Communication method, device and system

Abstract

The embodiment of the invention provides a communication method, device and system. The method may comprise: a reference signal receiving end receiving a first reference signal; and the reference signal receiving end inputs the first reference signal into a first target model for processing to obtain K first target branch adaptive layers in the first target model and weights of the K first target branch adaptive layers. Wherein the first target model comprises N first target branch adaptive layers. The reference signal receiving end also receives the first data. And then, the reference signal receiving end inputs the first data into the K first target branch adaptive layers for processing to obtain processed data. Wherein the maximum weight in the weights of the K first target branch adaptive layers is not less than a preset value. The actual channel distribution is represented based on the determined K first target branch adaptive layers and the K weights to cope with the dynamic change of the channel, so that the system performance can be improved.

Description

Communication method, device and system Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a communication method, device, and system. Background Art

[0002] In recent years, the development of deep learning has attracted research from academia and industry on wireless communication technologies based on deep learning. The research results have confirmed that deep learning technology can improve the performance of wireless communication systems and has the potential to be applied in the physical layer for interference adjustment, channel estimation and signal detection, signal processing and other aspects.

[0003] The neural network transceiver combines the transmitter and receiver to optimize specific performance metrics and channel models, eliminating the need for prior expert knowledge. This allows for customized, self-evolving air interfaces and approaches the Shannon limit. However, in real-world channel scenarios, channel distributions change dynamically. This requires the neural network model to be able to discern channel changes in the current environment and quickly adapt to new scenarios with minimal training overhead.

[0004] To address the issues caused by varying channel distributions, Park S et al. combined meta-learning methods to propose a meta-autoencoder network structure. This approach treats end-to-end training under various channels as separate subtasks, thereby training an initial transceiver network with potential for each channel. The main training task aims to obtain optimal network initialization parameters, enabling the system to converge on training for any channel with a minimum number of stochastic gradient descent (SGD) steps, thereby rapidly adapting to time-varying channel variations.

[0005] Since the channel distribution in actual scenarios changes dynamically, it may not exist in the existing subtasks in the above meta-learning method. Therefore, more subtasks need to be iteratively calculated, which makes the training computational overhead larger. Moreover, each round of iteration requires calculating the channel conditions in all subtasks, so each subtask needs to be trained until fitting and feedback is obtained, which is more time-consuming.

[0006] Summary of the Invention

[0007] The present application discloses a communication method, device and system that can solve the problems of neural network training and deduction caused by dynamic changes in channels in the environment.

[0008] In a first aspect, an embodiment of the present application provides a communication method. The method can be performed by a communication device, or by a component of the communication device (e.g., a chip (system)). The method includes: a reference signal receiving end receives a first reference signal. Then, the reference signal receiving end inputs the first reference signal into a first target model for processing, and obtains K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model. The first target model includes N first target branch adaptive layers, K is not greater than N, and K and N are both positive integers. The reference signal receiving end also sends first information. The first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers. The K first target branch adaptive layers and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be transmitted.

[0009] In this embodiment of the present application, a reference signal receiver determines K first target branch adaptive layers and weights for these K first target branch adaptive layers out of N first target branch adaptive layers of a first target model based on a received reference signal. This approach characterizes the actual channel distribution based on the determined K first target branch adaptive layers and K weights to address dynamic channel changes, thereby improving system performance.

[0010] In one possible implementation, the method further includes: receiving, at a reference signal receiving end, first data obtained by processing the coded data to be transmitted by the transmitting end; and then inputting, by the reference signal receiving end, the first data into the K first target branch adaptation layers for processing to obtain processed data.

[0011] In a possible implementation, the method further includes: a reference signal receiving end sending a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0012] By performing dictionary alignment, the transmitting end selects the optimal processing method under the corresponding channel distribution to process the encoded data, which can improve communication performance.

[0013] In another possible implementation, the method further includes: receiving, at a reference signal receiving end, a second channel dictionary set. Then, performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set. Finally, the reference signal receiving end transmits first indication information indicating the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0014] By performing dictionary alignment, the transmitting end selects the optimal processing method under the corresponding channel distribution to process the encoded data, which can improve communication performance.

[0015] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0016] In one possible implementation, the method further includes: receiving a second reference signal at a reference signal receiving end. Then, the reference signal receiving end transmits third information, where the third information is identical to the first information, or the third information instructs the transmitting end to transmit the reference signal after a first time interval. The third information is derived based on the second reference signal. Finally, the reference signal receiving end receives data transmitted by the transmitting end within the first time interval.

[0017] The transmitting end sends data within the first time, which can be understood as the transmitting end not sending a reference signal within the first time. In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0018] In a second aspect, an embodiment of the present application provides a communication method. The method can be performed by a communication device, or by a component of the communication device (e.g., a chip (system)). The method includes: a reference signal receiving end receives a first reference signal. Then, the reference signal receiving end inputs the first reference signal into a first target model for processing, and obtains K first target branch adaptive layers in the first target model and weights of the K first target branch adaptive layers. The first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers.

[0019] The reference signal receiving end also transmits fourth information indicating K second target branch adaptive layers and weights of the K second target branch adaptive layers. The weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be transmitted. The weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0020] In an embodiment of the present application, a reference signal receiver determines K first target branch adaptive layers and their weights among the N first target branch adaptive layers of a first target model based on a received reference signal. K second target branch adaptive layers and their weights are then indicated to the transmitter. This approach characterizes the actual channel distribution based on the determined K first target branch adaptive layers and K weights to address dynamic channel changes, thereby improving system performance. Furthermore, by indicating this to the transmitter, the transmitter processes the coded data to be transmitted, thereby improving communication performance.

[0021] In a possible implementation, the method further includes: the reference signal receiving end receiving first data, where the first data is obtained by the transmitting end processing the coded data to be transmitted.

[0022] The reference signal receiving end further inputs the first data into the K first target branch adaptive layers for processing to obtain processed data.

[0023] In one possible implementation, the method further includes: transmitting, by a reference signal receiving end, a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer. The reference signal receiving end further receives first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layers in the second channel dictionary set and the first target branch adaptation layers in the first channel dictionary set.

[0024] By performing dictionary alignment, the reference signal receiving end can determine the weights of the above-mentioned K second target branch adaptive layers and the K second target branch adaptive layers to instruct the transmitting end to process the encoded data according to the optimal processing method under the corresponding channel distribution, thereby improving communication performance.

[0025] In another possible implementation, the method further includes: receiving, at the reference signal receiving end, a second channel dictionary set. Then, performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0026] By performing dictionary alignment, the reference signal receiving end can determine the weights of the above-mentioned K second target branch adaptive layers and the K second target branch adaptive layers to instruct the transmitting end to process the encoded data according to the optimal processing method under the corresponding channel distribution, thereby improving communication performance.

[0027] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0028] In one possible implementation, the method further includes: receiving, at a reference signal receiving end, a second reference signal. The reference signal receiving end then transmits fifth information, where the fifth information is identical to the fourth information, or the fifth information instructs the transmitting end to transmit the reference signal after a first time interval, the fifth information being derived based on the second reference signal. The reference signal receiving end further receives data transmitted by the transmitting end within the first time interval.

