model;

the terminal equipment processes the first signal by adopting a second model to obtain first information, wherein the first information comprises channel information;

wherein the first model and the second model are obtained by joint training.

The embodiment of the application also provides a communication method, which comprises the following steps:

the network device transmits a first signal, the first signal being generated by a first model; the first signal is used for being processed by the second model to obtain first information, and the first information comprises channel information;

wherein the first model and the second model are obtained through joint training.

The embodiment of the application also provides a model training method, which comprises the following steps:

and carrying out joint training on the first initial model and the second initial model by adopting input information and/or a first channel simulation module to obtain a trained first model and trained second model.

The embodiment of the application also provides a terminal device, which comprises:

A first receiving module for receiving a first signal, the first signal being generated by a first model;

the first processing module is used for processing the first signal by adopting a second model to obtain first information, wherein the first information comprises channel information;

The embodiment of the application also provides a network device, which comprises:

a sixth transmitting module for transmitting a first signal, the first signal being generated by the first model; the first signal is used for being processed by the second model to obtain first information, and the first information comprises channel information;

The embodiment of the application also provides model training equipment, which comprises:

and the joint training module is used for jointly training the first initial model and the second initial model by adopting the input information and/or the first channel simulation module to obtain a trained first model and trained second model.

The embodiment of the application also provides a terminal device, which comprises: the system comprises a processor, a memory and a transceiver, wherein the memory is used for storing a computer program, and the processor is used for calling and running the computer program stored in the memory and controlling the transceiver to execute the method according to any one of the first communication methods.

The embodiment of the application also provides a network device, which comprises: the system comprises a processor, a memory and a transceiver, wherein the memory is used for storing a computer program, and the processor is used for calling and running the computer program stored in the memory and controlling the transceiver to execute the method according to any one of the second communication methods.

The embodiment of the application also provides model training equipment, which comprises: a processor and a memory for storing a computer program, the processor being adapted to invoke and run the computer program stored in the memory to perform the method according to any of the model training methods described above.

The embodiment of the application also provides a chip, which comprises: a processor for calling and running a computer program from a memory, causing a device on which the chip is mounted to perform the method according to any one of the first communication methods described above.

The embodiment of the application also provides a chip, which comprises: a processor for calling and running a computer program from a memory, causing a device on which the chip is mounted to perform the method according to any of the second communication methods described above.

The embodiment of the application also provides a chip, which comprises: a processor for calling and running a computer program from memory, causing a device on which the chip is mounted to perform the method as claimed in any one of the model training methods above.

The embodiments of the present application also propose a computer-readable storage medium storing a computer program for causing a computer to execute the method according to any one of the first communication methods described above.

The embodiment of the application also proposes a computer-readable storage medium for storing a computer program that causes a computer to execute the method according to any one of the second communication methods described above.

The embodiment of the application also proposes a computer readable storage medium for storing a computer program for causing a computer to perform the method according to any one of the model training methods described above.

The embodiments of the present application also propose a computer program product comprising computer program instructions for causing a computer to perform the method according to any one of the first communication methods described above.

The embodiments of the present application also propose a computer program product comprising computer program instructions for causing a computer to perform the method according to any one of the second communication methods described above.

The embodiments of the present application also propose a computer program product comprising computer program instructions for causing a computer to perform the method according to any one of the model training methods described above.

The embodiment of the application also proposes a computer program, which causes a computer to execute the method according to any one of the first communication methods described above.

The embodiment of the application also proposes a computer program, which makes a computer execute the method according to any one of the second communication methods.

The embodiment of the application also proposes a computer program, which makes a computer execute the method according to any one of the model training methods.

By adopting the embodiment of the application, the terminal equipment processes the received first signal by adopting the second model, the first signal is generated by the first model, and the first model and the second model are obtained by combined training, and as the first model and the second model are obtained by combined training, the performance requirements in the whole signal generation and the signal processing can be considered, and the overall performance of the network can be improved.

Drawings

Fig. 1 is a schematic diagram of an application scenario in an embodiment of the present application.

Fig. 2 is a schematic diagram of a neural network structure.

Fig. 3A is a schematic diagram of a CSI feedback scheme.

Fig. 3B is a schematic diagram of a way of performing channel estimation.

Fig. 3C is a schematic diagram of a positioning method.

Fig. 4 is a schematic flow chart of a communication method 400 according to an embodiment of the present application.

Fig. 5 is a schematic flow chart diagram of another communication method 500 according to an embodiment of the present application.

Fig. 6 is a schematic diagram of a model transmission mode according to an embodiment of the present application.

Fig. 7 is a schematic diagram of a model structure according to the present application.

Fig. 8 is a schematic diagram of another model transmission mode according to an embodiment of the present application.

Fig. 9 is a schematic diagram of a design scheme integrating AI-based wireless communication system multi-user reference signals and channel estimation according to an embodiment of the present application.

Fig. 10 is a schematic diagram of a channel information structure according to the present application.

Fig. 11 is a schematic diagram of another channel information structure according to the present application.

Fig. 12 is a schematic diagram of a structure of channel eigenvector information according to the present application.

Fig. 13 is a schematic diagram of a neural network structure according to the present application.

Fig. 14 is a schematic diagram of model structure and information transmission in an integrated design of multi-user reference signals and channel estimation for an AI-based wireless communication system according to an embodiment of the present application.

Fig. 15 is a schematic diagram of an integrated design of multi-user reference signal, channel estimation, and channel information feedback for an AI-based wireless communication system according to an embodiment of the present application.

Fig. 16 is a schematic diagram of another neural network structure according to the present application.

Fig. 17 is a schematic diagram of model structure and information transmission in an integrated design of multi-user reference signal, channel estimation and channel information feedback for an AI-based wireless communication system according to an embodiment of the present application.

Fig. 18 is a schematic flow chart diagram of another communication method 1800 according to an embodiment of the present application.

Fig. 19 is a schematic flow chart diagram of another communication method 1900 according to an embodiment of the present application.

Fig. 20 is a schematic flow chart diagram of another model training method 2000 according to an embodiment of the present application.

Fig. 21 is a schematic structural diagram of a terminal device 2100 according to an embodiment of the present application.

Fig. 22 is a schematic structural diagram of a network device 2200 according to an embodiment of the present application.

Fig. 23 is a schematic structural view of a model training apparatus 2300 according to an embodiment of the application.

FIG. 24 is a schematic block diagram of a communication device or model training device 700 according to an embodiment of the present application;

fig. 25 is a schematic structural diagram of a chip 800 according to an embodiment of the present application.

Detailed Description

The technical solutions in the embodiments of the present application will be described below with reference to the drawings in the embodiments of the present application.

It should be noted that the terms "first," "second," and the like in the description and the claims of the embodiments of the present application and the above figures are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. The objects described as "first" and "second" may be the same or different.

The technical solution of the embodiment of the application can be applied to various communication systems, for example: global system for mobile communications (Global System of Mobile communication, GSM), code division multiple access (Code Division Multiple Access, CDMA) system, wideband code division multiple access (Wideband Code Division Multiple Access, WCDMA) system, general packet Radio service (General Packet Radio Service, GPRS), long term evolution (Long Term Evolution, LTE) system, long term evolution advanced (Advanced long term evolution, LTE-a) system, new Radio (NR) system, evolution system of NR system, LTE-based access to unlicensed spectrum, LTE-U) system over unlicensed spectrum, NR (NR-based access to unlicensed spectrum, NR-U) system over unlicensed spectrum, universal mobile communication system (Universal Mobile Telecommunication System, UMTS), wireless local area network (Wireless Local Area Networks, WLAN), wireless fidelity (Wireless Fidelity, wiFi), next Generation communication (5 th-Generation, 5G) system, or other communication system, etc.

Generally, the number of connections supported by the conventional communication system is limited and easy to implement, however, with the development of communication technology, the mobile communication system will support not only conventional communication but also, for example, device-to-Device (D2D) communication, machine-to-machine (Machine to Machine, M2M) communication, machine type communication (Machine Type Communication, MTC), inter-vehicle (Vehicle to Vehicle, V2V) communication, and the like, to which the embodiments of the present application can also be applied.

Optionally, the communication system in the embodiment of the present application may be applied to a carrier aggregation (Carrier Aggregation, CA) scenario, a dual connectivity (Dual Connectivity, DC) scenario, and a Stand Alone (SA) fabric scenario.

The frequency spectrum of the application in the embodiments of the present application is not limited. For example, embodiments of the present application may be applied to licensed spectrum as well as unlicensed spectrum.

Embodiments of the present application describe various embodiments in connection with a network device and a terminal device, wherein: a terminal device may also be called a User Equipment (UE), an access terminal, a subscriber unit, a subscriber station, a mobile station, a remote terminal, a mobile device, a User terminal, a wireless communication device, a User agent, a User device, or the like. The terminal device may be a Station (ST) in a WLAN, may be a cellular telephone, a cordless telephone, a session initiation protocol (Session Initiation Protocol, SIP) phone, a wireless local loop (Wireless Local Loop, WLL) station, a personal digital assistant (Personal Digital Assistant, PDA) device, a handheld device with wireless communication functionality, a computing device or other processing device connected to a wireless modem, a vehicle mounted device, a wearable device, and a next generation communication system, such as a terminal device in an NR network or a terminal device in a future evolved public land mobile network (Public Land Mobile Network, PLMN) network, etc.

By way of example, and not limitation, in embodiments of the present application, the terminal device may also be a wearable device. The wearable device can also be called as a wearable intelligent device, and is a generic name for intelligently designing daily wear by applying wearable technology and developing wearable devices, such as glasses, gloves, watches, clothes, shoes and the like. The wearable device is a portable device that is worn directly on the body or integrated into the clothing or accessories of the user. The wearable device is not only a hardware device, but also can realize a powerful function through software support, data interaction and cloud interaction. The generalized wearable intelligent device includes full functionality, large size, and may not rely on the smart phone to implement complete or partial functionality, such as: smart watches or smart glasses, etc., and focus on only certain types of application functions, and need to be used in combination with other devices, such as smart phones, for example, various smart bracelets, smart jewelry, etc. for physical sign monitoring.

The network device may be a device for communicating with the mobile device, the network device may be an Access Point (AP) in WLAN, a base station (Base Transceiver Station, BTS) in GSM or CDMA, a base station (NodeB, NB) in WCDMA, an evolved base station (Evolutional Node B, eNB or eNodeB) in LTE, or a relay station or an Access Point, or a vehicle device, a wearable device, and a network device (gNB) in NR network, or a network device in future evolved PLMN network, etc.

In the embodiment of the present application, the network device provides services for a cell, and the terminal device communicates with the network device through a transmission resource (for example, a frequency domain resource, or a spectrum resource) used by the cell, where the cell may be a cell corresponding to the network device (for example, a base station), and the cell may belong to a macro base station, or may belong to a base station corresponding to a Small cell (Small cell), where the Small cell may include: urban cells (Metro cells), micro cells (Micro cells), pico cells (Pico cells), femto cells (Femto cells) and the like, and the small cells have the characteristics of small coverage area and low transmitting power and are suitable for providing high-rate data transmission services.

Fig. 1 illustrates one network device 110 and two terminal devices 120, alternatively, the wireless communication system 100 may include a plurality of network devices 110, and each network device 110 may include other numbers of terminal devices 120 within a coverage area, which is not limited in this embodiment. The embodiment of the present application may be applied to one terminal device 120 and one network device 110, and may also be applied to one terminal device 120 and another terminal device 120.

Optionally, the wireless communication system 100 may further include other network entities such as a mobility management entity (Mobility Management Entity, MME), an access and mobility management function (Access and Mobility Management Function, AMF), and the like, which are not limited in this embodiment of the present application.

It should be understood that the terms "system" and "network" are used interchangeably herein. The term "and/or" is herein merely an association relationship describing an associated object, meaning that there may be three relationships, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone. In addition, the character "/" herein generally indicates that the front and rear associated objects are an "or" relationship.

It should be understood that, in the embodiments of the present application, the "indication" may be a direct indication, an indirect indication, or an indication having an association relationship. For example, a indicates B, which may mean that a indicates B directly, e.g., B may be obtained by a; it may also indicate that a indicates B indirectly, e.g. a indicates C, B may be obtained by C; it may also be indicated that there is an association between a and B.

In the description of the embodiments of the present application, the term "corresponding" may indicate that there is a direct correspondence or an indirect correspondence between the two, or may indicate that there is an association between the two, or may indicate a relationship between the two and the indicated, configured, or the like.

In order to facilitate understanding of the technical solutions of the embodiments of the present application, the following description is given of related technologies of the embodiments of the present application, and the following related technologies may be optionally combined with the technical solutions of the embodiments of the present application as an alternative, which all belong to the protection scope of the embodiments of the present application. It should be understood that the basic flow and basic concepts described below are not limiting on the embodiments of the present application.

There are many designs for reference signals in the current wireless communication system, such as downlink reference signals including downlink demodulation reference signals (DMRS, demodulation Reference Signal), channel state information reference signals (CSI-RS, channel State Information Reference Signal), downlink phase tracking reference signals (PT-RS, phase Tracking Reference Signal), positioning reference signals (PRS, positioning Reference Signal), etc., and uplink reference signals including sounding reference signals (SRS, sounding Reference Signal), uplink DMRS, uplink PT-RS, etc. These reference signals are designed to perform various tasks such as channel estimation, phase tracking, positioning, etc.

In recent years, artificial intelligence research represented by neural networks has achieved very great results in many fields, and will also play an important role in the production and life of people for a long time in the future.

A simple neural network basic structure includes: input layer, hidden layer and output layer, as shown in fig. 2. The input layer is responsible for receiving data, the hidden layer processes the data, and the final result is generated at the output layer. In this case, each node represents a processing unit, which can be considered to simulate a neuron, and a plurality of neurons form a neural network, and the information transmission and processing of the plurality of layers form an integral neural network.

With the continuous development of the neural network research, a neural network deep learning algorithm is proposed in recent years, more hidden layers are introduced, and feature learning is performed through layer-by-layer training of the neural network with multiple hidden layers, so that the learning and processing capacity of the neural network is greatly improved, and the neural network deep learning algorithm is widely applied in the aspects of pattern recognition, signal processing, optimization combination, anomaly detection and the like.

The combination of a neural network and a wireless communication system is a current research direction, and there are many works related to applying the neural network to wireless communication problems such as channel estimation, phase tracking, positioning, beam management and the like, but source information relied on by the works is still various existing reference signals.

For example, in the CSI feedback problem, as shown in fig. 3A, CSI information obtained by acquiring and processing a current CSI-RS reference signal is recovered at the base station side after AI encoding and decoding. In the channel estimation problem, as shown in fig. 3B, the UE may implement high performance estimation of a given channel through an AI channel estimator based on the current CSI-RS, or DMRS. In the positioning problem, as shown in fig. 3C, the UE may acquire corresponding channel information through the current PRS, and then rely on positioning the channel information to acquire a high-precision positioning result through an AI-based positioning algorithm.

For wireless communication systems, current reference signals are not designed to be used later in AI-based, neural network approach wireless communication solutions. When the existing reference signals are directly applied to various wireless communication solutions based on AI and neural network methods, only the existing reference signals are borrowed, and a wireless communication system solution constructed under the condition of relying on the existing reference signals is provided.

This problem is questionable as to whether the existing reference signals are in optimal matching relation with AI-based, neural network-based wireless communication solutions, because the reference signal designs and wireless communication solutions herein are designed independently and it is difficult to achieve optimal matching results.

In addition, from another perspective, existing reference signals are designed to take into account the reference signal design of a common scenario rather than specific scenarios. For the wireless communication solutions based on AI and neural networks, the dependence on environment and scene related data in the technical construction process is considered, and the solutions are often based on scene optimization. In this case, the design of the existing reference signal is initially inconsistent with the design of AI-based wireless communication solutions for the problems of scene optimization and scene suitability.

In summary, the present solution considers that it is necessary to construct reference signal designs and functional matches in a wireless communication system based on AI technology, and the above-mentioned reference signal designs need to be compatible with the optimization of AI performance and the requirements of multi-user applications.

The present embodiment proposes a communication method, and fig. 4 is a schematic flow chart of a communication method 400 according to the present embodiment, and the method may alternatively be applied to the system shown in fig. 1, but is not limited thereto. The method includes at least some of the following.

S410: the terminal equipment receives a first signal, wherein the first signal is generated by a first model;

s420: the terminal equipment processes the first signal by adopting a second model to obtain first information, wherein the first information comprises channel information;

wherein the first model and the second model are obtained by joint training.

Alternatively, the first signal may include a reference signal. The reference signal is an uplink reference signal or a downlink reference signal, for example, a downlink DMRS, CSI-RS, downlink PT-RS, PRS, or other signals.

In step S410, the terminal device may receive a first signal from the network device, which may be generated by the network device using a first model.