[0029] The transmitting end sends data within the first time, which can be understood as the transmitting end not sending a reference signal within the first time. In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0030] In a third aspect, an embodiment of the present application provides a communication method. The method can be performed by a communication device, or by a component of the communication device (e.g., a chip (system)). The method includes: a transmitting end sends a first reference signal. Then, the transmitting end receives first information, where the first information indicates K first target branch adaptive layers and weights of the K first target branch adaptive layers in a first target model of the first reference signal receiving end. The first information is obtained based on the first reference signal, and K is a positive integer.

[0031] In this embodiment of the present application, the information received by the transmitting end indicates K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model of the first reference signal receiving end. This allows the actual channel distribution to be characterized based on the K first target branch adaptive layers and the K weights to cope with dynamic channel changes, thereby improving communication performance.

[0032] In one possible implementation, the method further includes: a transmitting end determining, based on the first information, K second target branch adaptive layers corresponding to the K first target branch adaptive layers and weights of the K second target branch adaptive layers. The transmitting end then inputs the coded data to be transmitted into the K second target branch adaptive layers of the second target model for processing, thereby obtaining first data. Finally, the transmitting end transmits the first data.

[0033] In one possible implementation, when M is less than K, the transmitting end updates the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers. The third target model includes M second target branch adaptive layers, where M is a positive integer.

[0034] In this way, the corresponding target branch adaptation layer and weight can be determined based on the received information to process the encoded data, thereby improving communication performance.

[0035] In one possible implementation, the method further includes: receiving, at a transmitting end, a first channel dictionary set; and performing dictionary alignment on the first channel dictionary set and a second channel dictionary set to obtain a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0036] By receiving the channel dictionary set, both ends can align the dictionaries and then select the optimal processing method under the corresponding channel distribution to process the encoded data, thereby improving communication performance.

[0037] In another possible implementation, the method further includes: the transmitting end transmitting a second channel dictionary set, and the transmitting end further receiving first indication information, the first indication information being used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0038] By performing dictionary alignment, the transmitting end selects the optimal processing method under the corresponding channel distribution to process the encoded data, which can improve communication performance.

[0039] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0040] In one possible implementation, the method further includes: the transmitting end transmitting a second reference signal. Then, the transmitting end receives third information, where the third information is the same as the first information, or the third information indicates that the reference signal should be transmitted after a first time interval. The third information is obtained based on the second reference signal. Finally, the transmitting end transmits data within the first time interval.

[0041] The transmitting end sends data within the first time, which can be understood as the transmitting end not sending a reference signal within the first time. In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0042] In a fourth aspect, embodiments of the present application provide a communication method. This method can be performed by a communication device, or by a component of the communication device (e.g., a chip (system)). The method includes: a transmitting end transmitting a first reference signal. The transmitting end then receives fourth information. The fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers. The fourth information is obtained based on the first reference signal, where K is a positive integer.

[0043] In this embodiment of the present application, the information received by the transmitting end indicates K second target branch adaptive layers and the weights of the K second target branch adaptive layers. In this way, the actual channel distribution is characterized based on the K target branch adaptive layers and the K weights to cope with dynamic changes in the channel, thereby improving communication performance.

[0044] In a possible implementation, the method further includes:

[0045] The transmitting end inputs the coded data to be sent into the K second target branch adaptive layers of the second target model for processing based on the fourth information to obtain the first data.

[0046] Then, the sending end sends the first data.

[0047] In this way, the transmitting end processes the coded data according to the optimal processing method under the corresponding channel distribution, which can improve the communication performance.

[0048] In a possible implementation, the method further includes:

[0049] The transmitting end receives the first channel dictionary set.

[0050] Then, the transmitting end performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0051] The transmitting end further sends first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0052] By receiving the channel dictionary set, both ends can align the dictionaries, and then instruct the reference signal receiving end to determine the corresponding target branch adaptation layer and weight, so that the transmitting end can process the encoded data based on the optimal processing method under the corresponding channel distribution, which can improve communication performance.

[0053] In another possible implementation, the method further includes:

[0054] The transmitting end sends a second channel dictionary set, which is used by the reference signal receiving end to determine the correspondence between the N first target branch adaptation layers of the reference signal receiving end and the M second target branch adaptation layers of the transmitting end, where M and N are both positive integers.

[0055] By sending a channel dictionary set, the reference signal receiving end can align the dictionaries at both ends to determine the corresponding target branch adaptation layer and weight, so that the transmitting end can process the encoded data based on the optimal processing method under the corresponding channel distribution, which can improve communication performance.

[0056] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0057] In a possible implementation, the method further includes:

[0058] The transmitting end sends a second reference signal.

[0059] Then, the transmitting end receives fifth information, where the fifth information is the same as the fourth information, or the fifth information indicates that the reference signal is sent again after a first time interval, and the fifth information is obtained based on the second reference signal;

[0060] The sending end sends data within the first time.

[0061] The transmitting end sends data within the first time, which can be understood as the transmitting end not sending a reference signal within the first time. In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0062] In a fifth aspect, an embodiment of the present application provides a communication method. The method can be executed by a communication device, or by a component of the communication device (such as a chip (system)). The method includes: a reference signal receiving end receives a first reference signal. The reference signal receiving end inputs the first reference signal into a first target model for processing, and obtains K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model. The first target model includes N first target branch adaptive layers, K is not greater than N, and K and N are both positive integers. The reference signal receiving end also receives first data. Then, the reference signal receiving end inputs the first data into the K first target branch adaptive layers for processing to obtain processed data. The maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value.

[0063] In an embodiment of the present application, a reference signal receiver determines K first target branch adaptive layers and weights for these K first target branch adaptive layers out of N first target branch adaptive layers of a first target model based on a received reference signal. When the maximum weight among the K first target branch adaptive layers is no less than a preset value, the received first data is input into these K first target branch adaptive layers for processing, thereby obtaining processed data. This approach characterizes the actual channel distribution based on the determined K first target branch adaptive layers and K weights to address dynamic channel changes, thereby improving system performance.

[0064] In one possible implementation, the first target model further includes a first target channel feature extraction network and a target sparse gating module, wherein the first target channel feature extraction network is used to process the first reference signal to obtain channel distribution information and channel category information corresponding to the channel distribution information.

[0065] The target sparse gating module is used to calculate the K first target branch adaptive layers and the weights of the K first target branch adaptive layers according to the channel distribution information and the channel category information.

[0066] The first target channel feature extraction network and the target sparse gating module can be two independent modules or integrated into one, and this solution does not impose any restrictions on this.

[0067] In a possible implementation, the K first target branch adaptive layers are used to process the first data respectively to obtain data processed by the K first target branch adaptive layers respectively.

[0068] The weights of the K first target branch adaptive layers are used to perform weighted sum processing on the data processed respectively by the K first target branch adaptive layers to obtain the processed data.

[0069] In a possible implementation, the method further includes: when a maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, the reference signal receiving end decodes the first data to obtain processed data.

[0070] In one possible implementation, the method further includes: a reference signal receiving end transmitting first information. The first information indicates the K first target branch adaptive layers and weights of the K first target branch adaptive layers. The K first target branch adaptive layers and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be transmitted to obtain the first data.

[0071] When the maximum weight among the K weights of the first target branch adaptive layer is not less than a preset value, a message is sent to the transmitter, prompting it to process the coded data before sending it. This allows the transmitter to process the coded data in the optimal manner for the corresponding channel distribution, improving communication performance.

[0072] In a possible implementation, the method further includes: when the maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, the reference signal receiving end sends second information, and the second information instructs the transmitting end to directly send the first data.

[0073] In a possible implementation, the method further includes: a reference signal receiving end sending a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0074] By sending a channel dictionary set, both ends can align the dictionaries, so that the transmitting end can select the optimal processing method under the corresponding channel distribution to process the encoded data, which can improve communication performance.

[0075] In another possible implementation, the method further includes: receiving, at a reference signal receiving end, a second channel dictionary set; performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; and transmitting, at the reference signal receiving end, first indication information indicating the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0076] By performing dictionary alignment, the transmitting end selects the optimal processing method under the corresponding channel distribution to process the encoded data, which can improve communication performance.

[0077] In a possible implementation, the method further includes:

[0078] The reference signal receiving end transmits fourth information indicating K second target branch adaptive layers and weights of the K second target branch adaptive layers. The weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be transmitted to obtain the first data. The weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0079] In a possible implementation, the method further includes:

[0080] The reference signal receiving end sends a first channel dictionary set.

[0081] The reference signal receiving end further receives first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0082] In a possible implementation, the method further includes:

[0083] The reference signal receiving end receives the second channel dictionary set.

[0084] The reference signal receiving end further performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0085] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0086] In one possible implementation, the method further includes: a reference signal receiving end receiving a second reference signal. Then, the reference signal receiving end sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send the reference signal after a first time interval. The third information is obtained based on the second reference signal. The reference signal receiving end also receives data sent by the transmitting end within the first time interval.

[0087] In another possible implementation, the method further includes: receiving a second reference signal at a reference signal receiving end. Then, the reference signal receiving end transmits fifth information, where the fifth information is the same as the fourth information, or the fifth information instructs the transmitting end to transmit the reference signal after a first time interval. The fifth information is obtained based on the second reference signal. The reference signal receiving end also receives data transmitted by the transmitting end within the first time interval.

[0088] The transmitting end sending data within the first time period can be understood as not sending a reference signal within the first time period. This reduces pilot transmission overhead and increases the spectral efficiency of data transmission. This example can be found in the description of the third information above and will not be further elaborated here.

[0089] In a sixth aspect, an embodiment of the present application provides a communication method. The method can be executed by a communication device, or by a component of a communication device (e.g., a chip (system)). The method includes: a transmitting end sends a first reference signal. When the transmitting end receives the first information, the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers in the first target model of the first reference signal receiving end, and based on the first information, the K second target branch adaptive layers corresponding to the K first target branch adaptive layers and the weights of the K second target branch adaptive layers are determined. The first information is obtained based on the first reference signal, and K is a positive integer. Then, the transmitting end inputs the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data. Finally, the transmitting end sends the first data.

[0090] In this embodiment of the present application, the transmitter processes the coded data to be transmitted based on the received information before transmitting it. This allows the actual channel distribution to be characterized based on the determined K first target branch adaptive layers and K weights to cope with dynamic channel changes. Furthermore, by selecting the optimal processing method for the corresponding channel distribution to process the coded data, communication performance can be improved.

[0091] In one possible implementation, when M is less than K, the transmitting end updates the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers. The third target model includes M second target branch adaptive layers, where M is a positive integer.

[0092] In this way, the corresponding target branch adaptation layer and weight can be determined based on the received information to process the encoded data, thereby improving communication performance.

[0093] In a possible implementation, the K second target branch adaptive layers are used to respectively process the coded data to be sent, to obtain data processed by the K second target branch adaptive layers.

[0094] The weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed respectively by the K second target branch adaptive layers to obtain the first data.

[0095] In one possible implementation, the method further includes: when the transmitting end receives second information indicating that the coded data to be transmitted should be directly transmitted, the transmitting end transmits the coded data to be transmitted, wherein the second information is obtained based on the first reference signal.

[0096] In one possible implementation, the method further includes: receiving, at a transmitting end, a first channel dictionary set; and performing dictionary alignment on the first channel dictionary set and a second channel dictionary set to obtain a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0097] By receiving the channel dictionary set, both ends can align the dictionaries and then select the optimal processing method under the corresponding channel distribution to process the encoded data, thereby improving communication performance.

[0098] In another possible implementation, the method further includes: the transmitting end transmitting a second channel dictionary set, and then receiving first indication information, the first indication information being used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0099] By performing dictionary alignment, the transmitting end selects the optimal processing method under the corresponding channel distribution to process the encoded data, which can improve communication performance.

[0100] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0101] In one possible implementation, the method further includes: the transmitting end transmitting a second reference signal. The transmitting end then receives third information, where the third information is identical to the first information; or, the third information indicates that the reference signal should be transmitted after a first time interval. The third information is obtained based on the second reference signal. Finally, the transmitting end transmits data within the first time interval.

[0102] The transmitting end sends data within the first time, which can be understood as the transmitting end not sending a reference signal within the first time. In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0103] In a seventh aspect, an embodiment of the present application provides a communication method. The method can be executed by a communication device, or by a component of a communication device (e.g., a chip (system)). The method includes: a transmitting end sends a first reference signal. Then, when the transmitting end receives fourth information, the fourth information indicates K second target branch adaptive layers and the weights of the K second target branch adaptive layers. According to the fourth information, the coded data to be sent is input into the K second target branch adaptive layers of the second target model for processing to obtain the first data. The fourth information is obtained based on the first reference signal, and K is a positive integer. The transmitting end also sends the first data.

[0104] In this embodiment of the present application, the information received by the transmitter indicates K second target branch adaptive layers and the weights of these K second target branch adaptive layers. This allows the actual channel distribution to be represented based on the K target branch adaptive layers and K weights to address dynamic channel changes, thereby improving communication performance. Furthermore, the transmitter processes the encoded data according to the optimal processing method for the corresponding channel distribution, thereby improving communication performance.

[0105] In one possible implementation, the K second target branch adaptive layers are used to respectively process the coded data to be sent to obtain the data respectively processed by the K second target branch adaptive layers; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data respectively processed by the K second target branch adaptive layers to obtain the first data.

[0106] In a possible implementation, the method further includes:

[0107] When receiving the second information, the second information indicates that the coded data to be sent should be sent directly, and the transmitting end sends the coded data to be sent. The second information is obtained based on the first reference signal.

[0108] In a possible implementation, the method further includes:

[0109] The transmitting end receives a first channel dictionary set. The transmitting end then performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set. The transmitting end also transmits first indication information indicating the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0110] By receiving the channel dictionary set, both ends can align the dictionaries, and then instruct the reference signal receiving end to determine the corresponding target branch adaptation layer and weight, so that the transmitting end can process the encoded data based on the optimal processing method under the corresponding channel distribution, which can improve communication performance.

[0111] In a possible implementation, the method further includes:

[0112] The transmitting end sends a second channel dictionary set, which is used by the reference signal receiving end to determine the correspondence between the N first target branch adaptation layers of the reference signal receiving end and the M second target branch adaptation layers of the transmitting end, where M and N are both positive integers.

[0113] By sending a channel dictionary set, the reference signal receiving end can align the dictionaries at both ends to determine the corresponding target branch adaptation layer and weight, so that the transmitting end can process the encoded data based on the optimal processing method under the corresponding channel distribution, which can improve communication performance.