Optionally, the second model may include a channel estimation sub-model;

the channel estimation sub-model is used for carrying out channel estimation based on a first signal (such as a reference signal) to obtain channel information.

The channel information may be used to characterize a channel quality, a channel state, or a channel estimation result obtained by performing channel estimation based on the first information.

The network device may be an access network device (such as a base station, eNB, or gNB) serving the terminal device, or may be an access network device that communicates with the terminal device, such as a base station, eNB, or gNB.

Another communication method is proposed in the present embodiment, and fig. 5 is a schematic flow chart of another communication method 500 according to the present embodiment, and the method may alternatively be applied to the system shown in fig. 1, but is not limited thereto. As shown in fig. 5, after the above step S420, it may further include:

s530: the terminal equipment processes the first information by adopting a third model to obtain second information;

s540: the terminal equipment sends the second information, and the second information is used for being processed by a fourth model to obtain third information;

the first model, the second model, the third model and the fourth model are obtained through combined training.

Optionally, the third model includes a compressible sub-model; the compression sub-model is used for compressing first information (such as channel information) to obtain compressed information of the first information; accordingly, the second information includes compressed information of the first information.

Accordingly, the fourth model may include a restorative sub-model; the recovery sub-model is used for carrying out recovery processing on the compressed information of the first information to obtain recovery information of the first information; accordingly, the third information includes recovery information of the first information.

With the above model, in step S530, the terminal device may compress the first information (e.g., channel information) with the compression sub-model to obtain the second information (e.g., compressed information of the channel information); in step S540, the terminal device may send the compressed information of the channel information to the network device, where the network device may be the network device that sent the first signal (such as the reference signal) before; and then, the network equipment can recover the compressed information of the channel information by adopting a recovery sub-model to obtain third information. The third model and the fourth model form a channel information feedback model.

Alternatively, the third model includes a generation sub-model and a compression sub-model; the generating sub-model is used for carrying out feature transformation on first information (such as channel information) to obtain a first feature vector corresponding to the first information; the compression sub-model is used for compressing the first feature vector to obtain compression information of the first feature vector; the second information includes compressed information of the first feature vector.

Correspondingly, the fourth model comprises a recovery sub-model; the restoration sub-model is used for restoring the compressed information of the first feature vector to obtain restoration information of the first feature vector; the third information includes restoration information of the first feature vector.

With the above model, in step S530, the terminal device may generate the first feature vector of the first information by using the generating sub-model in the third model, and then compress the first feature vector by using the compressing sub-model in the third model to obtain the second information (such as compressed information of the first feature vector); in step S540, the terminal device may send the compressed information of the first feature vector to the network device, where the network device may be the network device that sent the first signal (such as the reference signal) before; and then, the network equipment can recover the compressed information of the first feature vector by adopting a recovery sub-model to obtain third information. The third model and the fourth model form a channel information feedback model.

The first model may also be referred to as a signal generation sub-model, and is used to generate reference signals for multiple users, such as to generate a reference signal set containing multiple reference signals, where the reference signals in the reference signal set may be used by multiple terminal devices (e.g., terminal devices belonging to the same cell or served by the same access device).

The second model may also be referred to as a channel estimation sub-model, and is used for performing channel estimation based on the reference signal to obtain channel information.

The generating sub-model may also be referred to as a channel information generating sub-model, and is configured to process the received channel information by using a data transformation method or the like, to obtain a channel characteristic vector of the channel information, for example, using a singular value decomposition (SVD, singular Value Decomposition) method to obtain a channel information characteristic vector.

The compression sub-model may also be referred to as a channel information compression sub-model for compressing the received channel information or channel information feature vectors. The recovery sub-model may also be referred to as a channel information recovery sub-model for recovering the received compressed information.

The signal generation sub-model, the channel estimation sub-model, the channel information generation sub-model, the channel information compression sub-model or the recovery sub-model can be formed by one or more of a full connection network, a convolutional neural network, a residual network and a self-attention mechanism network.

It can be seen that when the signal generation sub-model and the channel estimation sub-model are jointly trained, the integrated design of multi-user reference signals and channel estimation of the wireless communication system based on AI is realized; when the signal generation sub-model, the channel estimation sub-model, the channel information generation sub-model, the channel information compression sub-model and the recovery sub-model are jointly trained, or the signal generation sub-model, the channel estimation sub-model, the channel information compression sub-model and the recovery sub-model are jointly trained, the integrated design of multi-user reference signals, channel estimation and channel information feedback of the wireless communication system based on AI is realized.

The training of the model can be accomplished by the terminal device or by the network device. Accordingly, there are at least two model training and transmission modes:

mode one:

the training process is completed by the network device (such as a base station), and the network device can send all or part of the trained signal generation sub-model, the channel estimation sub-model, the channel information generation sub-model, the channel information compression sub-model and the channel information recovery sub-model to the terminal device. In addition, the network device may send the first channel model module and/or the second channel simulation module to the terminal device, where the first channel model module and the second channel simulation module are respectively configured to simulate an influence of the downlink transmission channel and the uplink transmission channel on the signal.

Or, the network device may send the trained channel estimation sub-model, the channel information generation sub-model and the channel information compression sub-model to the terminal device as an integral channel estimation and channel information feedback module, for example, as a first coding model; alternatively, the network device may send the trained channel estimation sub-model and the channel information compression sub-model to the terminal device as an integral channel estimation and channel information feedback module, e.g. as a second coding model. Or, the network device may further send the trained channel information recovery submodel to the UE as a decoding module. The model transmission mode is shown in fig. 6.

The above modes of each model, sub-model and module of the network device transmission can be completed in independent transmission or can be completed in non-independent transmission (for example, all the information is transmitted through one signaling and message).

Since one network device may serve multiple terminal devices, the network device may send the above-described models, sub-models, and/or modules to all (or at least some) of the terminal devices it serves. Taking the network device as a base station and the terminal device as a mobile terminal as an example, for example, the base station a serves the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4, the base station a may send the above models, submodels and/or modules to the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4, respectively.

Fig. 7 is a schematic diagram of two model structures according to the present application. Referring to fig. 7, a terminal device may receive a channel estimation sub-model and use the channel estimation sub-model for channel estimation based on a received reference signal. Alternatively, the terminal device may receive the channel estimation sub-model and the channel information compression sub-model, respectively, and use the sub-models for channel estimation and channel information compression, respectively. Alternatively, the terminal device may receive the channel estimation sub-model, the channel information generation sub-model, and the channel information compression sub-model, respectively, and use the sub-models for channel estimation, generation of the channel information feature vector, and channel information compression, respectively.

Alternatively, the terminal device may receive the signal generation sub-model, the first channel simulation module, and the channel estimation sub-model, respectively, with which the terminal device may evaluate the performance of the signal generation sub-model and the channel estimation sub-model. For example, the terminal device generates a reference signal by using a signal generation sub-model, and processes the reference signal by using the first channel simulation module to simulate the reference signal received by the terminal device through a downlink channel; then, channel estimation is carried out by utilizing a channel estimation sub-model based on the signal processed by the first channel simulation module, so as to obtain a channel estimation result; and comparing the channel estimation result with parameters of the first channel simulation module, and evaluating the performances of the signal generation sub-model and the channel estimation sub-model based on the comparison result and the quality of the reference signal generated by the signal generation sub-model and/or the quality of the reference signal obtained after processing by the first channel simulation module. After the evaluation is qualified, the terminal equipment can adopt a channel estimation sub-model to carry out channel estimation based on the truly received reference signals. In addition, the first channel simulation module may be stored in the terminal device in advance, in which case the terminal device does not need to receive the first channel simulation module from the network device. In addition, the first channel simulation module can be simplified into an identity matrix, and in this case, after the first channel simulation module is adopted to process the reference signal, the obtained signal is the same as the reference signal; that is, it is simulated that the reference signal does not change in the downlink transmission process, and the reference signal received by the terminal device through the downlink channel is the same as the reference signal sent by the network device.

Alternatively, the terminal device may receive the signal generating sub-model, the first channel simulation module, the channel estimation sub-model, the channel information generating sub-model, the channel information compressing sub-model, the second channel simulation module, and the channel information recovering sub-model, respectively, and with these sub-models/modules, the terminal device may evaluate the performance of the signal generating sub-model, the channel estimation sub-model, the channel information generating sub-model, the channel information compressing sub-model, and the channel information recovering sub-model. For example, the terminal device generates a reference signal by using a signal generation sub-model, and processes the reference signal by using the first channel simulation module to simulate the reference signal received by the terminal device through a downlink channel; then, channel estimation is carried out by utilizing a channel estimation sub-model based on the signal processed by the first channel simulation module, so as to obtain a channel estimation result (such as channel information); processing the channel information by adopting a channel information generation sub-model to obtain vector characteristics of the channel information; compressing vector features of the channel information by adopting a channel information compression sub-model to obtain compressed channel information; processing the compressed channel information by using a second channel simulation module to simulate the compressed channel information received by the network equipment through an uplink channel; and processing the result processed by the second channel simulation module by using the channel information recovery sub-model so as to simulate the channel information obtained after the network equipment recovers the channel information. And then, continuously comparing the information output by the channel information recovery sub-model with the information input by the channel information compression sub-model, and evaluating the performances of the signal generation sub-model, the channel estimation sub-model, the channel information generation sub-model, the channel information compression sub-model and the channel information recovery sub-model based on the comparison result and the quality of the reference signal generated by the signal generation sub-model and/or the quality of the reference signal obtained after processing by the first channel simulation module (further based on the comparison result of the channel estimation result and the parameters of the first channel simulation module). After the evaluation is qualified, the terminal equipment can adopt the channel estimation sub-model to carry out channel evaluation based on the truly received reference signal, adopts the channel information generation sub-model and the channel information compression sub-model to compress information obtained after the evaluation, and sends the compressed channel information to the network equipment. In addition, the first channel simulation module and/or the second channel simulation module may be stored in the terminal device in advance, in which case the terminal device does not need to receive the first channel simulation module and/or the second channel simulation module from the network device. In addition, the first channel simulation module can be simplified into an identity matrix, and in this case, after the first channel simulation module is adopted to process the reference signal, the obtained signal is the same as the reference signal; that is, it is simulated that the reference signal does not change in the downlink transmission process, and the reference signal received by the terminal device through the downlink channel is the same as the reference signal sent by the network device. Similarly, the second channel simulation module may be simplified into an identity matrix, where the second channel simulation module is used to process the compressed channel information, and the obtained result is the same as the compressed channel information; that is, it is simulated that the compressed channel information does not change in the uplink transmission process, and the compressed channel information received by the network device through the uplink channel is the same as the compressed channel information sent by the terminal device.

In addition, the channel information generation sub-model is an optional model, and the channel information compression sub-model directly compresses the channel information generated by the channel estimation sub-model under the condition that the channel information generation sub-model does not exist, so as to obtain the compressed channel information. In addition, the processing manner of the other models/modules in the processes of receiving, evaluating, using and the like is consistent with that described above, and will not be described again here.

If the overall evaluation result of the received model/sub-model by the terminal device is poor (such as lower quality of reference signal or lower accuracy of channel estimation, etc.), the received model/sub-model may not be used, but the terminal device may perform joint training on the model/sub-model again by itself to update the model parameters of the model/sub-model, or the terminal device may perform training by itself to obtain a new model/sub-model. After the terminal equipment acquires a new model/sub-model after the joint training is carried out again or updated, the model/sub-model can be synchronized to the network equipment; accordingly, after receiving the new model/sub-model, the network device may replace the model/sub-model that was originally trained, and may also synchronize the new model/sub-model to other terminal devices. With respect to the other related processes thereafter, this embodiment is not listed one by one. Through the processing, the model/sub-model with optimal use performance in the whole communication system can be ensured, so that the overall performance of the whole system is further improved.

The above method of receiving the submodel/module by the terminal device may be by one or more of the following means: downlink Control signaling, media access Control (MAC, medica Access Control) Control Element (CE) messages, radio resource Control (RRC, radio Resource Control) messages, broadcast, downlink data transmission for artificial intelligence-like traffic or neural network-like transmission requirements.

In some embodiments, the present application further includes: the terminal equipment receives the second model.

Optionally, the terminal device may also receive the first model.

In some embodiments, the present application further includes: the terminal equipment receives the second model and the third model.

Optionally, the terminal device may further receive the first model and the fourth model.

The first model, the second model, the third model, the fourth model, the sub-model in the first model, the sub-model in the second model, the sub-model in the third model or the sub-model in the fourth model are carried by one of the following: downlink control signaling, MAC CE messages, RRC messages, broadcast messages, downlink data transmissions for artificial intelligence-like traffic transmission requirements.

Optionally, the method may further include: the terminal equipment receives a first coding model, wherein the first coding model comprises a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms the second model;

the generation sub-model and the compression sub-model constitute the third model described above.

Correspondingly, the process of processing the first signal by the end device by adopting the second model to obtain the first information and processing the first information by adopting the third model to obtain the second information can be combined into one step, which comprises the following steps: and the terminal equipment processes the first signal by adopting the first coding model to obtain second information.

In some embodiments, the first coding model is carried by one of: downlink control signaling, MAC CE messages, RRC messages, broadcast messages, downlink data transmissions for artificial intelligence-like traffic transmission requirements.

Optionally, the method may further include: the terminal equipment receives a second coding model, wherein the second coding model comprises a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms the second model;

the compressed submodel constitutes the third model described above.

Correspondingly, the process of processing the first signal by the end device by adopting the second model to obtain the first information and processing the first information by adopting the third model to obtain the second information can be combined into one step, which comprises the following steps: and the terminal equipment processes the first signal by adopting a second coding model to obtain second information.

In some embodiments, the second coding model is carried by one of: downlink control signaling, MAC CE messages, RRC messages, broadcast messages, downlink data transmissions for artificial intelligence-like traffic transmission requirements.

Mode two:

the training process is completed by the terminal equipment (such as UE), and the terminal equipment may send all or part of the trained signal generation sub-model, the channel estimation sub-model, the channel information generation sub-model, the channel information compression sub-model, and the channel information recovery sub-model to the network equipment. In addition, the terminal device may send the first channel model module and/or the second channel simulation module to the network device, where the first channel model module and the second channel simulation module are respectively configured to simulate the influence of the downlink transmission channel and the uplink transmission channel on the signal.

Or the terminal device can use the trained channel estimation sub-model, the channel information generation sub-model and the channel information compression sub-model as an integral channel estimation and channel information feedback module, if the integral channel estimation and channel information feedback module is used as a first coding model to be sent to the network device, the channel information recovery sub-model can be used as a first decoding model; or the terminal device may use the trained channel estimation sub-model and the channel information compression sub-model as an integral channel estimation and channel information feedback module, for example, as a second coding model to send to the network device, and the channel information recovery sub-model may be used as a second decoding model. The model transmission mode is shown in fig. 8.

The above modes of each model, sub-model and module of the terminal device transmission can be completed in independent transmission or in non-independent transmission (for example, all the information is transmitted through a signaling and message).

The terminal device may send a signal generation sub-model for use by the network device in generating the reference signal using the signal generation sub-model. Or the terminal equipment can send a signal generation sub-model and a channel information recovery sub-model, which are used for the network equipment to generate a reference signal by adopting the signal generation sub-model, and recover the compressed channel information sent by the terminal equipment by adopting the channel information recovery sub-model.

Or, the terminal device may send the signal generating sub-model and the first channel simulation module, and the channel estimation sub-model respectively, so that the network device may use these sub-models/modules to evaluate the performance of the signal generating sub-model and the channel estimation sub-model, where the evaluation manner is the same as the evaluation manner of the terminal device described above, and will not be described herein again. After the evaluation is qualified, the network device can use the signal generation sub-model to generate a reference signal, and send the reference signal to the terminal device for the terminal device to perform channel estimation based on the received reference signal. In addition, the first channel simulation module may be stored in the network device in advance, in which case the network device does not need to receive the first channel simulation module from the terminal device. In addition, the first channel simulation module can be simplified into an identity matrix, and in this case, after the first channel simulation module is adopted to process the reference signal, the obtained signal is the same as the reference signal; that is, it is simulated that the reference signal does not change in the downlink transmission process, and the reference signal received by the terminal device through the downlink channel is the same as the reference signal sent by the network device.