[0114] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0115] In a possible implementation, the method further includes:

[0116] The transmitting end sends a second reference signal.

[0117] Then, the transmitting end receives fifth information. The fifth information is the same as the fourth information; or the fifth information is used to indicate that the reference signal is sent after the first time interval, and the fifth information is obtained based on the second reference signal;

[0118] The sending end also sends data within the first time.

[0119] The transmitting end sends data within the first time, which can be understood as the transmitting end not sending a reference signal within the first time. In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0120] In an eighth aspect, an embodiment of the present application provides a target model training method. The method can be executed by a communication device, or by a component of a communication device (such as a chip (system)). The target model includes a target encoding network, a target decoding network and a first target model, and the first target model includes a first target channel feature extraction network, N first target branch adaptive layers and a target sparse gating module. The method includes: training the initial channel feature extraction network to obtain the first target channel feature extraction network and the first channel dictionary set. Then, according to the target encoding network and the target decoding network, the N initial branch adaptive layers are trained to obtain the N first target branch adaptive layers, where N is the number of first target branch adaptive layers in the first channel dictionary set. Finally, according to the target encoding network, the target decoding network, the first target channel feature extraction network and the N first target branch adaptive layers, the initial sparse gating module is trained to obtain the target sparse gating module.

[0121] In this embodiment of the present application, the first target channel feature extraction network is obtained by training the initial channel feature extraction network. Then, N first target branch adaptive layers are trained by training the N initial branch adaptive layers. Finally, the initial sparse gating module is trained based on the trained target encoding network, target decoding network, first target channel feature extraction network, and N first target branch adaptive layers to obtain a target sparse gating module. The target model obtained by this method can solve the problems of neural network training and deduction caused by dynamic changes in the channel environment.

[0122] On the other hand, this scheme adjusts the output of the branch adaptive layer by training the weights of the branch adaptive layer instead of training the target encoding network parameters, which can reduce the overhead of network training.

[0123] Optionally, the first channel dictionary set includes a correspondence between the first target branch adaptation layer and the channel label.

[0124] In one possible implementation, training the initial channel feature extraction network to obtain the first target channel feature extraction network and the first channel dictionary set includes: training the initial channel feature extraction network multiple times to obtain the first target channel feature extraction network and the first channel dictionary set. The initial channel feature extraction network is pre-trained based on labeled historical channel information or a preset channel model.

[0125] Among them, when the channel feature extraction network U t-1 During the t-th training, both the unlabeled channel information and the labeled historical channel information are input into the channel feature extraction network U t-1The channel distribution information corresponding to the unlabeled channel information and the labeled historical channel information is obtained by processing. t-1 Cluster centers. t-1 The predicted channel dictionary set is obtained by combining the cluster centers and the channel distribution information. Then, a first loss value is calculated based on the predicted channel dictionary set and the marked channel dictionary set, and the channel feature extraction network U is adjusted based on the first loss value. t-1 Parameters. Let t = t + 1, and repeat the above steps until the number of iterations reaches the preset number, and the channel feature extraction network U t-1 As the first target channel feature extraction network, the predicted channel dictionary set is used as the first channel dictionary set.

[0126] In one possible implementation, N initial branch adaptive layers are trained based on the target encoding network and the target decoding network to obtain the N first target branch adaptive layers, including: during the t-th training of the m-th initial branch adaptive layer, first sample data is input into the target encoding network and a preset channel for processing to obtain second sample data. The second sample data is input into the m-th initial branch adaptive layer for processing to obtain data processed by the m-th initial branch adaptive layer, where m is a positive integer and m is not greater than N. The data processed by the m-th initial branch adaptive layer is input into the target decoding network for processing to obtain third sample data. Then, a second loss value is calculated based on the first and third sample data, and parameters of the m-th initial branch adaptive layer are adjusted based on the second loss value. The above steps are repeated for t = t + 1 until a stopping condition is met, and the m-th initial branch adaptive layer is used as the m-th first target branch adaptive layer.

[0127] Let m=m+1, and repeat the above steps until m=N, to obtain the N first target branch adaptive layers.

[0128] In a possible implementation, the method further includes: sending second indication information, where the second indication information instructs the mth initial branch adaptive layer to perform training.

[0129] In one possible implementation, the initial sparse gating module is trained according to the target encoding network, the target decoding network, the first target channel feature extraction network and the N first target branch adaptive layers to obtain the target sparse gating module, including: training the initial sparse gating module multiple times to obtain the target sparse gating module.

[0130] Among them, when the sparse gating module X t-1During the t-th training, the fourth sample data is input into the target coding network and the preset channel for processing to obtain the fifth sample data. The fifth sample data is input into the first target channel feature extraction network for processing to obtain channel distribution information training data and channel category information training data. The channel distribution information training data and the channel category information training data are both input into the sparse gating module X. t-1 , and obtain R first target branch adaptive layer samples and R weight samples of the first target branch adaptive layer corresponding to the channel distribution information training data, where R is a positive integer and R is not greater than N. The fifth sample data is processed according to the R first target branch adaptive layer samples and the R weight samples of the first target branch adaptive layer to obtain processed training data. The processed training data is input into the target decoding network for processing to obtain sixth sample data. Then, a third loss value is calculated based on the fourth sample data, the sixth sample data, and the R weight samples of the first target branch adaptive layer, and the sparse gating module X is adjusted according to the third loss value. t-1 Let t = t + 1, and repeat the above steps until the stopping condition is reached. t-1 As the target sparse gating module.

[0131] In a ninth aspect, the present application provides a communication device, comprising: a communication module for receiving a first reference signal; a processing module for inputting the first reference signal into a first target model for processing to obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, K is not greater than N, and K and N are both positive integers; the communication module is also used to send first information, wherein the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the K first target branch adaptive layers and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent.

[0132] In one possible implementation, the communication module is further used to: receive first data, where the first data is obtained by the transmitting end processing the encoded data to be sent; the processing module is further used to input the first data into the K first target branch adaptive layers for processing to obtain processed data.

[0133] In one possible implementation, the communication module is further used to: send a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine the correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0134] In another possible implementation, the communication module is further used to: receive a second channel dictionary set; the processing module is further used to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; the communication module is further used to send first indication information, and the first indication information is used to indicate the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0135] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0136] In one possible implementation, the communication module is also used to: receive a second reference signal; send third information, which is the same as the first information, or the third information indicates that the transmitting end sends the reference signal after a first time interval, and the third information is obtained based on the second reference signal; and receive data sent by the transmitting end within the first time.

[0137] In a tenth aspect, the present application provides a communication device, comprising:

[0138] A communication module, configured to receive a first reference signal;

[0139] a processing module, configured to input the first reference signal into a first target model for processing, and obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, where the first target model includes N first target branch adaptive layers, K is not greater than N, and both K and N are positive integers;

[0140] The communication module is further used to send fourth information, where the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers, where the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0141] In a possible implementation, the communication module is further configured to:

[0142] receiving first data, where the first data is obtained by the transmitting end processing the coded data to be sent;

[0143] The processing module is further configured to input the first data into the K first target branch adaptive layers for processing to obtain processed data.