Alternatively, the terminal device may send the signal generating sub-model, the first channel simulation module, the channel estimation sub-model, the channel information generating sub-model, the channel information compressing sub-model, the second channel simulation module, and the channel information recovering sub-model, respectively, for the network device to evaluate the performance of the signal generating sub-model, the channel estimation sub-model, the channel information generating sub-model, the channel information compressing sub-model, and the channel information recovering sub-model by using these sub-models/modules. The evaluation method is the same as the evaluation method of the terminal device, and is not described in detail herein. After the evaluation is qualified, the network equipment can adopt the signal generation sub-model to generate a reference signal, and send the reference signal to the terminal equipment for the terminal equipment to perform channel estimation based on the received reference signal; and recovering the compressed channel information received from the terminal device by adopting the channel information recovery sub-model. In addition, the first channel simulation module and/or the second channel simulation module may be stored in the network device in advance, in which case the network device does not need to receive the first channel simulation module and/or the second channel simulation module from the terminal device. In addition, the first channel simulation module can be simplified into an identity matrix, and in this case, after the first channel simulation module is adopted to process the reference signal, the obtained signal is the same as the reference signal; that is, it is simulated that the reference signal does not change in the downlink transmission process, and the reference signal received by the terminal device through the downlink channel is the same as the reference signal sent by the network device. Similarly, the second channel simulation module may be simplified into an identity matrix, where the second channel simulation module is used to process the compressed channel information, and the obtained result is the same as the compressed channel information; that is, it is simulated that the compressed channel information does not change in the uplink transmission process, and the compressed channel information received by the network device through the uplink channel is the same as the compressed channel information sent by the terminal device.

The method for transmitting the submodel/module by the terminal device may be one or more of the following modes: downlink control signaling, MAC CE messages, RRC messages, broadcast, downlink data transmission for artificial intelligence class services or neural network class transmission requirements.

Since one network device may serve a plurality of terminal devices, the network device may receive the above-mentioned model, sub-model and/or module transmitted by one of the terminal devices served by the network device, and use the received model, sub-model and/or module for reference signal generation and channel feedback processes of all or part of the terminal devices served by the network device. Taking network equipment as a base station and terminal equipment as a mobile terminal for example, the base station a serves the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4, and the mobile terminal 1 sends the above models, submodels and/or modules to the base station a, and the base station a uses the above models, submodels and/or modules for reference signal generation and channel feedback procedures of all or part of the mobile terminals (such as one or more of the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4) served by the base station.

Alternatively, the network device may receive the above models, sub-models, and/or modules sent by each terminal device served by the network device, and use the received models, sub-models, and/or modules for reference signal generation and channel feedback processes of the corresponding terminal device. Taking network equipment as a base station and terminal equipment as a mobile terminal as an example, for example, the base station a serves the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4, the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4 respectively send the models, the submodels and/or the modules trained by the base station a to the base station a, and the base station respectively adopts the models, the submodels and/or the modules received from each mobile terminal to execute the reference signal generation and the channel feedback process of the corresponding terminal equipment.

Or, the network device may receive the above models, sub-models and/or modules sent by at least two terminal devices served by the network device, and combine or optimize the received models, sub-models and/or modules to obtain new models, sub-models and/or modules, respectively, and use the new models, sub-models and/or modules for the reference signal generation and channel feedback process of all or part of the terminal devices served by the network device. Taking network equipment as a base station, terminal equipment as a mobile terminal, for example, the base station a serves the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4, the mobile terminal 1 and the mobile terminal 2 respectively send the models, the submodels and/or the modules trained by the base station a to the base station a, the base station combines or optimizes the received models, submodels and/or modules to obtain new models, submodels and/or modules, and uses the new models, submodels and/or modules for generating reference signals and channel feedback processes of all or part of the mobile terminals (such as one or more of the mobile terminal 1, the mobile terminal 2, the mobile terminal 3 and the mobile terminal 4) served by the new models, submodels and/or modules.

In some embodiments, the present application further includes: and the terminal equipment sends the first model. Optionally, the terminal device may also send the second model.

In some embodiments, the present application further includes: the terminal device transmits the first model and the fourth model. Optionally, the terminal device may also send the second model and/or the third model.

Optionally, the first model, the second model, the third model, the fourth model, the sub-model in the first model, the sub-model in the second model, the sub-model in the third model, or the sub-model in the fourth model are carried by one of: uplink control signaling, MAC CE messages, RRC messages, broadcast messages, uplink data transmissions, and uplink data transmissions for artificial intelligence-like traffic transmission requirements.

Optionally, the method may further include: the terminal equipment sends a first coding model which comprises a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

the generation sub-model and the compression sub-model constitute a third model.

In some embodiments, the first coding model is carried by one of: uplink control signaling, MAC CE messages, RRC messages, broadcast messages, uplink data transmissions, and uplink data transmissions for artificial intelligence-like traffic transmission requirements.

Optionally, the method may further include: the terminal equipment sends a second coding model which comprises a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

the compressed submodel constitutes a third model.

In some embodiments, the second coding model is carried by one of: uplink control signaling, MAC CE messages, RRC messages, broadcast messages, uplink data transmissions, and uplink data transmissions for artificial intelligence-like traffic transmission requirements.

The above describes two training subjects of the model, and the transmission, evaluation and use of the model/sub-model in case of model training by different subjects. The embodiment of the application can also adopt other devices to train the model, and send the trained model to the terminal device and the network device respectively. The model can be transmitted by a wired connection mode or a wireless connection mode. For example, the second first model (or model parameters of the second model) is transmitted to the terminal device via a wired connection with the terminal device, or the second model (or model parameters of the second model) is transmitted to the terminal device via another wireless connection with the terminal device. The wireless connection mode may be bluetooth or wireless fidelity (Wi-Fi, wireless Fidelity) connection mode, etc.

The foregoing embodiments are described by taking the case of channel estimation of a downlink channel as an example, and the embodiments of the present application are also applicable to generating an uplink reference signal and evaluating the uplink channel, where a specific manner corresponds to the foregoing embodiments and is not described herein again.

The specific modes of the terminal equipment for model training in the embodiment of the application at least have the following two modes:

example one, AI-based wireless communication system multi-user reference signal, channel estimation integrated design scheme:

example two: AI-based wireless communication system multi-user reference signal, channel estimation and channel information feedback integrated design scheme.

The two examples described above are each described below.

Example one:

optionally, the terminal device may use input information and/or a first channel simulation module to perform joint training on the first initial model and the second initial model, so as to obtain the first model and the second model after training. The first model can be obtained after the first initial model is jointly trained, and the second model can be obtained after the second initial model is jointly trained.

Specifically, the terminal device may input the input information into the first initial model to obtain a first set output by the first initial model, where the first set includes a plurality of first reference signals;

Inputting any one of the first reference signals in the first set into the first channel simulation module to obtain a second reference signal;

inputting the second reference signal into the second initial model to obtain channel information;

determining a first loss function based on at least one of the first set, the second reference signal, the channel information, and parameters in the first channel simulation module;

updating the first initial model and the second initial model according to the first loss function.

In some implementations, determining the first loss function may include:

determining a degree of difference and/or a reference signal quality of the channel information and parameters of the first channel simulation module based on at least one of the first set, the second reference signal, the channel information and parameters of the first channel simulation module;

the first loss function is determined based on a degree of difference of the channel information from a parameter of the first channel simulation module and/or a reference signal quality.

The above-described manner of determining the quality of the reference signal will be described in detail in the following embodiments.

Fig. 9 is a schematic diagram of a design scheme integrating AI-based wireless communication system multi-user reference signals and channel estimation according to an embodiment of the present application. In fig. 9, the first model (the first model is a first initial model before training is completed) is specifically a signal generation sub-model, and the second model (the second model is a second initial model before training is completed) is specifically a channel estimation sub-model. The first channel simulation module may not participate in training, for example, parameters of the first channel simulation module are fixed, so as to simulate the reference signal received by the terminal device after the reference signal is transmitted through the channel. The first channel simulation module may be pre-stored in the terminal device and the network device, respectively, or may be sent to the terminal device by the network device or other devices prior to the joint training.

In fig. 9, the signal generation sub-model and the channel estimation sub-model are also distinguished, but the above models are not generated independently, but are implemented by a neural network of joint design and training, and are supervised in the above-described integrated solution by a specific loss function design. Specifically, firstly, the scheme gives input, output, loss function and model structure information of the multi-user reference signal design, and meanwhile, the scheme gives input, output, loss function and model structure information of a channel estimation model matched with the reference signal design.

The input information includes at least one of: no input, noise, random number, sequence in a preset sequence set, channel type indication information, channel data sample information, wireless channel or scene related information.

The method comprises the following specific steps:

(a) No input: there is no independent input information.

(b) Noise: the input may be noise, which may be from the real environment or may be artificially generated.

(c) Random number: the input may be a random number sequence, or a pseudo-random number sequence.

(d) Sequence: the input may be a sequence in a given set of sequences, where the set of sequences may be one or more of a set of sequences, a set of golden sequences, a set of zc sequences, and the like.

(e) Channel type indication information: the input of the above joint scheme may further include indication information of a channel type, for example, indication frequency information, environment information, and scene information corresponding to the channel, for example: high frequency, low frequency, indoor, outdoor, dense cell, open field, internet of things scene, industrial scene, etc.

(f) Channel data sample information: the input of the above joint scheme may also include channel data samples.

For (b) (c) and (d) using noise, pseudo-random numbers or predefined sequences as inputs to the multi-user signal joint construction scheme, the format of the noise, random numbers or predefined sequences may be one-dimensional vectors, or two-dimensional matrices, or high-dimensional noise, random numbers or predefined sequence sets. The format of the noise, pseudo-random number, predefined sequence may be agreed upon in advance by a protocol or signaling. The format of the noise, pseudo-random number, predefined sequence may be consistent with the reference signal sequence format desired to be generated.

When the noise, pseudo-random number or predefined sequence is used as the input of the multi-user signal joint construction scheme in (b) (c) and (d), different noise, pseudo-random number or predefined sequence inputs form basic variables for constructing different reference signals, the diversity of the different input sequences can form the diversity of the signal generation submodel output, and the diversity and the reference signals meeting the constraint of the subsequent loss function form a multi-user reference signal set facing the scene and the task in the scheme.

The channel type indication information and the channel data sample information described in (e) and (f) can be directly used as input of a signal generation sub-model or can be used as input of one or more of a first channel simulation module and a channel estimation sub-model.

In addition, the input may also include other information related to the wireless channel or scene, such as signal-to-noise ratio, signal-to-interference-and-noise ratio, channel type, bandwidth information, latency information, etc., of the channel, as input to one or more of the various sub-models described above.

In the process of joint training, one or more of the above information may be input according to actual situations or actual scenes, which is not limited herein.

The output of the multi-user reference signal design proposed by this example includes the following:

first term, reference signal set:

for the signal generation submodel, the input is the input described in the previous section, the output is a reference signal set, and the reference signal set includes a plurality of reference signals. For example, the input of the signal generating sub-model may be a set of random numbers, or a set of given sequences, the output of the signal generating sub-model may be a set of corresponding sequences that are output after the set of random numbers or sequences pass through the signal generating sub-model, and the set of output sequences is an output reference signal set, where a plurality of reference signals may be included in the reference signal set. Each reference signal may be applied to a different UE.

Second item, channel information:

the output of the channel estimation sub-model may include channel information. The channel information may be complete channel information, such as time domain channel information, or frequency domain channel information.

Alternatively, the channel information may be distributed in the first dimension and/or the second dimension.

Alternatively, the channel information may be distributed in at least one of the first dimension, the second dimension, and the third dimension.

In particular, a single sample of channel information may be formed by a matrix of size M x N having M first granularity in a first dimension and N second granularity in a second dimension, M and N may or may not be equal, with a specific numerical indication within the matrix representing channel quality. Wherein the channel quality may be characterized by a signal strength value; the units of signal strength values may be dBm; alternatively, the signal strength value may be expressed in terms of a numerical value obtained after normalization without units. In addition, two-dimensional data of m×n may be synthesized into one-dimensional data of 1×1 (m×n) or (m×n) ×1, and the specific transformation may be a difference in terms of expression from the first dimension to the second dimension, or from the second dimension to the first dimension.

In some embodiments, the first dimension may be a frequency domain dimension, and the channel information includes data distributed over M1 (M1 is a value of M) frequency domain granularity of the frequency domain dimension; wherein M1 is a positive integer.

Optionally, the frequency domain granularity includes a RBs and/or b subcarriers, where a or b is a positive integer.

Specifically, a single sample of channel information may be distributed over a first dimension having M1 granularity (e.g., denoted as M), which may be a frequency domain dimension, and when the first dimension is a frequency domain dimension, granularity M may be a RBs (a is 1 or more, e.g., 2 RBs, 4 RBs, 8 RBs), or b subcarriers (b is greater than 1, e.g., 4 subcarriers, 6 subcarriers, 18 subcarriers). When the first dimension is a frequency domain dimension, the frequency domain range indicated by a single sample of the channel information is a frequency domain range of M1; for example, if granularity M is 4RB, the frequency domain range indicated by a single sample of channel information is M1 x 4RB.

In some embodiments, the first dimension may be a time domain dimension, and the channel information includes data distributed over M2 (M2 is a value of M, and M2 may be the same as or different from M1) delay granularity in the time domain dimension; wherein M2 is a positive integer.

Optionally, the latency granularity includes at least one of: p1 microseconds, p2 symbol lengths, and p3 symbol sampling points, wherein p1, p2 or p3 is a positive integer. The symbols may include orthogonal frequency division multiplexing (OFDM, orthogonal Frequency Division Multiplexing) symbols, among others.

In particular, a single sample of channel information may be distributed over a first dimension of M2 granularity (e.g., denoted as p), which may be a time domain dimension, and when the first dimension is a time domain dimension, granularity p may be a delay granularity, e.g., one delay granularity is p1 microseconds, or p2 symbols long, or the number of sampling points of p3 symbols. When the first dimension is a time domain dimension, the time domain range (or delay range) indicated by a single sample of the training set is a time domain range of M2 x p; for example, if granularity p is 8 symbols long, the time domain range indicated by a single sample of channel information is M2 x 8 symbols long.

In some embodiments, the second dimension may be a spatial domain dimension.

Optionally, the spatial domain dimension may be an antenna dimension, and the channel information includes data distributed over N1 first granularity of the antenna dimension, where N1 is a positive integer.

Wherein, the first granularity may include a pair of transceiving antennas.

Alternatively, the spatial domain dimension may be an angular domain dimension, and the channel information includes data distributed over N2 second granularities of the angular domain dimension, where N2 is a positive integer.

Wherein the second granularity comprises an angular interval. Such as the angular spacing of the transmit and receive antennas, and/or the angular spacing of the reception of the channel information.

In particular, a single sample of channel information may be distributed over a second dimension of N1 granularity (e.g., denoted as N), which may be a spatial domain dimension, and in particular may be an antenna dimension, e.g., over a second dimension consisting of N1 antenna pairs, the second granularity being a pair of transceiving antennas.

Furthermore, a single sample of channel information may also be distributed over a second dimension of N2 granularity (e.g. denoted q), which may be a spatial domain dimension, in particular an angular domain dimension, e.g. consisting of N2 angles over the second dimension, the second granularity being the size of the angular interval between the N angles.

The data on a particular combination of dimensions in a single sample of channel information indicates the channel quality indication for that particular combination of dimensions. For example, fig. 10 is a schematic diagram of a channel information structure according to the present application, where fig. 10 shows a matrix structure with a first dimension being a frequency domain dimension and a second dimension being m×n of a spatial dimension, and an indication value X on a 3 rd row and a 6 th column may be used to represent a channel quality case on a 3 rd frequency domain granularity (e.g., a frequency domain granularity of 2RB as shown in fig. 10) on a 6 th spatial granularity (e.g., a spatial granularity of 1 antenna pair as shown in fig. 10). As another example, fig. 11 is a schematic diagram of another channel information structure according to the present application, where fig. 11 shows a matrix structure with a first dimension being a time domain dimension and a second dimension being m×n of a space dimension, and an indicated value Y on a 5 th column of a 4 th row may be used to represent a channel quality condition at a 4 th delay granularity at a 5 th space granularity (the space granularity is 1 arrival angle as shown in fig. 11). In fig. 10 and 11, K represents the number of channel information, and K is a positive integer.

In some embodiments, the channel information includes a signature sequence of S groups of length U, where S or U is a positive integer.

Alternatively, S may be 2, 4 or 8.

Alternatively, the U may be 16, 32, 48, 64, 128 or 256.

Specifically, the output information of the channel estimation sub-model may be channel feature information obtained by mathematical transformation of the original channel information, for example, channel feature vector information obtained by singular value decomposition (SVD, singular Value Decomposition), which may be channel feature vector information decomposed into a single stream, or channel feature vector information decomposed into multiple streams. For example, the output information of the channel estimation sub-model is an S-stream eigenvector, and each stream is composed of an eigenvalue with a length of U. For example, it may be 2-stream, 4-stream, or 8-stream channel feature vector information, each stream being composed of a feature sequence of 16, 32, 48, 64, 128, or 256 lengths. Fig. 12 is a schematic diagram of a structure of channel eigenvector information according to the present application, and in the example of fig. 12, the output of the channel estimation sub-model is 4-stream eigenvectors, each eigenvector being composed of an eigenvalue sequence of length 32.

It should be noted that, because the channel information output by the signal generating sub-model and the channel information output after being processed by the first channel simulation module may be presented in complex form, the channel information may be further provided with a dimension based on the above description, where the dimension is caused by independently presenting the imaginary part and the real part of the channel information output by the signal generating sub-model (or the channel information output after being processed by the first channel simulation module). For example, there may be a third dimension, in addition to the first and second dimensions, as described above, the third dimension being derived from the real and imaginary parts of the channel information.

That is, the third dimension includes a complex dimension including 2 elements for carrying real and imaginary parts, respectively, in the data included in the channel information.

In addition, it should be noted that the output of the channel information may be split and combined based on the first dimension, the second dimension, and the third dimension. For example, the channel information is distributed in a T-dimensional matrix, where the T-dimensional matrix is a matrix formed by splitting and/or combining at least one of the first dimension, the second dimension, and the third dimension, and T is a positive integer.

For example, when the second dimension is an antenna pair dimension, the second dimension may be further split into a transmit antenna sub-dimension and a receive antenna sub-dimension, thereby expanding the dimensions of the virtual channel output form described above. The embodiment is not exhaustive of various possible dimensions after splitting

In the following description, two-dimensional channel information composed of a first dimension and a second dimension is taken as an example for simplicity of description, but it should be clear that the dimensions of the channel information are not limited to two dimensions.

The second output of the multi-user reference signal design, channel information, is introduced above. The multi-user reference signal design may also have an output of the overall joint scheme, as follows:

third, output of the entire joint scheme:

the output of the overall joint training scheme includes a trained signal generation sub-model, and/or a channel estimation sub-model. The system also comprises a first channel simulation module, wherein the first channel simulation module can be preset and used for simulating the change condition of the reference signal passing through the wireless channel, and the first channel simulation module can not participate in model training; alternatively, the first channel simulation module may be obtained through joint training.

The input of the channel estimation sub-model is the output of the reference signal sequence output by the signal generation sub-model after passing through the first channel simulation module. The first channel simulation module is configured to simulate a change condition of the reference signal passing through the wireless channel, for example, the first channel simulation module may be configured of a fully-connected network, and the weight of the fully-connected network is a channel matrix H of the wireless channel, that is, after the reference signal S output by the signal generating module passes through the first channel simulation module, a reference signal result of simulating that S passes through the channel H is obtained, for example, S' =h×s.

The various models, sub-models or modules described above may employ neural network structures, such as one or more of a fully connected network, a convolutional neural network, a residual network, a self-attention mechanism network. Fig. 13 is a schematic structural diagram of a neural network according to the present application, and as shown in fig. 13, a multi-user reference signal design scheme according to an embodiment of the present application includes a signal generation sub-model, a first channel simulation module, and a channel estimation sub-model, each sub-model/model including one or more fully connected layers. The input information can be a random number or an original sequence with the length of 64, and the input information input signal generates a submodel; the signal generation sub-model generates a reference signal set comprising a plurality of reference signals, and the reference signals output by the signal generation sub-model are used as the input of the first channel simulation module; the first channel simulation module outputs a result obtained by processing the received reference signal and inputs the result into the channel estimation sub-model; the channel estimation sub-model performs channel estimation based on the received data to obtain a final channel estimation result, for example, the size of the channel estimation result may be 8192, and may be converted into a three-dimensional matrix form of [128,32,2 ].

The above describes several output contents of the multi-user reference signal design scheme proposed in this example, and in the subsequent model use process, the terminal device or the network device may use the third output content, that is, the trained signal generation sub-model and the channel estimation sub-model, to generate the reference signal and perform the channel estimation.

The following describes the training process of the terminal device on the model and the design scheme of the loss function in this example.

In some embodiments, the terminal device adopts input information and/or a first channel simulation module to perform joint training on a first initial model and a second initial model, so as to obtain the first model and the second model after training.

Specifically, the above-mentioned joint training method may include:

the terminal equipment inputs the input information into the first initial model to obtain a first set output by the first initial model, wherein the first set comprises a plurality of first reference signals;

Specifically, the determining the first loss function may include:

Referring to the drawings in detail, fig. 14 is a schematic diagram of model structure and information transmission in an integrated design scheme of multi-user reference signals and channel estimation of an AI-based wireless communication system according to an embodiment of the present application. In fig. 14, the first model (the first model is the first initial model before training is completed) is specifically a signal generation sub-model, and the second model (the second model is the second initial model before training is completed) is specifically a channel estimation sub-model. The first channel simulation module may not participate in the training, such as parameter fixing of the first channel simulation module.

As shown in fig. 14, the parameters of the first channel simulation module are represented by matrix H. In the model training process, the signal generation sub-model outputs a first set, and the first set comprises a plurality of first reference signals S. After the first reference signal S passes through the first channel simulation module, a second reference signal S ', S' =h×s is output, where the symbol "×" indicates multiplication of two matrices; s' represents the reference signal received by the receiving end after the original reference signal passes through the wireless channel. S ' is input into a channel estimation sub-model, and channel estimation is carried out on the basis of the S ' by the channel estimation sub-model, so that channel information H ' is obtained. The third reference signal may be obtained by processing S with H ', and S "=h ' ×s, where S" is a reference signal obtained by processing the original reference signal (S) with the result (H ') of channel estimation, and may also be a scene-based reference signal.

The loss function in the embodiment of the application can be designed based on the difference degree between H 'and H and/or the quality of the reference signal, and the smaller the difference between H' and H is, the higher the quality of the reference signal is, the better the model effect is shown. The reference signal herein may refer to the original reference signal S (i.e., the first reference signal) generated by the signal generating sub-model, and/or the reference signal S' (i.e., the second reference signal) processed by the first channel simulation module, and/or the reference signal S "(i.e., the third reference signal) obtained by processing the original reference signal using the result of the channel estimation. Quality may be manifested as cross-correlation between different reference signals, and/or peak-to-average power ratio of reference signals; the lower the cross-correlation between different reference signals, the lower the peak-to-average power ratio of the reference signals, the higher the quality of the reference signals.

For example, the reference signal quality may be represented by at least one of:

cross-correlation between different first reference signals in the first set;

cross-correlation between a first reference signal in a first set and other reference signals;

peak-to-average power ratio of the first reference signals in the first set.

The other reference signals of the first reference signal may be pre-stored reference signals, such as reference signals in another reference signal set generated in the current training, or reference signals in another reference signal set generated in the previous N (N is a positive integer) times of training.

As another example, the quality of the reference signal is represented by at least one of:

cross-correlation between different second reference signals;

cross-correlation between the second reference signal and other reference signals;

peak-to-average power ratio of the second reference signal.

The other reference signals of the second reference signal may be pre-stored reference signals, for example, reference signals obtained by processing a reference signal in another reference signal set generated in the current training by the first channel simulation module, or reference signals obtained by processing a reference signal in another reference signal set generated in the previous N times of training by the first channel simulation module.

As another example, the above reference signal quality is represented by at least one of:

cross-correlation between different third reference signals;

cross-correlation between the third reference signal and other reference signals;

peak-to-average power ratio of the third reference signal;

and the third reference signal is obtained by processing the first reference signal based on the channel information.

The other reference signals of the third reference signal may be pre-stored reference signals, for example, reference signals obtained by processing the channel information by using reference signals in another reference signal set generated in the current training, or reference signals obtained by processing the channel information by using reference signals in another reference signal set generated in the previous N times of training.

In some embodiments, the degree of difference between the channel information (representing the estimated channel) and the parameters of the first channel simulation module (representing the actual channel) may be measured by a specific distance, such as mean square error (MSE, mean Squared Error) or Normalized Mean Square Error (NMSE); the degree of similarity may also be measured as a measure, such as cosine similarity, cosine similarity squared, and the like.

The above several metrics in the first loss function may be joint metrics, such as joint metrics with equal weights added, or joint metrics with unequal weights added (for example, giving a greater weight to the proportion of the cross correlation of the reference signals in the joint metrics, or giving a greater weight to the accuracy of the channel estimation result, or giving equal weights each accounting for 50%); or by way of multiplication or cross entropy calculation.

For example, the first loss function=x×log (reference signal cross correlation) +y×log (degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel). Referring to fig. 14, wherein reference signal cross-correlation may refer to cross-correlation between different S, between different S', or between different s″ in fig. 14; the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is represented as the degree of difference of H' and H in fig. 14. Wherein x and y may be positive numbers, and if a greater weight is given to the cross correlation of the reference signals, x > y may be taken; x < y can be taken if the channel estimate is given more weight.

As another example, the first loss function=x×log (reference signal cross correlation) +y×log (degree of difference between the channel estimation result of the output of the channel estimation submodel and the actual channel) +z×log (peak-to-average power ratio of the reference signal). Referring to fig. 14, in which the reference signal may refer to S, S 'or S "in fig. 14, the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is expressed as the degree of difference of H' and H in fig. 14. Where x, y and z may be positive numbers.

As another example, the first loss function = x log (reference signal cross correlation) +z log (peak-to-average power ratio of the reference signal);

alternatively, the first loss function = log (reference signal cross correlation);

alternatively, the first loss function=log (the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel).

Referring to fig. 14, in which the reference signal may refer to S, S 'or S "in fig. 14, the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is expressed as the degree of difference of H' and H in fig. 14. Where x, y and z may be positive numbers.

The above describes an example one of model training of the terminal device in the embodiment of the present application, that is, an integrated design scheme of multi-user reference signals and channel estimation of the AI-based wireless communication system; another example of an AI-based wireless communication system multi-user reference signal, channel estimation, channel information feedback integrated design is presented below.

Example two:

optionally, the terminal device may perform joint training on the first initial model, the second initial model, the third initial model, and the fourth initial model by using at least one of the input information, the first channel simulation module, and the second channel simulation module, to obtain the trained first model, second model, third model, and fourth model. The first model can be obtained after the first initial model is jointly trained, the second model can be obtained after the second initial model is jointly trained, the third model can be obtained after the third initial model is jointly trained, and the fourth model can be obtained after the fourth initial model is jointly trained.

inputting the channel information into the third initial model to obtain compressed information of the channel information; wherein the third initial model comprises a generated initial sub-model and a compressed initial sub-model, and the input of the generated initial sub-model is used as the input of the third initial model, the output of the generated initial sub-model is used as the output of the compressed initial sub-model, and the output of the compressed initial sub-model is used as the output of the third initial model; alternatively, the third initial model comprises a compressed initial sub-model;

inputting the compressed information of the channel information into the second channel simulation module to obtain equivalent receiving information of the compressed information of the channel information;

Inputting equivalent receiving information of the compressed information of the channel information into the fourth initial model to obtain output information of the fourth initial model;

determining a second loss function based on at least one of the first set, the second reference signal, the channel information, parameters in the first channel simulation module, input information of the compressed initial sub-model, and output information of the fourth initial model;

updating the first initial model, the second initial model, the third initial model and the fourth initial model according to the second loss function.

In some implementations, determining the second loss function may include:

determining at least one of a reference signal quality, a degree of difference of the channel information from the parameters of the first channel simulation module, and a degree of difference of the output information of the fourth initial model from the input information of the compressed initial sub-model based on at least one of the first set, the second reference signal, the channel information, the parameters in the first channel simulation module, the input information of the compressed initial sub-model, and the output information of the fourth initial model;

The second loss function is determined based on at least one of the reference signal quality, the degree of difference of the channel information and the parameters of the first channel simulation module, and the degree of difference of the output information of the fourth initial model and the input information of the compressed initial sub-model.

Fig. 15 is a schematic diagram of an integrated design of multi-user reference signal, channel estimation, and channel information feedback for an AI-based wireless communication system according to an embodiment of the present application. In fig. 15, the first model (the first model is a first initial model before training is completed) is specifically a signal generation sub-model, and the second model (the second model is a second initial model before training is completed) is specifically a channel estimation sub-model. The first channel simulation module may not participate in training, for example, parameters of the first channel simulation module are fixed, so as to simulate the reference signal received by the terminal device after the reference signal is transmitted through the channel. The first channel simulation module may be pre-stored in the terminal device and the network device, respectively, or may be sent to the terminal device by the network device or other devices prior to the joint training. The third model (before training is completed, the third model is a third initial model) is specifically a channel information compression sub-model, or specifically includes a channel information generation sub-model and a channel information compression sub-model; the block diagram of the channel information generation sub-model in fig. 15 is a broken line, indicating that the channel information generation sub-model is an option. When the third model includes a channel information generation sub-model, the channel information output from the channel estimation sub-model is input to the channel information generation sub-model, and the channel information generation sub-model converts the channel information into a channel information feature vector and inputs the channel information feature vector to the channel information compression sub-model. In the case where the channel information generation sub-model is not included in the third model, the channel information output from the channel estimation sub-model is input to the channel information compression sub-model. The information output by the channel information compression sub-model is input to the second channel simulation module, and the second channel simulation module may not participate in training, for example, parameters of the second channel simulation module are fixed, and the compressed information output by the channel information compression sub-model is used for compressing information received by the network device after being transmitted through a channel. The second channel simulation module may be pre-stored in the terminal device and the network device, respectively, or may be sent to the terminal device by the network device or other devices prior to the joint training. The fourth model (before training is completed, the fourth model is a fourth initial model) is specifically a channel information recovering sub-model, and is used for recovering the received compressed information. The function of the channel information compression sub-model can be to compress the input information, and the function of the channel information recovery sub-model can be to decompress the input information; in an ideal situation, the channel information recovery sub-model should be able to recover the data before compression by the channel information compression sub-model.

In fig. 15, although the signal generation sub-model, the channel estimation sub-model, the channel information generation sub-model, the channel information compression sub-model, and the channel information restoration sub-model are also distinguished, the respective models are not independently generated but are realized by a neural network which is jointly designed and trained, and the above-described integrated solution is supervised by a specific loss function design.

The input information in this embodiment may be the same as the input information in the above example one, and will not be described here again.

first term, reference signal set:

the reference signal set may be the same as the reference signal set in the above example one, and will not be described here again.

Second item, channel information:

the channel information may be the same as that in the above example one, and will not be described here again.

Third, output of the entire joint scheme:

the output of the overall joint training scheme includes a trained signal generation sub-model, a channel estimation sub-model, a channel information generation sub-model, a channel information compression sub-model, and/or a channel information recovery sub-model. The system further comprises a first channel simulation module, wherein the first channel simulation module can be preset and used for simulating the change condition of the reference signal passing through the wireless channel, and the first channel simulation module can not participate in model training. The system further comprises a second channel simulation module, wherein the second channel simulation module can be preset and used for simulating the change condition of the compressed channel information (or the compressed channel information feature vector) passing through the wireless channel, and the second channel simulation module can not participate in model training.

The output of the above-described channel estimation sub-model may be the result of channel estimation by the reference signal, e.g. complete channel information. The result of channel estimation (e.g., complete channel information) may be directly input to the channel information compression sub-model, or the result of channel estimation may be input to the channel information compression sub-model after being converted by the channel information generation sub-model (e.g., channel feature vector information obtained by SVD decomposition). The output of the channel information compression sub-model may be directly input to the channel information recovery sub-model, or may be input to the channel information recovery sub-model after being processed by the second channel simulation module. The second channel simulation module may simulate a wireless channel environment, for example, a real channel, a channel generated by a channel scene predetermined in a protocol, or a channel obtained by modeling and fitting a channel may be adopted, and then the output of the channel information compression submodel may be passed through the channel, or the output of the channel information compression submodel may be convolved with the channel or a data process equivalent to the convolution may be performed (for example, the output may be multiplied after being transformed into a frequency domain by fourier transform, and then transformed into a time domain by inverse fourier transform, thereby equivalently obtaining a result of the time domain convolution). The channel information restoration sub-model outputs restored channel information, which may be complete channel information or channel feature vector information obtained after conversion (e.g., by SVD decomposition) of the channel information generation sub-model.

The various models, sub-models or modules described above may employ neural network structures, such as one or more of a fully connected network, a convolutional neural network, a residual network, a self-attention mechanism network. Fig. 16 is a schematic diagram of another neural network structure according to the present application, and as shown in fig. 16, a multi-user reference signal design scheme according to an embodiment of the present application includes a signal generation sub-model, a first channel simulation module, a channel estimation sub-model, a channel information compression sub-model, a second channel simulation module, and a channel information recovery sub-model, each sub-model/model including one or more fully connected layers. The input information may be a random number of length 64 or an original sequence, and the input information input signal generates a submodel. The signal generation sub-model generates a reference signal set comprising a plurality of reference signals, the reference signals output by the signal generation sub-model being input to the first channel simulation module. The first channel simulation module outputs a result of processing the received reference signal and inputs the result to the channel estimation sub-model. And the channel estimation sub-model carries out channel estimation based on the received data to obtain channel information. And inputting the channel information into a channel information compression sub-model to obtain compressed channel information. The compressed channel information is input to a second channel simulation module, the second channel simulation module outputs a result obtained by processing the received compressed channel information, and the processing result is input to a channel information recovery sub-model to obtain the finally recovered channel information. For example, the recovered channel information may be 8192 in size and may be converted into a three-dimensional matrix form of [128,32,2 ].