[0144] In a possible implementation, the communication module is further configured to:

[0145] Sending a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer;

[0146] First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0147] In a possible implementation, the communication module is further configured to:

[0148] receiving a second channel dictionary set;

[0149] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptive layer in the second channel dictionary set and the first target branch adaptive layer in the first channel dictionary set.

[0150] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0151] In a possible implementation, the communication module is further configured to:

[0152] receiving a second reference signal;

[0153] sending fifth information, where the fifth information is the same as the fourth information, or the fifth information instructs the transmitting end to send the reference signal after a first time interval, and the fifth information is obtained based on the second reference signal;

[0154] Receive data sent by the sending end within the first time.

[0155] In the eleventh aspect, the present application provides a communication device, including: a communication module for sending a first reference signal; the communication module is also used to receive first information, the first information indicating K first target branch adaptive layers and the weights of the K first target branch adaptive layers in the first target model of the first reference signal receiving end, the first information is obtained based on the first reference signal, and K is a positive integer.

[0156] In one possible implementation, the device also includes a processing module, which is used to: determine, based on the first information, K second target branch adaptive layers corresponding to the K first target branch adaptive layers and the weights of the K second target branch adaptive layers; the processing module is also used to input the encoded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data; and the communication module is also used to send the first data.

[0157] In one possible implementation, when M is less than K, the processing module is also used to update the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

[0158] In one possible implementation, the communication module is further used to: receive a first channel dictionary set; the processing module is further used to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0159] In another possible implementation, the communication module is further used to: send a second channel dictionary set; receive first indication information, where the first indication information is used to indicate the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0160] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0161] In one possible implementation, the communication module is also used to: send a second reference signal; receive fifth information, which is the same as the fourth information, or the fifth information indicates that the reference signal is sent after a first time interval, and the fifth information is obtained based on the second reference signal; and send data within the first time.

[0162] In a twelfth aspect, the present application provides a communication device, including:

[0163] A communication module, configured to send a first reference signal;

[0164] The communication module is further configured to receive fourth information, where the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers. The fourth information is obtained based on the first reference signal, where K is a positive integer.

[0165] In a possible implementation, the apparatus further includes a processing module configured to:

[0166] Inputting the to-be-sent coded data into the K second target branch adaptive layers of the second target model for processing according to the fourth information to obtain first data;

[0167] The communication module is further configured to send the first data.

[0168] In a possible implementation, the communication module is further configured to:

[0169] receiving a first channel dictionary set;

[0170] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptive layer in the second channel dictionary set and the first target branch adaptive layer in the first channel dictionary set;

[0171] The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0172] In a possible implementation, the communication module is further configured to:

[0173] A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

[0174] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0175] In a possible implementation, the communication module is further configured to:

[0176] sending a second reference signal;

[0177] receiving fifth information, where the fifth information is the same as the fourth information, or the fifth information indicates that the reference signal is sent after a first time interval, and the fifth information is obtained based on the second reference signal;

[0178] The data is sent within the first time.

[0179] In a thirteenth aspect, the present application provides a communication device, including: a communication module for receiving a first reference signal; a processing module for inputting the first reference signal into a first target model for processing, and obtaining K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, K is not greater than N, and K and N are both positive integers; the communication module is also used to receive first data; the processing module is also used to input the first data into the K first target branch adaptive layers for processing, and obtain processed data, wherein the maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value.

[0180] In one possible implementation, the first target model also includes a first target channel feature extraction network and a target sparse gating module. The first target channel feature extraction network is used to process the first reference signal to obtain channel distribution information and channel category information corresponding to the channel distribution information; the target sparse gating module is used to calculate the K first target branch adaptive layers and the weights of the K first target branch adaptive layers based on the channel distribution information and the channel category information.

[0181] In one possible implementation, the K first target branch adaptive layers are used to process the first data respectively to obtain the data processed by the K first target branch adaptive layers respectively; the weights of the K first target branch adaptive layers are used to perform weighted sum processing on the data processed by the K first target branch adaptive layers respectively to obtain the processed data.

[0182] In a possible implementation, the processing module is further configured to: when a maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, decode the first data to obtain processed data.

[0183] In one possible implementation, the communication module is further used to: send first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data.

[0184] In a possible implementation, the communication module is further configured to: when a maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, send second information instructing the transmitter to directly send the first data.

[0185] In a possible implementation, the communication module is further used to: send a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine the correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0186] In another possible implementation, the communication module is further used to: receive a second channel dictionary set; the processing module is further used to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; the communication module is further used to send first indication information, and the first indication information is used to indicate the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0187] In a possible implementation, the communication module is further configured to:

[0188] Send fourth information, where the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers, where the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0189] In a possible implementation, the communication module is further configured to:

[0190] Sending a first channel dictionary set;

[0191] First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0192] In another possible implementation, the communication module is further configured to:

[0193] receiving a second channel dictionary set;

[0194] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptive layer in the second channel dictionary set and the first target branch adaptive layer in the first channel dictionary set.

[0195] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0196] In one possible implementation, the communication module is also used to: receive a second reference signal; send third information, which is the same as the first information, or the third information indicates that the sending end sends the reference signal after a first time interval, and the third information is obtained based on the second reference signal; and receive data sent by the sending end within the first time.

[0197] In another possible implementation, the communication module is also used to: receive a second reference signal; send fifth information, which is the same as the fourth information, or the fifth information indicates that the sending end sends the reference signal after a first time interval, and the fifth information is obtained based on the second reference signal; and receive data sent by the sending end within the first time.

[0198] In a fourteenth aspect, the present application provides a communication device, including: a communication module, configured to send a first reference signal; a processing module, configured to, when receiving first information, indicate K first target branch adaptive layers and weights of the K first target branch adaptive layers in a first target model of a receiving end of the first reference signal, and determine, based on the first information, K second target branch adaptive layers corresponding to the K first target branch adaptive layers and weights of the K second target branch adaptive layers, wherein the first information is obtained based on the first reference signal, and K is a positive integer; the processing module is further configured to input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data; and the communication module is further configured to send the first data.

[0199] In one possible implementation, when M is less than K, the processing module is also used to update the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

[0200] In one possible implementation, the K second target branch adaptive layers are used to respectively process the coded data to be sent to obtain the data respectively processed by the K second target branch adaptive layers; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data respectively processed by the K second target branch adaptive layers to obtain the first data.

[0201] In a possible implementation, the communication module is further configured to: when receiving second information, the second information indicating that the coded data to be sent is to be sent directly, send the coded data to be sent, and the second information is obtained based on the first reference signal.

[0202] In one possible implementation, the communication module is further used to: receive a first channel dictionary set; the processing module is further used to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0203] In another possible implementation, the communication module is further used to: send a second channel dictionary set; receive first indication information, where the first indication information is used to indicate the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0204] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0205] In one possible implementation, the communication module is also used to: send a second reference signal; receive third information, which is the same as the first information; or, the third information is used to indicate that the reference signal is sent after a first time interval, and the third information is obtained based on the second reference signal; and send data within the first time.

[0206] In a fifteenth aspect, the present application provides a communication device, including:

[0207] A communication module, configured to send a first reference signal;

[0208] a processing module, configured to, upon receiving fourth information indicating K second target branch adaptive layers and weights of the K second target branch adaptive layers, input, based on the fourth information, coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data, wherein the fourth information is obtained based on the first reference signal, and K is a positive integer;

[0209] The communication module is further configured to send the first data.