In the following use process, the terminal device or the network device may use the third output content, that is, the trained signal generating sub-model, the channel estimation sub-model, the channel information generating sub-model (optional), the channel information compressing sub-model and the channel information recovering sub-model, to generate the reference signal, perform channel estimation and perform channel information feedback.

In some embodiments, the terminal device performs joint training on the first initial model, the second initial model, the third initial model and the fourth initial model by adopting at least one of input information, a first channel simulation module and a second channel simulation module, so as to obtain the first model, the second model, the third model and the fourth model after training.

Specifically, the above-mentioned joint training method may include:

Specifically, the determining the second loss function may include:

Referring to the drawings in detail, fig. 17 is a schematic diagram of model structure and information transmission in an integrated design scheme of multi-user reference signal, channel estimation and channel information feedback of an AI-based wireless communication system according to an embodiment of the present application. In fig. 17, the first model (the first model is the first initial model before training is completed) is specifically a signal generation sub-model, and the second model (the second model is the second initial model before training is completed) is specifically a channel estimation sub-model. The first channel simulation module may not participate in the training, such as parameter fixing of the first channel simulation module.

As shown in fig. 17, the parameters of the first channel simulation module are represented by matrix H. In the model training process, the signal generation sub-model outputs a first set, and the first set comprises a plurality of first reference signals S. After the first reference signal S passes through the first channel simulation module, a second reference signal S ', S' =h×s is output, where the symbol "×" indicates multiplication of two matrices; s' represents the reference signal received by the receiving end after the original reference signal passes through the wireless channel. S ' is input into a channel estimation sub-model, and channel estimation is carried out on the basis of the S ' by the channel estimation sub-model, so that channel information H ' is obtained. The third reference signal may be obtained by processing S with H ', and S "=h ' ×s, where S" is a reference signal obtained by processing the original reference signal (S) with the result (H ') of channel estimation, and may also be referred to as a scene-based reference signal. And after channel information H' is subjected to channel information generation, channel information compression, simulation transmission in a channel and channel information recovery, recovery information output by a channel information recovery sub-model is obtained.

The loss function in the embodiment of the application can be designed based on the difference degree between H' and H, and/or the quality of the reference signal, and/or the difference degree between the channel information output by the channel information recovery sub-model and the input information of the channel information compression sub-model. Specifically, the difference degree between H' and H represents the quality of channel estimation, and the difference degree between the channel information output by the channel information recovery sub-model and the input information of the channel information compression sub-model represents the quality of channel information feedback. The smaller the difference between H' and H, the higher the quality of the reference signal, and the smaller the difference between the channel information output by the channel information recovery sub-model and the input information of the channel information compression sub-model, the better the model effect is indicated. The quality of the reference signal is the same as that of the first example, and will not be described here again.

In some embodiments, the degree of difference between the channel information (representing the estimated channel) and the parameters of the first channel simulation module (representing the actual channel) and the degree of difference between the channel information output by the channel information recovery sub-model and the input information of the channel information compression sub-model may be measured by a specific distance, for example, MSE or NMSE; the degree of similarity may also be measured as a measure, such as cosine similarity, cosine similarity squared, and the like.

The above several metrics in the second loss function may be joint metrics, such as joint metrics with equal weights added, or joint metrics with unequal weights added (for example, giving a greater weight to the proportion of the cross-correlation of the reference signals in the joint metrics, or giving a greater weight to the accuracy of the channel estimation result, or giving a greater weight to the accuracy of the channel information feedback); or by way of multiplication or cross entropy calculation.

For example, the second loss function=x×log (reference signal cross correlation) +y×log (degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel) +z×log (degree of difference between the channel information of the output of the channel information recovery sub-model and the input information of the channel information compression sub-model). Referring to fig. 17, wherein reference signal cross-correlation may refer to cross-correlation between different S, between different S', or between different s″ in fig. 17; the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is represented as the degree of difference of H' and H in fig. 17. Wherein x, y, z may be positive numbers.

As another example, the second loss function=x×log (reference signal cross correlation) +y×log (degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel) +z×log (degree of difference between the channel information output of the channel information recovery sub-model and the input information of the channel information compression sub-model) +h×log (peak-to-average power ratio of the reference signal). Referring to fig. 17, wherein reference signal cross-correlation may refer to cross-correlation between different S, between different S', or between different s″ in fig. 17; the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is represented as the degree of difference of H' and H in fig. 17. Wherein x, y, z, h can be a positive number.

As another example, the second loss function=x×log (reference signal cross correlation) +z×log (degree of difference between the channel information output from the channel information recovery sub-model and the input information of the channel information compression sub-model). Referring to fig. 17, wherein reference signal cross-correlation may refer to cross-correlation between different S, between different S', or between different s″ in fig. 17; the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is represented as the degree of difference of H' and H in fig. 17. Wherein x and z may be positive numbers.

As another example, the second loss function=x×log (reference signal cross correlation) +z×log (degree of difference between the channel information output by the channel information recovering sub-model and the input information of the channel information compressing sub-model) +h×log (peak-to-average power ratio of the reference signal). Referring to fig. 17, wherein reference signal cross-correlation may refer to cross-correlation between different S, between different S', or between different s″ in fig. 17; the degree of difference between the channel estimation result of the output of the channel estimation sub-model and the actual channel is represented as the degree of difference of H' and H in fig. 17. Wherein x, z, h may be positive numbers.

The above describes an example two of model training of the terminal device in the embodiment of the present application, that is, an integrated design scheme of multi-user reference signal, channel estimation and channel information feedback of the AI-based wireless communication system.

The manner of convergence with respect to the above training may include at least one of: judging whether the number of iterative training reaches the preset number of iterative training, and judging whether the difference degree is smaller than a preset threshold value. The preset times and the preset threshold value can be set according to actual conditions. When the training is determined to be completed based on the above manner, the first initial model, the second initial model, the third initial model and the fourth initial model after the training is completed can be respectively used as the first model, the second model, the third model and the fourth model.

The determination manner of the model training convergence in this example is the same as that in the foregoing example, and will not be described here again.

Therefore, through the combined training mode, the scheme for making the multi-user reference signal of the wireless communication system based on the AI is provided, so that the overall advantages of better reference signal design, wireless communication solution design and scene adaptation are obtained. Specifically, an integrated design scheme for multi-user reference signals and channel estimation of the AI-based wireless communication system is provided, and a corresponding scheme input, output, model structure division and a loss function design for the integrated design in the corresponding scheme are provided. The designs can form training schemes of corresponding reference signal generation modules and reference signal generation schemes under different task targets, and the multi-user reference signal set facing the scene and the task in the scheme is formed by various reference signals meeting the constraint of the subsequent loss function.

The joint design scheme provided by the present disclosure has at least the following advantages: (1) Instead of applying existing reference signals to AI-based wireless communication solutions (including channel estimation, channel information feedback, etc.), AI-based wireless communication solutions and optimally adapted reference signal construction are implemented as a unitary solution, thereby achieving optimal matching of reference signal designs with wireless communication solutions; (2) The AI-based solution is beneficial to achieving scene adaptation and obtaining corresponding adaptation gain, while the solution is beneficial to considering scene factors when constructing the reference signal so as to obtain the overall advantages of better reference signal design, wireless communication solution design and scene adaptation.

The present application also proposes another communication method, fig. 18 being a schematic flow chart of another communication method 1800 according to an embodiment of the present application, which may alternatively be applied to the system shown in fig. 1, but is not limited thereto. The method includes at least some of the following.

S1810: the network device transmits a first signal, the first signal being generated by a first model; the first signal is used for processing by the second model to obtain first information, which includes channel information.

The present application also proposes another communication method, fig. 19 is a schematic flow chart of a communication method 1900 according to an embodiment of the present application, which may alternatively be applied to the system shown in fig. 1, but is not limited thereto. The method includes at least some of the following.

As shown in fig. 19, the communication method further includes, after S1810:

s1920: the network equipment receives second information, and the second information is obtained by processing the first information through a third model;

s1930: the network equipment processes the second information by adopting a fourth model to obtain third information;

The first model, the second model, the third model and the fourth model are obtained through joint training.

Optionally, the first signal includes a reference signal.

Optionally, the second model includes a channel estimation sub-model, and the first information includes channel information;

the channel estimation sub-model is used for carrying out channel estimation based on the first signal to obtain channel information.

Optionally, the third model includes a compression sub-model;

the compression sub-model is used for compressing the first information to obtain compressed information of the first information; the second information includes compressed information of the first information.

Correspondingly, the fourth model comprises a recovery sub-model;

the recovery sub-model is used for carrying out recovery processing on the compressed information of the first information to obtain recovery information of the first information; the third information includes recovery information of the first information.

In another embodiment, the third model may include a generation sub-model and a compression sub-model; wherein,

the generation sub-model is used for carrying out feature transformation on the first information to obtain a first feature vector corresponding to the first information;

The compression sub-model is used for compressing the first feature vector to obtain compression information of the first feature vector; the second information includes compressed information of the first feature vector.

Correspondingly, the fourth model comprises a recovery sub-model;

the recovery sub-model is used for recovering the compressed information of the first feature vector to obtain recovery information of the first feature vector; the third information includes restoration information of the first feature vector.

In some embodiments, the above method further comprises: the network device receives the first model.

Optionally, the method may further include: the network device receives the second model.

In some embodiments, the above method further comprises: the network device receives the first model and the fourth model.

Optionally, the method may further include: the network device receives the second model and/or the third model.

Optionally, the method may further include: the network device receives a first coding model, wherein the first coding model comprises a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

That is, the terminal device packages the channel estimation sub-model, the generation sub-model, and the compression sub-model into one model, i.e., a first coding model, which is transmitted and used as a whole.

Optionally, the first coding model is carried by one of: uplink control signaling, MAC CE messages, RRC messages, broadcast messages, uplink data transmissions, and uplink data transmissions for artificial intelligence-like traffic transmission requirements.

Optionally, the method may further include: the network device receives a second coding model, the second coding model comprising a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

the compression sub-model constitutes a third model.

That is, the terminal device packages the channel estimation sub-model and the compression sub-model into one model, i.e., a second coding model, which is transmitted and used as a whole.

Optionally, the second coding model is carried by one of: uplink control signaling, MAC CE messages, RRC messages, broadcast messages, uplink data transmissions, and uplink data transmissions for artificial intelligence-like traffic transmission requirements.

Further, the method may further include training the model/sub-model by the network device.

In some embodiments, the above method further comprises: and the network equipment adopts input information and/or a first channel simulation module to perform joint training on the first initial model and the second initial model to obtain the first model and the second model after training.

The network device adopts input information and/or a first channel simulation module to perform joint training on the first initial model and the second initial model, which may specifically include:

the network equipment inputs the input information into the first initial model to obtain a first set output by the first initial model, wherein the first set comprises a plurality of first reference signals;

Optionally, the determining the first loss function includes:

In other embodiments, the above method further comprises: and the network equipment adopts at least one of input information, a first channel simulation module and a second channel simulation module to perform joint training on a first initial model, a second initial model, a third initial model and a fourth initial model to obtain the trained first model, second model, third model and fourth model.

The specific training mode can be as follows: comprising the following steps:

Optionally, determining the second loss function may include:

In some embodiments, the above reference signal quality is represented by at least one of:

cross-correlation between different first reference signals in the first set;

cross-correlation between a first reference signal in the first set and other reference signals;

a peak-to-average power ratio of a first reference signal in the first set.

Alternatively, in some embodiments, the quality of the reference signal is represented by at least one of:

cross-correlation between different ones of the second reference signals;

the peak-to-average power ratio of the second reference signal.

Alternatively, in some embodiments, the reference signal quality is represented by at least one of:

cross-correlation between different ones of said third reference signals;

A peak-to-average power ratio of the third reference signal;

Optionally, the degree of difference is measured by distance and/or similarity.

In some embodiments, the input information includes at least one of: no input, noise, random number, sequence in a preset sequence set, channel type indication information, channel data sample information, wireless channel or scene related information.

Optionally, the preset sequence set includes at least one of the following: a set of m sequences, a set of golden sequences, a set of zc sequences.

Optionally, the channel type indication information indicates at least one of the following: frequency information corresponding to the channel, environment information corresponding to the channel and scene information corresponding to the channel.

Optionally, the wireless channel or scene related information includes at least one of: signal to noise ratio, signal to interference noise ratio, channel type, bandwidth information, delay information of the channel.

Optionally, the format of the sequence in the noise, the random number or the preset sequence set is the same as the format of the output data of the first initial model.

Optionally, the format of the sequences in the noise, random number or preset sequence set includes at least one of the following formats: one-dimensional vectors, two-dimensional matrices, high-dimensional matrices.

Optionally, the format of the sequences in the noise, the random number or the preset sequence set is agreed by a protocol or signaling.

Optionally, the input information is used for inputting at least one of the following: the system comprises a first initial model, a first channel simulation module and a second initial model.

In some embodiments, the channel information is distributed in a first dimension and/or a second dimension.

Optionally, the first dimension is a frequency domain dimension, and the channel information includes data distributed on M1 frequency domain granularities of the frequency domain dimension; and M1 is a positive integer.

The frequency domain granularity comprises a RB and/or b sub-carriers, wherein a or b is a positive integer.

Optionally, the first dimension is a time domain dimension, and the channel information includes data distributed on M2 delay granularities of the time domain dimension; and M2 is a positive integer.

Wherein the latency granularity includes at least one of: p1 microseconds, p2 symbol lengths, and p3 symbol sampling points, wherein p1, p2 or p3 is a positive integer.

The symbols may include OFDM symbols.

Optionally, the second dimension is a spatial domain dimension.

For example, the spatial domain dimension is an antenna dimension, the channel information includes data distributed over N1 first granularity of the antenna dimension, and N1 is a positive integer.

Specifically, the first granularity may include a pair of transceiver antennas.

For another example, the spatial domain dimension is an angular domain dimension, the channel information includes data distributed over N2 second granularities of the angular domain dimension, where N2 is a positive integer.

In particular, the second granularity may include an angular interval.

In some embodiments, the third dimension includes a complex dimension, where the complex dimension includes 2 elements for respectively carrying a real part and an imaginary part in data included in the channel information.

In some embodiments, the channel information is distributed in a T-dimensional matrix, where the T-dimensional matrix is a matrix formed after splitting and/or combining at least one of the first dimension, the second dimension, and the third dimension, and T is a positive integer.

In some embodiments, the channel information includes a characteristic sequence of S groups having a length U, where S or U is a positive integer.

Specifically, S may be 2, 4 or 8.

Specifically, the U may be 16, 32, 48, 64, 128 or 256.

In some embodiments, the above method may further comprise: the network device transmits the second model.

Further, the method may further include: the network device transmits the first model.

In other embodiments, the above method may further comprise: the network device transmits the second model and the third model.

Further, the method may further include: the network device transmits the first model and/or the fourth model.

In some embodiments, the first model, the second model, the third model, the fourth model, a sub-model in the first model, a sub-model in the second model, a sub-model in the third model, or a sub-model in the fourth model described above is carried by one of: downlink control signaling, MAC CE messages, RRC messages, broadcast messages, downlink data transmissions for artificial intelligence-like traffic transmission requirements.

In some embodiments, the above method may further comprise: the network equipment sends a first coding model which comprises a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model forms a second model;

That is, the network device may package the channel estimation sub-model, the generation sub-model, and the compression sub-model into one entity, i.e., the first coding model, which is transmitted and used as one entity.

The first coding model may be carried by one of: downlink control signaling, MAC CE messages, RRC messages, broadcast messages, downlink data transmissions for artificial intelligence-like traffic transmission requirements.

In other embodiments, the above method may further comprise: the network equipment sends a second coding model which comprises a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

the compression sub-model constitutes a third model.

That is, the network device may package the channel estimation sub-model and the compression sub-model into one entity, i.e., the second coding model, which is transmitted and used as one entity.

The second coding model may be carried by one of: downlink control signaling, MAC CE messages, RRC messages, broadcast messages, downlink data transmissions for artificial intelligence-like traffic transmission requirements.

The present embodiments also provide a model training method, and fig. 20 is a schematic flow chart of another communication method 2000 according to the present embodiments, which may alternatively be applied to the system shown in fig. 1, but is not limited thereto. The model training method may be performed by the terminal device, or by the network device, or by other electronic devices. The method includes at least some of the following.

S2010: and carrying out joint training on the first initial model and the second initial model by adopting input information and/or a first channel simulation module to obtain a trained first model and trained second model.