[0210] In one possible implementation, the K second target branch adaptive layers are used to respectively process the coded data to be sent to obtain the data respectively processed by the K second target branch adaptive layers; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data respectively processed by the K second target branch adaptive layers to obtain the first data.

[0211] In a possible implementation, the communication module is further configured to:

[0212] When second information is received, the second information indicates that the coded data to be sent is to be sent directly, and the coded data to be sent is sent, and the second information is obtained according to the first reference signal.

[0213] In a possible implementation, the communication module is further configured to:

[0214] receiving a first channel dictionary set;

[0215] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptive layer in the second channel dictionary set and the first target branch adaptive layer in the first channel dictionary set;

[0216] The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0217] In another possible implementation, the communication module is further configured to:

[0218] A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

[0219] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0220] In a possible implementation, the processing module is further configured to:

[0221] sending a second reference signal;

[0222] receiving fifth information, where the fifth information is the same as the fourth information; or, the fifth information is used to indicate that the reference signal is to be sent after a first time interval, and the fifth information is obtained based on the second reference signal;

[0223] The data is sent within the first time.

[0224] In the sixteenth aspect, an embodiment of the present application provides a target model training device. The target model includes a target encoding network, a target decoding network and a first target model, and the first target model includes a first target channel feature extraction network, N first target branch adaptive layers and a target sparse gating module. The device includes: a first training module for training the initial channel feature extraction network to obtain the first target channel feature extraction network and the first channel dictionary set; a second training module for training the N initial branch adaptive layers according to the target encoding network and the target decoding network to obtain the N first target branch adaptive layers, where N is the number of first target branch adaptive layers in the first channel dictionary set; a third training module for training the initial sparse gating module according to the target encoding network, the target decoding network, the first target channel feature extraction network and the N first target branch adaptive layers to obtain the target sparse gating module.

[0225] Optionally, the first channel dictionary set includes a correspondence between the first target branch adaptation layer and the channel label.

[0226] In one possible implementation, the first training module is configured to train the initial channel feature extraction network multiple times to obtain the first target channel feature extraction network and the first channel dictionary set. The initial channel feature extraction network is pre-trained based on labeled historical channel information or a preset channel model.

[0227] Among them, when the channel feature extraction network U t-1 During the t-th training, both the unlabeled channel information and the labeled historical channel information are input into the channel feature extraction network U t-1 The channel distribution information corresponding to the unlabeled channel information and the labeled historical channel information is obtained by processing. t-1 Cluster centers. t-1 The predicted channel dictionary set is obtained by combining the cluster centers and the channel distribution information. Then, a first loss value is calculated based on the predicted channel dictionary set and the marked channel dictionary set, and the channel feature extraction network U is adjusted based on the first loss value. t-1 Parameters. Let t = t + 1, and repeat the above steps until the number of iterations reaches the preset number, and the channel feature extraction network U t-1 As the first target channel feature extraction network, the predicted channel dictionary set is used as the first channel dictionary set.

[0228] In one possible implementation, the second training module is configured to: during the t-th training of the m-th initial branch adaptive layer, input the first sample data into the target encoding network and a preset channel for processing to obtain second sample data. Input the second sample data into the m-th initial branch adaptive layer for processing to obtain data processed by the m-th initial branch adaptive layer, where m is a positive integer and m is not greater than N. Input the data processed by the m-th initial branch adaptive layer into the target decoding network for processing to obtain third sample data. Then, a second loss value is calculated based on the first and third sample data, and parameters of the m-th initial branch adaptive layer are adjusted based on the second loss value. Set t = t + 1, and repeat the above steps until a stopping condition is met, whereupon the m-th initial branch adaptive layer serves as the m-th first target branch adaptive layer.

[0229] Let m=m+1, and repeat the above steps until m=N, to obtain the N first target branch adaptive layers.

[0230] In a possible implementation, the apparatus further includes a communication module configured to send second indication information, where the second indication information instructs the mth initial branch adaptive layer to perform training.

[0231] In a possible implementation, the third training module is configured to perform multiple training on the initial sparse gating module to obtain the target sparse gating module.

[0232] Among them, when the sparse gating module X t-1 During the t-th training, the fourth sample data is input into the target coding network and the preset channel for processing to obtain the fifth sample data. The fifth sample data is input into the first target channel feature extraction network for processing to obtain channel distribution information training data and channel category information training data. The channel distribution information training data and the channel category information training data are both input into the sparse gating module X. t-1 , and obtain R first target branch adaptive layer samples and R weight samples of the first target branch adaptive layer corresponding to the channel distribution information training data, where R is a positive integer and R is not greater than N. The fifth sample data is processed according to the R first target branch adaptive layer samples and the R weight samples of the first target branch adaptive layer to obtain processed training data. The processed training data is input into the target decoding network for processing to obtain sixth sample data. Then, a third loss value is calculated based on the fourth sample data, the sixth sample data, and the R weight samples of the first target branch adaptive layer, and the sparse gating module X is adjusted according to the third loss value. t-1Let t = t + 1, and repeat the above steps until the stopping condition is reached. t-1 As the target sparse gating module.

[0233] With respect to any of the ninth to fifteenth aspects above, in one possible implementation, the processing module may be a processor, and the communication module may be a transceiver module, a transceiver, or a communication interface. It is understood that the communication module may be a transceiver in the device, for example, implemented by an antenna, a feeder, and a codec in the device. Alternatively, if the communication device is a chip provided in the device, the communication module may be an input / output interface of the chip, such as an input / output circuit, a pin, etc.

[0234] For any implementation of the above-mentioned sixteenth aspect, in a possible implementation, the first training module, the second training module and the third training module can all be processors.

[0235] In one possible implementation, the communication module may be a transceiver module, a transceiver, or a communication interface. It is understood that the communication module may be a transceiver in the device, for example, implemented by an antenna, a feeder, and a codec in the device. Alternatively, if the communication device is a chip provided in the device, the communication module may be an input / output interface of the chip, such as an input / output circuit, a pin, etc.

[0236] In aspect 17, an embodiment of the present application provides a communication device, comprising one or more processors; wherein the one or more processors are used to execute computer programs stored in one or more memories, so that the communication device implements the method as described in any one of the first aspect, or implements the method as described in any one of the second aspect, or implements the method as described in any one of the third aspect, or implements the method as described in any one of the fourth aspect, or implements the method as described in any one of the fifth aspect, or implements the method as described in any one of the sixth aspect, or implements the method as described in any one of the seventh aspect.

[0237] In a possible implementation manner, the communication device further includes the one or more memories.

[0238] In a possible implementation, the communication device is a chip or a chip system.

[0239] In aspect 18, an embodiment of the present application provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions. When the instructions are executed by a processor, the method described in any one of the first aspects is implemented, or the method described in any one of the second aspects is implemented, or the method described in any one of the third aspects is implemented, or the method described in any one of the fourth aspects is implemented, or the method described in any one of the fifth aspects is implemented, or the method described in any one of the sixth aspects is implemented, or the method described in any one of the seventh aspects is implemented.

[0240] In aspect 19, an embodiment of the present application provides a computer program product, comprising a computer program. When the computer program is executed, it implements the method described in any one of the first aspects, or implements the method described in any one of the second aspects, or implements the method described in any one of the third aspects, or implements the method described in any one of the fourth aspects, or implements the method described in any one of the fifth aspects, or implements the method described in any one of the sixth aspects, or implements the method described in any one of the seventh aspects.

[0241] In aspect 20, an embodiment of the present application provides a communication system, comprising the apparatus as described in any one of aspect 9 or the apparatus as described in any one of aspect 10, and also comprising the apparatus as described in any one of aspect 12; or, the communication system comprises the apparatus as described in any one of aspect 11, and the apparatus as described in any one of aspect 13; or, the communication system comprises the apparatus as described in any one of aspect 14, and the apparatus as described in any one of aspect 15, or the apparatus as described in any one of aspect 16.

[0242] It can be understood that the apparatus described in the ninth aspect, the apparatus described in the tenth aspect, the apparatus described in the eleventh aspect, the apparatus described in the twelfth aspect, the apparatus described in the thirteenth aspect, the apparatus described in the fourteenth aspect, the apparatus described in the fifteenth aspect, the apparatus described in the sixteenth aspect, the apparatus described in the seventeenth aspect, the computer storage medium described in the eighteenth aspect, or the computer program product described in the nineteenth aspect, and the system described in the twentieth aspect are all used to execute any of the methods provided in the first aspect, any of the methods provided in the second aspect, any of the methods provided in the third aspect, any of the methods provided in the fourth aspect, any of the methods provided in the fifth aspect, any of the methods provided in the sixth aspect, any of the methods provided in the seventh aspect, or any of the methods provided in the eighth aspect. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding methods and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0243] The following is an introduction to the drawings used in the embodiments of this application.

[0244] FIG1a is a schematic diagram of the architecture of a communication system provided by an embodiment of the present application;

[0245] FIG1b is a schematic diagram of the architecture of a communication system provided in an embodiment of the present application;

[0246] FIG1c is a schematic diagram of the architecture of a communication system provided in an embodiment of the present application;

[0247] FIG2a is a flow chart of a communication method provided in an embodiment of the present application;

[0248] FIG2 b is a schematic diagram of a target sparse gating module provided in an embodiment of the present application;

[0249] FIG3 is a schematic diagram of a communication system provided in an embodiment of the present application;

[0250] FIG4 is a flow chart of another communication method provided in an embodiment of the present application;

[0251] FIG5a is a schematic diagram of another communication system provided in an embodiment of the present application;

[0252] FIG5b is a schematic diagram of a frame structure provided in an embodiment of the present application;

[0253] FIG6 a is a flow chart of a target model training method provided in an embodiment of the present application;

[0254] FIG6b is a schematic diagram of another frame structure provided in an embodiment of the present application;

[0255] FIG7a is a schematic structural diagram of a communication device provided in an embodiment of the present application;

[0256] FIG7b is a schematic structural diagram of another communication device provided in an embodiment of the present application;

[0257] FIG8 is a schematic structural diagram of another communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0258] The following describes the embodiments of the present application in conjunction with the accompanying drawings. The terms used in the implementation methods of the embodiments of the present application are only used to explain the specific embodiments of the present application and are not intended to limit the present application.

[0259] Since existing neural network algorithms cannot solve the problem of dynamic channel changes in actual environments, in view of this, the present application provides a communication method, device and system that can cope with dynamic channel changes and achieve migration adaptation of unknown channels.

[0260] The system architecture of the embodiments of the present application will be described in detail below with reference to the accompanying drawings. The communication method of the present application can be applied to a cellular communication system. Referring to FIG. 1a , FIG. 1a is a schematic diagram of a communication system applicable to the embodiments of the present application, comprising a network device 101 and a terminal 102 .

[0261] The network device 101 can be a base station, an evolved NodeB (eNodeB), a transmission reception point (TRP), a next generation NodeB (gNB) in a fifth generation (5G) mobile communication system, a next generation base station in a sixth generation (6G) mobile communication system, a base station in a future mobile communication system, or an access node in a WiFi system, etc.; it can also be a module or unit that performs some of the functions of a base station, for example, a centralized unit (CU) or a distributed unit (DU). The network device 101 can be a macro base station, a micro base station or an indoor station, a relay node or a donor node, etc. The embodiments of the present application do not limit the specific technology and specific device form adopted by the network device 101. For ease of description, the following description uses a base station as an example of a network device.

[0262] The terminal 102 may also be referred to as a terminal device, user equipment (UE), mobile station, mobile terminal, etc. The terminal can be widely used in various scenarios, such as device-to-device (D2D), vehicle-to-everything (V2X) communication, machine-type communication (MTC), Internet of Things (IOT), virtual reality, augmented reality, industrial control, autonomous driving, telemedicine, smart grid, smart furniture, smart office, smart wearable, smart transportation, smart city, etc. The terminal can be a mobile phone, tablet computer, computer with wireless transceiver function, wearable device, vehicle, drone, helicopter, airplane, ship, robot, robotic arm, smart home device, etc. The embodiments of the present application do not limit the specific technology and specific device form adopted by the terminal.

[0263] Base stations and terminals can be fixed or mobile. They can be deployed on land, indoors or outdoors, handheld or vehicle-mounted; on water; or in the air on aircraft, balloons, and satellites. The embodiments of this application do not limit the application scenarios of base stations and terminals.

[0264] Communication between base stations and terminals, between base stations, and between terminals can be carried out through authorized spectrum, unauthorized spectrum, or both; communication can be carried out through spectrum below 6 gigahertz (GHz), spectrum above 6 GHz, or spectrum below 6 GHz and spectrum above 6 GHz. The embodiments of the present application do not limit the spectrum resources used for wireless communication.

[0265] In the embodiments of the present application, the functions of the base station may also be performed by a module (such as a chip) in the base station, or by a control subsystem that includes the base station functions. The control subsystem that includes the base station functions here may be a control center in the above-mentioned application scenarios such as smart grid, industrial control, and intelligent transportation. The functions of the terminal may also be performed by a module (such as a chip or modem) in the terminal, or by a device that includes the terminal functions.

[0266] For example, this solution can be applied to wireless communication systems such as 5G and satellite communications. A wireless communication system is usually composed of cells, each of which contains a base station (BS), which provides communication services to multiple mobile stations (MS). The base station contains a baseband unit (BBU) and a remote radio unit (RRU). The BBU and RRU can be placed in different places, for example: the RRU is remote and placed in an area with high traffic volume, and the BBU is placed in a central computer room. The BBU and RRU can also be placed in the same computer room. The BBU and RRU can also be different components under the same rack.

[0267] The wireless communication systems mentioned above include, but are not limited to, narrowband Internet of Things (NB-IoT), global system for mobile communications (GSM), enhanced data rate for GSM evolution (EDGE), wideband code division multiple access (WCDMA), code division multiple access 2000 (CDMA2000), time division-synchronization code division multiple access (TD-SCDMA), long term evolution (LTE), and three major application scenarios of the next-generation 5G mobile communication system: enhanced mobile broadband (eMBB), ultra-reliable and low latency communications (URLLC), and LTE-enhanced machine-to-machine communication (eMTC).