In some embodiments, the above-described joint training includes:

inputting the input information into the first initial model to obtain a first set of output of the first initial model, wherein the first set comprises a plurality of first reference signals;

Optionally, the determining the first loss function includes:

In some embodiments, the above-described joint training includes:

and performing joint training on the first initial model, the second initial model, the third initial model and the fourth initial model by adopting at least one of the input information, the first channel simulation module and the second channel simulation module to obtain a trained first model, a trained second model, a trained third model and a trained fourth model.

Specifically, the joint training may include:

Optionally, the determining the second loss function may include:

In one embodiment, the reference signal quality is represented by at least one of:

cross-correlation between different first reference signals in the first set;

a peak-to-average power ratio of a first reference signal in the first set.

In another embodiment, the quality of the reference signal is represented by at least one of:

cross-correlation between different ones of the second reference signals;

the peak-to-average power ratio of the second reference signal.

In another embodiment, the reference signal quality is represented by at least one of:

cross-correlation between different ones of said third reference signals;

a peak-to-average power ratio of the third reference signal;

Alternatively, the degree of difference may be measured using distance and/or similarity.

Wherein the set of preset sequences may include at least one of: a set of m sequences, a set of golden sequences, a set of zc sequences.

Wherein the channel type indication information may indicate at least one of: frequency information corresponding to the channel, environment information corresponding to the channel and scene information corresponding to the channel.

Wherein the wireless channel or scene related information may include at least one of: signal to noise ratio, signal to interference noise ratio, channel type, bandwidth information, delay information of the channel.

The format of the noise, the random number or the sequence in the preset sequence set may be the same as the format of the output data of the first initial model.

The format of the sequence in the noise, random number or preset sequence set may include at least one of the following formats: one-dimensional vectors, two-dimensional matrices, high-dimensional matrices.

The format of the sequences in the noise, random number or preset sequence set can be agreed by a protocol or signaling.

Wherein the input information may be used to input at least one of: the system comprises a first initial model, a first channel simulation module and a second initial model.

In some embodiments, the channel information is distributed in at least one of a first dimension, a second dimension, and a third dimension.

In one embodiment, the first dimension is a frequency domain dimension, and the channel information includes data distributed over M1 frequency domain granularities of the frequency domain dimension; and M1 is a positive integer.

For example, the frequency domain granularity includes a RBs and/or b subcarriers, where a or b is a positive integer.

In another embodiment, the first dimension is a time domain dimension, and the channel information includes data distributed over M2 delay granularities of the time domain dimension; and M2 is a positive integer.

For example, the latency granularity includes at least one of: p1 microseconds, p2 symbol lengths, and p3 symbol sampling points, wherein p1, p2 or p3 is a positive integer.

Optionally, the symbol comprises an OFDM symbol.

In some embodiments, the second dimension is a spatial domain dimension.

Optionally, the first granularity includes a pair of transceiver antennas.

Optionally, the second granularity includes an angular interval.

For example, S may be 2, 4 or 8.

For example, the U may be 16, 32, 48, 64, 128, or 256.

After the terminal device executes and the network device performs model training by using the training method, the trained model can be sent to the required device executing and/or network device, and the transmission method of the model is the same as that of the model in the communication method and is not described herein.

The embodiment of the present application further proposes a terminal device, and fig. 21 is a schematic structural diagram of a terminal device 2100 according to an embodiment of the present application, including:

A first receiving module 2110 for receiving a first signal, the first signal being generated by a first model;

a first processing module 2120, configured to process the first signal by using a second model to obtain first information;

Optionally, the terminal device further includes:

the second processing module is used for processing the first information by adopting a third model to obtain second information;

the first sending module is used for sending the second information, and the second information is used for being processed by the fourth model to obtain third information;

Optionally, the first signal includes a reference signal.

Optionally, the third model includes a compression sub-model;

Optionally, the fourth model includes a restorative sub-model;

Optionally, the third model includes a generation sub-model and a compression sub-model; wherein,

Optionally, the fourth model includes a restorative sub-model;

In some embodiments, the terminal device further includes: and the second receiving module is used for receiving the second model.

Optionally, the second receiving module is further configured to receive the first model.

In other embodiments, the terminal device further includes: and the third receiving module is used for receiving the second model and the third model.

Optionally, the third receiving module is further configured to receive the first model and/or the fourth model.

In some embodiments, the terminal device further includes:

a fourth receiving module, configured to receive a first coding model, where the first coding model includes a channel estimation sub-model, a generation sub-model, and a compression sub-model; wherein,

the channel estimation sub-model forms the second model;

the generation sub-model and compression sub-model constitute the third model.

In some embodiments, the terminal device further includes:

a fifth receiving module, configured to receive a second coding model, where the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms the second model;

the compression sub-model constitutes the third model.

In some embodiments, the terminal device further includes:

and the first training module is used for jointly training the first initial model and the second initial model by adopting the input information and/or the first channel simulation module to obtain the first model and the second model after training.

In some embodiments, the terminal device further includes:

the second training module is used for jointly training the first initial model, the second initial model, the third initial model and the fourth initial model by adopting at least one of the input information, the first channel simulation module and the second channel simulation module to obtain the trained first model, the trained second model, the trained third model and the trained fourth model.

The specific manner of the combined training performed by the first training module or the second training module is the same as that of the foregoing method embodiment, and will not be described herein.

In some embodiments, the terminal device further includes: and the second sending module is used for sending the first model.

Optionally, the second sending module is further configured to send the second model.

In some embodiments, the terminal device further includes: and the third sending module is used for sending the first model and the fourth model.

Optionally, the third sending module is further configured to send the second model and/or the third model.

In some embodiments, the terminal device further includes:

a fourth transmitting module, configured to transmit a first coding model, where the first coding model includes a channel estimation sub-model, a generation sub-model, and a compression sub-model; wherein,

The channel estimation sub-model forms a second model;

In some embodiments, the terminal device further includes:

a fifth transmitting module, configured to transmit a second coding model, where the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

the compression sub-model constitutes a third model.

It should be understood that the foregoing and other operations and/or functions of the modules in the terminal device according to the embodiments of the present application are respectively for implementing the corresponding flow of the terminal device in the method 400 of fig. 4, and are not described herein for brevity.

The embodiment of the present application further proposes a network device, and fig. 22 is a schematic structural diagram of a network device 2200 according to an embodiment of the present application, including:

a sixth transmitting module 2210 for transmitting a first signal, the first signal being generated by the first model; the first signal is used for being processed by the second model to obtain first information;

optionally, the first model and the second model are obtained through joint training.

In some embodiments, the network device further includes:

The sixth receiving module is used for receiving second information, and the second information is obtained by processing the first information through a third model;

the third processing module is used for processing the second information by adopting a fourth model to obtain third information;

Optionally, the first signal includes a reference signal.

In some embodiments, the second model includes a channel estimation sub-model, and the first information includes channel information;

In some embodiments, the third model comprises a compression sub-model;

In some embodiments, the fourth model includes a restorative sub-model;

In some embodiments, the third model includes a generative sub-model and a compressive sub-model; wherein,

Optionally, the fourth model includes a restorative sub-model;

In some embodiments, the network device further includes: and a seventh receiving module, configured to receive the first model.

Optionally, the seventh receiving module is further configured to receive the second model.

In some embodiments, the network device further includes: and an eighth receiving module, configured to receive the first model and the fourth model.

Optionally, the eighth receiving module is further configured to receive the second model and/or the third model.

In some embodiments, the network device further includes: a ninth receiving module, configured to receive a first coding model, where the first coding model includes a channel estimation sub-model, a generation sub-model, and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

In some embodiments, the network device further includes: a tenth receiving module, configured to receive a second coding model, where the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

the compression sub-model constitutes a third model.

In some embodiments, the network device further includes: and the third training module is used for jointly training the first initial model and the second initial model by adopting the input information and/or the first channel simulation module to obtain the first model and the second model after training.

In some embodiments, the network device further includes: and the fourth training module is used for jointly training the first initial model, the second initial model, the third initial model and the fourth initial model by adopting at least one of the input information, the first channel simulation module and the second channel simulation module to obtain the trained first model, the trained second model, the trained third model and the trained fourth model.

The specific manner of the joint training performed by the third training module or the fourth training module is the same as that of the foregoing method embodiment, and will not be described herein.

In some embodiments, the network device further includes: and a seventh sending module, configured to send the second model.

Optionally, the seventh sending module is further configured to send the first model.

In some embodiments, the network device further includes: and an eighth sending module, configured to send the second model and the third model.

Optionally, the eighth sending module is further configured to send the first model and/or the fourth model.

In some embodiments, the network device further includes:

a ninth transmitting module, configured to transmit a first coding model, where the first coding model includes a channel estimation sub-model, a generation sub-model, and a compression sub-model; wherein,

the channel estimation sub-model forms a second model;

In some embodiments, the network device further includes:

a tenth transmitting module, configured to transmit a second coding model, where the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model forms a second model;

the compression sub-model constitutes a third model.

It should be appreciated that the foregoing and other operations and/or functions of the modules in the network device according to the embodiments of the present application are respectively for implementing the corresponding flow of the network device in the method 1800 of fig. 18, and are not described herein for brevity.

The embodiment of the application further provides a model training device, and fig. 23 is a schematic structural diagram of a model training device 2300 according to the embodiment of the application, including:

the joint training module 2310 is configured to perform joint training on the first initial model and the second initial model by using the input information and/or the first channel simulation module, to obtain a trained first model and a trained second model.

In some embodiments, joint training module 2310 described above is used to:

Optionally, the determining the first loss function includes:

Optionally, the joint training module 2310 is configured to:

Optionally, the determining the second loss function includes:

Optionally, the reference signal quality is represented by at least one of:

cross-correlation between different first reference signals in the first set;

a peak-to-average power ratio of a first reference signal in the first set.

Optionally, the quality of the reference signal is represented by at least one of:

cross-correlation between different ones of the second reference signals;

the peak-to-average power ratio of the second reference signal.

Optionally, the reference signal quality is represented by at least one of:

cross-correlation between different ones of said third reference signals;

a peak-to-average power ratio of the third reference signal;

The specific mode of the combined training module is the same as the training mode in the embodiment of the method, and is not repeated here.

It should be appreciated that the foregoing and other operations and/or functions of the modules in the model training apparatus according to the embodiments of the present application are respectively for implementing the corresponding flow of the model training apparatus in the method 2000 of fig. 20, and are not described herein for brevity.

It should be noted that, regarding the functions described in the respective modules (sub-model, unit, or component, etc.) in the terminal device 2100, the network device 2200, and the model training device 2300 of the embodiments of the present application, the functions may be implemented by different modules (sub-model, unit, or component, etc.), or may be implemented by the same module (sub-model, unit, or component, etc.), for example, the first sending module and the second sending module may be different modules, or may be the same module, and all the functions thereof in the embodiments of the present application may be implemented. In addition, the transmitting module and the receiving module in the embodiments of the present application may be implemented by a transceiver of the device, and some or all of the remaining modules may be implemented by a processor of the device.

Fig. 24 is a schematic block diagram of a communication device or model training device 700 according to an embodiment of the present application. The communication device or model training device 700 shown in fig. 24 comprises a processor 710, from which the processor 710 may call and run a computer program to implement the methods in embodiments of the present application.

Optionally, as shown in fig. 24, the communication device or model training device 700 may also include a memory 720. Wherein the processor 710 may call and run a computer program from the memory 720 to implement the methods in embodiments of the present application.

Wherein the memory 720 may be a separate device from the processor 710 or may be integrated into the processor 710.

Optionally, as shown in fig. 24, the communication device or model training device 700 may further include a transceiver 730, and the processor 710 may control the transceiver 730 to communicate with other devices, and in particular, may send information or data to other devices, or receive information or data sent by other devices.

Among other things, transceiver 730 may include a transmitter and a receiver. Transceiver 730 may further include antennas, the number of which may be one or more.

Optionally, the communication device or the model training device 700 may be a terminal device in the embodiments of the present application, and the communication device or the model training device 700 may implement corresponding flows implemented by the terminal device in the methods in the embodiments of the present application, which are not described herein for brevity.

Optionally, the communication device or the model training device 700 may be a network device in the embodiments of the present application, and the communication device or the model training device 700 may implement corresponding flows implemented by the network device in the methods in the embodiments of the present application, which are not described herein for brevity.

Fig. 25 is a schematic structural diagram of a chip 800 according to an embodiment of the present application. The chip 800 shown in fig. 25 includes a processor 810, and the processor 810 may call and run a computer program from a memory to implement the method in the embodiments of the present application.

Optionally, as shown in fig. 25, chip 800 may also include a memory 820. Wherein the processor 810 may call and run a computer program from the memory 820 to implement the methods in embodiments of the present application.

Wherein the memory 820 may be a separate device from the processor 810 or may be integrated into the processor 810.

Optionally, the chip 800 may also include an input interface 830. The processor 810 may control the input interface 830 to communicate with other devices or chips, and in particular, may obtain information or data sent by other devices or chips.

Optionally, the chip 800 may further include an output interface 840. The processor 810 may control the output interface 840 to communicate with other devices or chips, and in particular, may output information or data to other devices or chips.

Optionally, the chip may be applied to a terminal device in the embodiment of the present application, and the chip may implement a corresponding flow implemented by the terminal device in each method in the embodiment of the present application, which is not described herein for brevity.

Optionally, the chip may be applied to a network device in the embodiment of the present application, and the chip may implement a corresponding flow implemented by the network device in each method in the embodiment of the present application, which is not described herein for brevity.

It should be understood that the chips referred to in the embodiments of the present application may also be referred to as system-on-chip chips, or the like.

The processors mentioned above may be general purpose processors, digital signal processors (digital signal processor, DSP), off-the-shelf programmable gate arrays (field programmable gate array, FPGA), application specific integrated circuits (application specific integrated circuit, ASIC) or other programmable logic devices, transistor logic devices, discrete hardware components, etc. The general-purpose processor mentioned above may be a microprocessor or any conventional processor.

The memory mentioned above may be volatile memory or nonvolatile memory, or may include both volatile and nonvolatile memory. The nonvolatile memory may be a read-only memory (ROM), a Programmable ROM (PROM), an Erasable PROM (EPROM), an electrically Erasable EPROM (EEPROM), or a flash memory. The volatile memory may be random access memory (random access memory, RAM).

It should be understood that the above memory is exemplary but not limiting, and for example, the memory in the embodiments of the present application may be Static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), synchronous Link DRAM (SLDRAM), direct RAM (DR RAM), and the like. That is, the memory in embodiments of the present application is intended to comprise, without being limited to, these and any other suitable types of memory.

In the above embodiments, it may be implemented in whole or in part by software, hardware, firmware, or any combination thereof. When implemented in software, may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When loaded and executed on a computer, produces a flow or function in accordance with embodiments of the present application, in whole or in part. The computer may be a general purpose computer, a special purpose computer, a computer network, or other programmable apparatus. The computer instructions may be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another computer-readable storage medium, for example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center by a wired (e.g., coaxial cable, fiber optic, digital subscriber line (Digital Subscriber Line, DSL)) or wireless (e.g., infrared, wireless, microwave, etc.). The computer readable storage medium may be any available medium that can be accessed by a computer or a data storage device such as a server, data center, etc. that contains an integration of one or more available media. The usable medium may be a magnetic medium (e.g., a floppy Disk, hard Disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium (e.g., solid State Disk (SSD)), or the like.

It should be understood that, in various embodiments of the present application, the sequence numbers of the foregoing processes do not mean the order of execution, and the order of execution of the processes should be determined by the functions and internal logic thereof, and should not constitute any limitation on the implementation process of the embodiments of the present application.

It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, and are not repeated herein.

The foregoing is merely a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any person skilled in the art can easily think about changes or substitutions within the technical scope of the present application, and the changes and substitutions should be covered in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.

Claims

A method of communication including:

The terminal device receives a first signal, the first signal is generated by the first model;

The terminal device uses a second model to process the first signal to obtain first information; the first information includes channel information;

The first model and the second model are jointly trained.

The method of claim 1, further comprising:

The terminal device processes the first information using a third model to obtain second information;

The terminal device sends the second information, and the second information is used for processing by the fourth model to obtain the third information;

Wherein, the first model, the second model, the third model and the fourth model are obtained by joint training.

The method of claim 1 or 2, wherein the first signal includes a reference signal.

The method according to any one of claims 1 to 3, wherein the second model includes a channel estimation sub-model;

The channel estimation sub-model is used to perform channel estimation based on the first signal to obtain channel information.

The method according to claim 2 or 3, wherein the third model comprises a compression sub-model;

The compression sub-model is used to compress the first information to obtain the compressed information of the first information; the second information includes the compressed information of the first information.

The method of claim 5, wherein the fourth model includes a recovery sub-model;

The recovery sub-model is used to recover the compressed information of the first information to obtain the recovered information of the first information; and the third information includes the recovered information of the first information.

The method according to claim 2 or 3, wherein the third model includes a generation sub-model and a compression sub-model; wherein,

The generating sub-model is used to perform feature transformation on the first information to obtain a first feature vector corresponding to the first information;

The compression sub-model is used to compress the first feature vector to obtain the compressed information of the first feature vector; the second information includes the compressed information of the first feature vector.

The method of claim 7, wherein the fourth model includes a recovery sub-model;

The recovery sub-model is used to restore the compressed information of the first feature vector to obtain the recovery information of the first feature vector; the third information includes the recovery information of the first feature vector.

The method according to any one of claims 1 to 8, further comprising:

The terminal device receives the second model.

The method of claim 9, further comprising:

The terminal device receives the first model.

The method according to any one of claims 1 to 8, further comprising:

The terminal device receives the second model and the third model.

The method of claim 11, further comprising:

The terminal device receives the first model and/or the fourth model.

The method of claim 12, wherein the first model, the second model, the third model, the fourth model, sub-models of the first model, the second model The sub-model in, the sub-model in the third model or the sub-model in the fourth model is carried by one of the following: downlink control signaling, media access control MAC control element CE message, radio resource control RRC message , broadcast messages, downlink data transmission, and downlink data transmission for artificial intelligence business transmission requirements.

The method of claim 2, further comprising:

The terminal device receives a first coding model, the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes the second model;

The generation sub-model and the compression sub-model constitute the third model.

The method according to claim 14, wherein the terminal device uses a second model to process the first signal to obtain the first information, and the terminal device uses a third model to process the first information, Obtain second information, including:

The terminal device uses the first coding model to process the first signal to obtain the second information.

The method according to claim 14 or 15, wherein the first coding model is carried by one of the following: downlink control signaling, MAC CE message, RRC message, broadcast message, downlink data transmission, artificial intelligence type service transmission requirements Downstream data transmission.

The method of claim 2, further comprising:

The terminal device receives a second coding model, wherein the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes the second model;

The compressed sub-model constitutes the third model.

The method according to claim 17, wherein the terminal device processes the first signal using a second model to obtain first information, and the terminal device processes the first information using a third model to obtain second information, comprising:

The terminal device uses the second coding model to process the first signal to obtain the second information.

The method according to claim 17 or 18, wherein the second coding model is carried by one of the following: downlink control signaling, MAC CE message, RRC message, broadcast message, downlink data transmission, artificial intelligence type service transmission requirements Downstream data transmission.

The method according to any one of claims 1 to 8, further comprising:

The terminal device uses the input information and/or the first channel simulation module to jointly train the first initial model and the second initial model to obtain the trained first model and the second model.

The method according to claim 20, wherein the terminal device uses input information and/or the first channel simulation module to jointly train the first initial model and the second initial model, including:

The terminal device inputs the input information into the first initial model to obtain a first set of outputs of the first initial model, where the first set includes a plurality of first reference signals;

Input any first reference signal in the first set into the first channel simulation module to obtain a second reference signal;

Input the second reference signal into the second initial model to obtain channel information;

The first initial model and the second initial model are updated according to the first loss function.

The method of claim 21, wherein determining the first loss function includes:

Determine the channel information and parameters of the first channel simulation module based on at least one of the first set, the second reference signal, the channel information, and parameters in the first channel simulation module the degree of difference and/or quality of the reference signal;

The first loss function is determined based on the degree of difference between the channel information and parameters of the first channel simulation module and/or reference signal quality.

The method according to any one of claims 1 to 8, further comprising:

The terminal device uses at least one of the input information, the first channel simulation module and the second channel simulation module to jointly train the first initial model, the second initial model, the third initial model and the fourth initial model to obtain The first model, the second model, the third model and the fourth model after training.

The method according to claim 23, wherein the terminal device adopts at least one of the input information, the first channel simulation module and the second channel simulation module to perform the first initial model, the second initial model, the third initial model and the third initial model. model and the fourth initial model for joint training, including:

Input the channel information into the third initial model to obtain the compressed information of the channel information; wherein the third initial model includes a generated initial sub-model and a compressed initial sub-model, and the input of the generated initial sub-model As the input of the third initial model, the output of the generated initial sub-model is used as the output of the compressed initial sub-model, and the output of the compressed initial sub-model is used as the output of the third initial model; or, The third initial model includes a compressed initial submodel;

Inputting the compressed information of the channel information into the second channel simulation module to obtain equivalent received information of the compressed information of the channel information;

Inputting equivalent received information of the compressed information of the channel information into the fourth initial model to obtain output information of the fourth initial model;

Determine a second loss function based on at least one of the first set, the second reference signal, the channel information, parameters in the first channel simulation module, input information of the compressed initial submodel, and output information of the fourth initial model;

The first initial model, the second initial model, the third initial model and the fourth initial model are updated according to the second loss function.

The method of claim 24, wherein determining the second loss function includes:

Based on the first set, the second reference signal, the channel information, parameters in the first channel simulation module, input information of the compressed initial sub-model and output information of the fourth initial model Determine at least one of the reference signal quality, the degree of difference between the channel information and the parameters of the first channel simulation module, and the degree of difference between the output information of the fourth initial model and the input information of the compressed initial sub-model. at least one of;

Based on at least one of the reference signal quality, the degree of difference between the channel information and the parameters of the first channel simulation module, and the degree of difference between the output information of the fourth initial model and the input information of the compressed initial sub-model. One item, determine the second loss function.

The method according to claim 22 or 25, wherein the reference signal quality is represented by at least one of the following:

Cross-correlation between different first reference signals in the first set;

Cross-correlation between the first reference signal and other reference signals in the first set;

A peak-to-average power ratio of a first reference signal in the first set.

The method according to claim 22 or 25, wherein the quality of the reference signal is represented by at least one of the following:

Cross-correlation between different second reference signals;

The cross-correlation between the second reference signal and other reference signals;

The peak-to-average power ratio of the second reference signal.

Cross-correlation between different third reference signals;

A cross-correlation between the third reference signal and other reference signals;

The peak-to-average power ratio of the third reference signal;

Wherein, the third reference signal is obtained by processing the first reference signal based on the channel information.

The method according to claim 22 or 25, wherein the degree of difference is measured by distance and/or similarity.

The method according to any one of claims 20 to 29, wherein the input information includes at least one of the following: no input, noise, random numbers, sequences in a preset sequence set, channel type indication information, channel data samples information, wireless channel, or scene-related information.

The method according to claim 30, wherein the preset sequence set includes at least one of the following: m sequence set, golden sequence set, and zc sequence set.

The method according to claim 30, wherein the channel type indication information indicates at least one of the following: frequency information corresponding to the channel, environmental information corresponding to the channel, and scene information corresponding to the channel.

The method according to claim 30, wherein the wireless channel or scene-related information includes at least one of the following: channel signal-to-noise ratio, signal-to-interference-noise ratio, channel type, bandwidth information, and delay information.

The method according to claim 30, wherein the format of the sequence in the noise, random number or preset sequence set is the same as the format of the first initial model output data.

The method according to claim 30 or 34, wherein the format of the noise, random number or sequence in the preset sequence set includes at least one of the following formats: one-dimensional vector, two-dimensional matrix, and high-dimensional matrix.

The method according to claim 30, 34 or 35, wherein the format of the noise, random number or sequence in the preset sequence set is agreed upon through a protocol or signaling.

The method according to any one of claims 30 to 36, wherein the input information is used to input at least one of the following: a first initial model, a first channel simulation module, and a second initial model.

The method according to any one of claims 21, 22, 24 or 25, wherein the channel information is distributed in the first dimension and/or the second dimension.

The method according to any one of claims 21, 22, 24 or 25, wherein the channel information is distributed in at least one of a first dimension, a second dimension and a third dimension.

The method according to claim 38 or 39, wherein

The first dimension is a frequency domain dimension, and the channel information includes data distributed on M1 frequency domain granularities of the frequency domain dimension; the M1 is a positive integer.

The method according to claim 40, wherein the frequency domain granularity includes a RBs and/or b subcarriers, and a or b is a positive integer.

The method according to claim 38 or 39, wherein,

The first dimension is a time domain dimension, and the channel information includes data distributed on M2 delay granularities of the time domain dimension; the M2 is a positive integer.

The method according to claim 42, wherein the delay granularity includes at least one of the following: p1 microseconds, p2 symbol lengths, and p3 number of sampling points of a symbol, and p1, p2 or p3 is a positive integer.

The method of claim 43, wherein the symbols comprise OFDM symbols.

The method according to claim 38 or 39, wherein the second dimension is a spatial domain dimension.

The method according to claim 45, wherein the spatial domain dimension is an antenna dimension, the channel information includes data distributed on N1 first granularities of the antenna dimension, and the N1 is a positive integer.

The method of claim 46, wherein the first granularity includes a pair of transceiver antennas.

The method according to claim 45, wherein the spatial domain dimension is an angle domain dimension, the channel information includes data distributed on N2 second granularities of the angle domain dimension, and the N2 is a positive integer.

The method of claim 48, wherein the second granularity comprises angular spacing.

The method according to claim 39, wherein the third dimension includes a complex dimension, and the complex dimension includes 2 elements, which are respectively used to carry the real part and the imaginary part of the data included in the channel information.

The method according to any one of claims 39 to 50, wherein the channel information is distributed in a T-dimensional matrix, and the T-dimensional matrix is split into at least one of the first dimension, the second dimension and the third dimension. The matrix formed after dividing and/or combining, the T is a positive integer.

The method according to any one of claims 21, 22, 24, 25 or 38 to 51, wherein the channel information includes S groups of feature sequences with a length of U, and the S or U is a positive integer.

The method of claim 52, wherein S is 2, 4 or 8.

The method of claim 52 or 53, wherein U is 16, 32, 48, 64, 128 or 256.

The method according to any one of claims 20 to 54, further comprising:

The terminal device sends the first model.

The method of claim 55, further comprising:

The terminal device sends the second model.

The method according to any one of claims 23 to 54, further comprising:

The terminal device sends the first model and the fourth model.

The method of claim 57, further comprising:

The terminal device sends the second model and/or the third model.

The method of claim 58, wherein the first model, the second model, the third model, the fourth model, sub-models of the first model, the second model The sub-model in, the sub-model in the third model or the sub-model in the fourth model is carried by one of the following: uplink control signaling, MAC CE message, RRC message, broadcast message, uplink data transmission, for Uplink data transmission for artificial intelligence business transmission requirements.

The method according to any one of claims 20 to 54, further comprising:

The terminal device sends a first coding model, wherein the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The method according to claim 60, wherein the first coding model is carried by one of the following: uplink control signaling, MAC CE message, RRC message, broadcast message, uplink data transmission, and uplink data transmission for artificial intelligence service transmission requirements.

The method according to any one of claims 20 to 54, further comprising:

The terminal device sends a second coding model, the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The compressed sub-model constitutes the third model.

The method according to claim 62, wherein the second coding model is carried by one of the following: uplink control signaling, MAC CE message, RRC message, broadcast message, uplink data transmission, uplink data for artificial intelligence service transmission requirements transmission.

A method of communication including:

The network device sends a first signal, the first signal is generated by the first model; the first signal is used for processing by the second model to obtain the first information; the first information includes channel information;

The first model and the second model are jointly trained.

The method of claim 64, further comprising:

The network device receives second information, and the second information is obtained by processing the first information by a third model;

The network device uses a fourth model to process the second information to obtain third information;

A method according to claim 64 or 65, wherein said first signal comprises a reference signal.

The method according to any one of claims 64 to 66, wherein the second model includes a channel estimation sub-model;

The method of claim 65 or 66, wherein the third model includes a compression sub-model;

The method of claim 68, wherein the fourth model includes a recovery submodel;

The recovery sub-model is used to perform recovery processing on the compressed information of the first information to obtain the recovery information of the first information; the third information includes the recovery information of the first information.

The method of claim 65 or 66, wherein the third model includes a generation sub-model and a compression sub-model; wherein,

The method of claim 70, wherein the fourth model includes a recovery sub-model;

The recovery sub-model is used to recover the compressed information of the first feature vector to obtain the recovery information of the first feature vector; the third information includes the recovery information of the first feature vector.

The method according to any one of claims 64 to 71, further comprising:

The network device receives the first model.

The method of claim 72, further comprising:

The network device receives the second model.

The method according to any one of claims 64 to 71, further comprising:

The network device receives the first model and the fourth model.

The method of claim 74, further comprising:

The network device receives the second model and/or the third model.

The method of claim 75, wherein the first model, the second model, the third model, the fourth model, a sub-model of the first model, the second model The sub-model in, the sub-model in the third model or the sub-model in the fourth model is carried by one of the following: uplink control signaling, MAC CE message, RRC message, broadcast message, uplink data transmission, for Uplink data transmission for artificial intelligence business transmission requirements.

The method according to any one of claims 64 to 71, further comprising:

The network device receives a first coding model, the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The method according to claim 77, wherein the first coding model is carried by one of the following: uplink control signaling, MAC CE message, RRC message, broadcast message, uplink data transmission, uplink data for artificial intelligence service transmission requirements transmission.

The method according to any one of claims 64 to 71, further comprising:

The network device receives a second coding model, wherein the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The compression sub-model constitutes a third model.

The method of claim 79, wherein the second coding model is carried by one of the following: uplink control signaling, MAC CE message, RRC message, broadcast message, uplink data transmission, uplink data for artificial intelligence service transmission requirements transmission.

The method according to any one of claims 64 to 71, further comprising:

The network device uses the input information and/or the first channel simulation module to jointly train the first initial model and the second initial model to obtain the trained first model and the second model.

The method according to claim 81, wherein the network device uses input information and/or the first channel simulation module to jointly train the first initial model and the second initial model, including:

The network device inputs the input information into the first initial model to obtain a first set of outputs of the first initial model, where the first set includes a plurality of first reference signals;

Determine a first loss function based on at least one of the first set, the second reference signal, the channel information, and parameters in the first channel simulation module;

The method of claim 82, wherein determining the first loss function comprises:

The method according to any one of claims 64 to 71, further comprising:

The network device uses at least one of the input information, the first channel simulation module and the second channel simulation module to jointly train the first initial model, the second initial model, the third initial model and the fourth initial model to obtain The first model, the second model, the third model and the fourth model after training.

The method of claim 84, wherein the network device uses at least one of input information, a first channel simulation module, and a second channel simulation module to perform the first initial model, the second initial model, the third initial model, and the third initial model. model and the fourth initial model for joint training, including:

Input the compressed information of the channel information into the second channel simulation module to obtain equivalent reception information of the compressed information of the channel information;

Enter the equivalent received information of the compressed information of the channel information into the fourth initial model to obtain the output information of the fourth initial model;

Based on the first set, the second reference signal, the channel information, parameters in the first channel simulation module, input information of the compressed initial sub-model and output information of the fourth initial model At least one item of determines the second loss function;

The method of claim 85, wherein determining the second loss function includes:

The method according to claim 83 or 86, wherein the reference signal quality is represented by at least one of the following:

Cross-correlation between different first reference signals in the first set;

The peak-to-average power ratio of the first reference signal in the first set.

The method according to claim 83 or 86, wherein the quality of the reference signal is represented by at least one of the following:

Cross-correlation between different second reference signals;

The peak-to-average power ratio of the second reference signal.

Cross-correlation between different third reference signals;

The cross-correlation between the third reference signal and other reference signals;

The peak-to-average power ratio of the third reference signal;

The third reference signal is obtained by processing the first reference signal based on the channel information.

The method according to claim 83 or 86, wherein the degree of difference is measured by distance and/or similarity.

The method according to any one of claims 81 to 90, wherein the input information includes at least one of the following: no input, noise, random numbers, sequences in a preset sequence set, channel type indication information, channel data samples information, wireless channel, or scene-related information.

The method according to claim 91, wherein the preset sequence set includes at least one of the following: m sequence set, golden sequence set, and zc sequence set.

The method according to claim 91, wherein the channel type indication information indicates at least one of the following: frequency information corresponding to the channel, environmental information corresponding to the channel, and scene information corresponding to the channel.

The method according to claim 91, wherein the wireless channel or scene related information includes at least one of the following: signal-to-noise ratio, signal-to-interference-to-noise ratio, channel type, bandwidth information, and delay information of the channel.

The method according to claim 91, wherein the format of the sequence in the noise, random number or preset sequence set is the same as the format of the first initial model output data.

The method according to claim 91 or 95, wherein the format of the noise, random number or sequence in the preset sequence set includes at least one of the following formats: one-dimensional vector, two-dimensional matrix, and high-dimensional matrix.

The method according to claim 91, 95 or 96, wherein the format of the noise, random number or sequence in the preset sequence set is agreed upon through a protocol or signaling.

The method according to any one of claims 91 to 97, wherein the input information is used to input at least one of the following: a first initial model, a first channel simulation module, and a second initial model.

The method according to any one of claims 82, 83, 85 or 86, wherein the channel information is distributed in a first dimension and/or a second dimension.

The method according to any one of claims 82, 83, 85 or 86, wherein the channel information is distributed in at least one of a first dimension, a second dimension and a third dimension.

The method of claim 99 or 100, wherein,

The method according to claim 101, wherein the frequency domain granularity includes a RBs and/or b subcarriers, and a or b is a positive integer.

The method according to claim 99 or 100, wherein:

The method according to claim 103, wherein the delay granularity includes at least one of the following: p1 microseconds, p2 symbol length, p3 number of sampling points of a symbol, and p1, p2 or p3 is a positive integer.

The method of claim 104, wherein the symbols comprise OFDM symbols.

The method of claim 99 or 100, wherein the second dimension is a spatial domain dimension.

The method according to claim 106, wherein the spatial domain dimension is an antenna dimension, and the channel information includes data distributed at N1 first granularities of the antenna dimension, and N1 is a positive integer.

The method of claim 107, wherein the first granularity includes a pair of transceiver antennas.

The method according to claim 106, wherein the spatial domain dimension is an angle domain dimension, the channel information includes data distributed on N2 second granularities of the angle domain dimension, and the N2 is a positive integer.

The method of claim 109, wherein the second granularity includes angular spacing.

The method according to claim 100, wherein the third dimension includes a complex dimension, and the complex dimension includes two elements, respectively used to carry a real part and an imaginary part in the data included in the channel information.

The method according to any one of claims 100 to 111, wherein the channel information is distributed in a T-dimensional matrix, and the T-dimensional matrix is split into at least one of the first dimension, the second dimension and the third dimension. The matrix formed after dividing and/or combining, the T is a positive integer.

The method according to any one of claims 82, 83, 85, 86 or 99 to 112, wherein the channel information includes S groups of feature sequences with a length of U, and the S or U is a positive integer.

The method according to claim 113, wherein S is 2, 4 or 8.

The method of claim 113 or 114, wherein U is 16, 32, 48, 64, 128 or 256.

The method according to any one of claims 81 to 115, further comprising:

The network device sends the second model.

The method of claim 116, further comprising:

The network device sends the first model.

The method of any one of claims 84 to 115, further comprising:

The network device sends the second model and the third model.

The method of claim 118, further comprising:

The network device sends the first model and/or the fourth model.

The method according to claim 119, wherein the first model, the second model, the third model, the fourth model, the sub-model in the first model, the sub-model in the second model, the sub-model in the third model or the sub-model in the fourth model is carried by one of the following: downlink control signaling, MAC CE message, RRC message, broadcast message, downlink data transmission, downlink data transmission for artificial intelligence service transmission requirements.

The method according to any one of claims 81 to 115, further comprising:

The network device sends a first coding model, wherein the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The method according to claim 121, wherein the first coding model is carried by one of the following: downlink control signaling, MAC CE message, RRC message, broadcast message, downlink data transmission, downlink data for artificial intelligence service transmission requirements transmission.

The method according to any one of claims 81 to 115, further comprising:

The network device sends a second coding model, the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The compression sub-model constitutes a third model.

The method according to claim 123, wherein the second coding model is carried by one of the following: downlink control signaling, MAC CE message, RRC message, broadcast message, downlink data transmission, downlink data for artificial intelligence service transmission requirements transmission.

A model training method including:

Using the input information and/or the first channel simulation module, the first initial model and the second initial model are jointly trained to obtain the trained first model and the second model.

The method according to claim 125, wherein the joint training of the first initial model and the second initial model using input information and/or the first channel simulation module includes:

Inputting the input information into the first initial model to obtain a first set of outputs of the first initial model, wherein the first set includes a plurality of first reference signals;

The method of claim 126, wherein determining the first loss function includes:

Determine, based on at least one of the first set, the second reference signal, the channel information and the parameters in the first channel simulation module, a degree of difference between the channel information and the parameters of the first channel simulation module and/or a reference signal quality;

The method of claim 125, wherein the joint training comprises:

Using at least one of the input information, the first channel simulation module and the second channel simulation module, jointly train the first initial model, the second initial model, the third initial model and the fourth initial model to obtain the trained third First model, second model, third model and fourth model.

The method of claim 128, wherein the first initial model, the second initial model, the third initial model, The fourth initial model is jointly trained, including:

Enter the input information into the first initial model to obtain a first set of outputs from the first initial model, where the first set includes a plurality of first reference signals;

Inputting the channel information into the third initial model to obtain compressed information of the channel information; wherein the third initial model includes a generated initial sub-model and a compressed initial sub-model, and the input of the generated initial sub-model is used as the input of the third initial model, the output of the generated initial sub-model is used as the output of the compressed initial sub-model, and the output of the compressed initial sub-model is used as the output of the third initial model; or, the third initial model includes a compressed initial sub-model;

The method of claim 129, wherein determining the second loss function comprises:

The method according to claim 127 or 130, wherein the reference signal quality is represented by at least one of the following:

a mutual correlation between different first reference signals in the first set;

The peak-to-average power ratio of the first reference signal in the first set.

The method according to claim 127 or 130, wherein the quality of the reference signal is represented by at least one of the following:

cross-correlations between different second reference signals;

The peak-to-average power ratio of the second reference signal.

Cross-correlation between different third reference signals;

The peak-to-average power ratio of the third reference signal;

The method according to claim 127 or 130, wherein the degree of difference is measured by distance and/or similarity.

The method according to any one of claims 125 to 134, wherein the input information includes at least one of the following: no input, noise, random numbers, sequences in a preset sequence set, channel type indication information, channel data samples information, wireless channel, or scene-related information.

The method according to claim 134, wherein the preset sequence set includes at least one of the following: m sequence set, golden sequence set, and zc sequence set.

The method according to claim 134, wherein the channel type indication information indicates at least one of the following: frequency information corresponding to the channel, environmental information corresponding to the channel, and scene information corresponding to the channel.

The method according to claim 134, wherein the wireless channel or scene related information includes at least one of the following: signal-to-noise ratio, signal-to-interference-to-noise ratio, channel type, bandwidth information, and delay information of the channel.

The method of claim 134, wherein the format of the noise, random numbers or sequences in the preset sequence set is the same as the format of the first initial model output data.

The method according to claim 135 or 139, wherein the format of the noise, random number or sequence in the preset sequence set includes at least one of the following formats: one-dimensional vector, two-dimensional matrix, and high-dimensional matrix.

The method according to claim 135, 139 or 140, wherein the format of the noise, random number or sequence in the preset sequence set is agreed through a protocol or signaling.

The method according to any one of claims 125 to 141, wherein the input information is used to input at least one of the following: a first initial model, a first channel simulation module, and a second initial model.

The method of claim 126, 127, 129 or 130, wherein the channel information is distributed in a first dimension and/or a second dimension.

The method of claim 126, 127, 129 or 130, wherein the channel information is distributed in at least one of a first dimension, a second dimension and a third dimension.

The method of claim 143 or 144, wherein,

The method according to claim 145, wherein the frequency domain granularity includes a RBs and/or b subcarriers, and a or b is a positive integer.

The method of claim 143 or 144, wherein,

The method according to claim 147, wherein the delay granularity includes at least one of the following: p1 microseconds, p2 symbol length, p3 symbol sampling point number, and the p1, p2 or p3 is positive integer.

The method of claim 87, wherein the symbols comprise OFDM symbols.

The method of claim 143 or 144, wherein the second dimension is a spatial domain dimension.

The method according to claim 150, wherein the spatial domain dimension is an antenna dimension, and the channel information includes data distributed at N1 first granularities of the antenna dimension, and N1 is a positive integer.

The method of claim 151, wherein the first granularity includes a pair of transceiver antennas.

The method according to claim 150, wherein the spatial domain dimension is an angle domain dimension, the channel information includes data distributed on N2 second granularities of the angle domain dimension, and the N2 is a positive integer.

The method of claim 153, wherein the second granularity includes angular spacing.

The method of claim 144, wherein the third dimension includes a complex dimension, and the complex dimension includes 2 elements, respectively used to carry a real part and an imaginary part in the data included in the channel information.

The method according to any one of claims 144 to 155, wherein the channel information is distributed in a T-dimensional matrix, and the T-dimensional matrix is split into at least one of the first dimension, the second dimension and the third dimension. The matrix formed after dividing and/or combining, the T is a positive integer.

The method according to any one of claims 143 to 156, wherein the channel information includes S groups of feature sequences with a length of U, and the S or U is a positive integer.

The method of claim 157, wherein S is 2, 4 or 8.

The method of claim 157 or 158, wherein U is 16, 32, 48, 64, 128 or 256.

A terminal device including:

A first receiving module, configured to receive a first signal, where the first signal is generated by the first model;

A first processing module used to process the first signal using a second model to obtain first information; the first information includes channel information;

The first model and the second model are jointly trained.

The terminal device according to claim 160, further comprising:

The second processing module is used to process the first information using a third model to obtain the second information;

The first sending module is used to send the second information, and the second information is used for processing by the fourth model to obtain the third information;

Among them, the first model, the second model, the third model and the fourth model are obtained by joint training.

The terminal device according to claim 160 or 161, wherein the first signal includes a reference signal.

The terminal device according to any one of claims 160 to 162, wherein the second model includes a channel estimation sub-model;

The terminal device according to claim 161 or 162, wherein the third model includes a compression sub-model;

The compression sub-model is used to compress the first information to obtain compressed information of the first information; the second information includes the compressed information of the first information.

The terminal device according to claim 164, wherein the fourth model includes a recovery sub-model;

The terminal device according to claim 161 or 162, wherein the third model includes a generation sub-model and a compression sub-model; wherein,

The terminal device according to claim 166, wherein the fourth model includes a recovery sub-model;

The terminal device according to any one of claims 160 to 167, further comprising:

The second receiving module is used to receive the second model.

The terminal device according to claim 168, wherein,

The second receiving module is also used to receive the first model.

The third receiving module is used to receive the second model and the third model.

The terminal device according to claim 170, wherein,

The third receiving module is also used to receive the first model and/or the fourth model.

The terminal device according to claim 161, further comprising:

The fourth receiving module is used to receive the first coding model, the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes the second model;

The terminal device according to claim 161, further comprising:

A fifth receiving module is used to receive a second coding model, wherein the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes the second model;

The compressed sub-model constitutes the third model.

The terminal device according to any one of claims 160 to 173, further comprising:

The first training module is configured to use the input information and/or the first channel simulation module to jointly train the first initial model and the second initial model to obtain the trained first model and the second model.

The second training module is used to use at least one of the input information, the first channel simulation module and the second channel simulation module to combine the first initial model, the second initial model, the third initial model and the fourth initial model. Training is performed to obtain the trained first model, the second model, the third model and the fourth model.

The terminal device according to claim 174 or 175, further comprising:

The second sending module is used to send the first model.

The terminal device according to claim 176, wherein,

The second sending module is further used to send the second model.

The method according to claim 174 or 175, further comprising:

The third sending module is used to send the first model and the fourth model.

The terminal device according to claim 178, wherein,

The third sending module is also used to send the second model and/or the third model.

The terminal device according to claim 174 or 175, further comprising:

The fourth sending module is used to send the first coding model, the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The terminal device according to claim 174 or 175, further comprising:

The fifth sending module is used to send the second coding model, the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The compression sub-model constitutes a third model.

A network device that includes:

a sixth sending module, configured to send a first signal, wherein the first signal is generated by a first model; the first signal is used for being processed by a second model to obtain first information; the first information includes channel information;

The first model and the second model are jointly trained.

The network device of claim 182, further comprising:

A sixth receiving module, configured to receive second information, the second information being obtained by processing the first information by the third model;

A third processing module, configured to use a fourth model to process the second information to obtain third information;

Wherein, the first model, the second model, the third model and the fourth model are obtained through joint training.

A network device according to claim 182 or 183, wherein the first signal includes a reference signal.

The network device according to any one of claims 182 to 184, wherein the second model includes a channel estimation sub-model;

The network device of claim 183 or 184, wherein the third model includes a compression sub-model;

The network device of claim 186, wherein the fourth model includes a recovery sub-model;

The network device according to claim 183 or 184, wherein the third model includes a generation sub-model and a compression sub-model; wherein,

The network device of claim 188, wherein the fourth model comprises a recovery sub-model;

The network device according to any one of claims 182 to 189, further comprising:

A seventh receiving module is used to receive the first model.

The network device of claim 190, wherein:

The seventh receiving module is also used to receive the second model.

An eighth receiving module is used to receive the first model and the fourth model.

The network device of claim 192, wherein:

The eighth receiving module is also used to receive the second model and/or the third model.

The ninth receiving module is used to receive the first coding model, the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The tenth receiving module is used to receive the second coding model, the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The compression sub-model constitutes a third model.

The third training module is configured to use the input information and/or the first channel simulation module to jointly train the first initial model and the second initial model to obtain the trained first model and the second model.

The fourth training module is used to jointly train the first initial model, the second initial model, the third initial model, and the fourth initial model by using input information, at least one of the first channel simulation module and the second channel simulation module to obtain the first model, the second model, the third model, and the fourth model after training.

The network device according to claim 196 or 197, further comprising:

The seventh sending module is used to send the second model.

The network device of claim 198, wherein:

The seventh sending module is also used to send the first model.

The network device according to claim 196 or 197, further comprising:

An eighth sending module is used to send the second model and the third model.

The network device of claim 200, wherein:

The eighth sending module is further used to send the first model and/or the fourth model.

The network device according to claim 196 or 197, further comprising:

The ninth sending module is used to send the first coding model, the first coding model includes a channel estimation sub-model, a generation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The network device according to claim 196 or 197, further comprising:

The tenth sending module is used to send the second coding model, the second coding model includes a channel estimation sub-model and a compression sub-model; wherein,

The channel estimation sub-model constitutes a second model;

The compression sub-model constitutes a third model.

A model training device including:

The joint training module is used to jointly train the first initial model and the second initial model using the input information and/or the first channel simulation module to obtain the trained first model and the second model.

The device of claim 204, wherein the joint training module is used for:

Inputting any of the first reference signals in the first set into the first channel simulation module to obtain a second reference signal;

The device of claim 205, wherein determining the first loss function includes:

The device of claim 204, wherein the joint training module is used for:

The first initial model, the second initial model, the third initial model and the fourth initial model are jointly trained using at least one of the input information, the first channel simulation module and the second channel simulation module to obtain the trained first model, the second model, the third model and the fourth model.

The apparatus of claim 207, wherein the joint training module is configured to:

The apparatus of claim 208, wherein determining the second loss function includes:

The second loss function is determined based on at least one of the reference signal quality, the degree of difference between the channel information and the parameters of the first channel simulation module, and the degree of difference between the output information of the fourth initial model and the input information of the compressed initial submodel.

The device according to claim 206 or 209, wherein the reference signal quality is represented by at least one of the following:

Cross-correlation between different first reference signals in the first set;

The peak-to-average power ratio of the first reference signal in the first set.

The apparatus according to claim 206 or 209, wherein the quality of the reference signal is represented by at least one of the following:

Cross-correlation between different second reference signals;

a cross-correlation between the second reference signal and other reference signals;

The peak-to-average power ratio of the second reference signal.

Cross-correlation between different third reference signals;

a peak-to-average power ratio of the third reference signal;

A terminal device includes: a processor, a memory and a transceiver, the memory is used to store a computer program, the processor is used to call and run the computer program stored in the memory, and control the transceiver to execute as follows: The method of any one of claims 1 to 63.

A network device includes: a processor, a memory and a transceiver, the memory is used to store a computer program, the processor is used to call and run the computer program stored in the memory, and control the transceiver to execute as follows: The method of any one of claims 64 to 124.

A model training device, including: a processor and a memory, the memory is used to store a computer program, the processor is used to call and run the computer program stored in the memory, and execute as claimed in any one of claims 125 to 159 method described.

A chip includes: a processor for calling and running a computer program from a memory, so that a device installed with the chip executes the method according to any one of claims 125 to 159.

A chip includes: a processor for calling and running a computer program from a memory, so that a device installed with the chip executes the method according to any one of claims 64 to 124.

A computer-readable storage medium for storing a computer program, wherein the computer program causes a computer to execute the method according to any one of claims 1 to 63.

A computer-readable storage medium for storing a computer program that causes a computer to perform the method according to any one of claims 64 to 124.

A computer-readable storage medium for storing a computer program, wherein the computer program causes a computer to execute the method as claimed in any one of claims 125 to 159.

A computer program product comprising computer program instructions, the computer program instructions causing a computer to perform the method according to any one of claims 1 to 63.

A computer program product comprising computer program instructions, the computer program instructions causing a computer to execute the method as claimed in any one of claims 64 to 124.

A computer program product comprising computer program instructions causing a computer to perform the method according to any one of claims 125 to 159.

A computer program causing a computer to perform the method according to any one of claims 1 to 63.

A computer program causing a computer to perform the method of any one of claims 64 to 124.

A computer program causing a computer to perform the method of any one of claims 125 to 159.