[0268] Please refer to Figure 1b, which is a schematic diagram of another communication system applicable to an embodiment of the present application, and the system includes a satellite base station 103 and a terminal 104.

[0269] Satellite base station 103 may be a drone, hot air balloon, low-orbit satellite, medium-orbit satellite, or high-orbit satellite. Alternatively, satellite base station 103 may refer to a non-terrestrial base station or non-terrestrial device. Satellite base station 103 may function as both a network device and a terminal device. Satellite base station 103 may not have base station functionality, or may have some or all base station functionality, and this application does not limit this.

[0270] For an introduction to the terminal 104 , please refer to the description of the terminal 102 in FIG. 1 a , which will not be repeated here.

[0271] Satellite base station 103 can provide communication services for terminal 104. Satellite base station 103 transmits downlink data to terminal 104. This data is encoded using channel coding. The channel-coded data undergoes constellation modulation and is then transmitted to terminal 104. Terminal 104 transmits uplink data to satellite base station 103. This uplink data can also be encoded using channel coding. The encoded data undergoes constellation modulation and is then transmitted to satellite base station 103.

[0272] The communication method of the present application can also be applied to inter-satellite communication systems. Referring to FIG. 1 c , FIG. 1 c is a schematic diagram of another communication system applicable to the embodiment of the present application, the system including satellites 105 and 106 .

[0273] Traditional intersatellite link communication systems can be divided into two major components: the acquisition, pointing, and tracking (APT) subsystem and the communication subsystem. The communication subsystem is responsible for intersatellite information transmission and is the core of the intersatellite communication system. The APT subsystem is responsible for acquisition, pointing, and tracking between satellites. Acquisition refers to determining the incoming signal's direction; pointing refers to adjusting the transmitted beam's aiming direction toward the receiving beam; and tracking involves continuously adjusting the alignment and acquisition throughout the communication process. To minimize channel attenuation and interference while ensuring high confidentiality and transmission rates, the APT subsystem must be adjusted in real time to adapt to changes. Existing APT systems are all optical, a disadvantage of which is the difficulty of optical alignment and the need for mechanical pointing adjustments. Existing communication subsystems are mostly optical, with some microwave-band systems also available, often utilizing a single high-gain antenna. Existing APT systems and communication subsystems are independent systems. Disadvantages include optical communication being susceptible to vibration and other factors, resulting in unstable transmission rates; and the low frequency and communication capacity of millimeter waves, which require mechanical pointing adjustments for antennas.

[0274] As shown in Figure 1c, satellites 105 and 106 each include a communication module, a transceiver antenna, an APT module, and an APT transmitter / receiver. The communication method of the present application can be applied to the communication module.

[0275] In summary, the communication method of this solution can be deployed at the network device 101 end, or the satellite base station 103 end, or the satellite 105 end, and can also be deployed at the terminal 102, the terminal 104 or the satellite 106, etc. This solution does not impose any restrictions on this.

[0276] For example, when deployed at the satellite base station 103, a multi-branch adaptive layer is used to correspond to terminal networks under different channel distributions, thereby improving performance.

[0277] It should be understood that the number of devices shown in the above figures is schematic. According to the needs of the actual scenario, there can be any number of designs, and this solution does not limit this.

[0278] The above describes the architecture of the embodiment of the present application. The following describes the method of the embodiment of the present application in detail.

[0279] Referring to Figure 2a, it is a flow chart of a communication method provided by an embodiment of the present application. Optionally, the method can be applied to the aforementioned communication system, such as the communication system shown in Figure 1a. The communication method shown in Figure 2a may include steps 201-206. It should be understood that this application is described in the order of 201-206 for the convenience of description, and is not intended to limit execution to the above order. The embodiment of the present application does not limit the order of execution, execution time, number of executions, etc. of the above one or more steps. The following description is based on the example that the execution subjects of steps 201 and 204 of the communication method are base stations, and the execution subjects of steps 202, 203, 205 and 206 are terminals. This application is also applicable to other execution subjects. Steps 201-206 are as follows:

[0280] 201. A base station sends a first reference signal;

[0281] The first reference signal may be a pilot signal. A pilot signal is an unmodulated direct sequence spread spectrum signal continuously transmitted by a base station. The pilot signal enables the terminal to obtain the timing of the forward code division multiple access channel and provides a relevant demodulation phase reference.

[0282] 202. The terminal receives the first reference signal;

[0283] The terminal receives the first reference signal from the base station.

[0284] 203. The terminal inputs the first reference signal into a first target model for processing, and obtains K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, where the first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers.

[0285] In the embodiment of the present application, each first target branch adaptive layer can be understood as corresponding to a different type of channel distribution, that is, N first target branch adaptive layers correspond to N types of channel distribution.

[0286] Of course, it can also be understood that N first target branch adaptive layers correspond to N' types of clustered channel distributions, where N is greater than N'. In other words, at least one first target branch adaptive layer may correspond to one type of clustered channel distribution. This solution does not impose strict restrictions on this. Clustering can be understood as merging channel distributions with similar or identical channel distributions into one type of channel distribution.

[0287] That is, by inputting the first reference signal into the first target model for processing, K first target branch adaptive layers among the N first target branch adaptive layers of the first target model and the weights of the K first target branch adaptive layers can be determined. In other words, the current channel distribution can be comprehensively represented by the channel distribution types corresponding to the determined K first target branch adaptive layers and their corresponding weights.

[0288] In a possible implementation, the first target model further includes a first target channel feature extraction network and a target sparse gating module.

[0289] The first target channel feature extraction network is used to process the first reference signal to obtain channel distribution information and channel category information corresponding to the channel distribution information.

[0290] The channel distribution information may be a feature vector used to characterize the current channel distribution in the feature space.

[0291] The channel category information corresponding to the channel distribution information may be a probability of predicting the channel type to which the current channel distribution belongs.

[0292] The target sparse gating module is used to calculate the K first target branch adaptive layers and the weights of the K first target branch adaptive layers according to the channel distribution information and the channel category information.

[0293] The target sparse gating module can derive the type corresponding to the current channel distribution based on the above channel distribution information and channel category information, which is comprehensively represented by K first target branch adaptive layers and their corresponding weights.

[0294] Optionally, the target sparse gating module may calculate the weights of each first target branch adaptive layer based on the following method:

[0295] α=softmax(TopK(W SG *h+p cluster ));

[0296] Where softmax is a normalized exponential function; W SG is the parameter of the target sparse gating module. h is the channel distribution information. pcluster is the channel category information corresponding to the channel distribution information. α is the weight of the branch adaptive layer. TopK represents the number of branches finally selected. For example, K = 3, which means α1, α2, ..., α N Only 3 of them have corresponding values, and the rest are 0. This ensures that the network deduction time does not increase with the increase in the number of branches.

[0297] Of course, other methods may also be used to calculate the weights of the first target branch adaptive layers, and this solution does not limit this.

[0298] In a possible implementation, the K first target branch adaptive layers are used to process the first data respectively to obtain data processed by the K first target branch adaptive layers respectively.

[0299] For example, the processing performed by each first target branch adaptive layer is to perform an affine transformation on the first data. For example, the first data is f(x)∈C 72×14×1×1 , the dimension represents the complex domain of 72 subcarriers, 14 symbols, 1 receiving antenna, and 1 transmitting antenna. Each first target branch adaptation layer corresponds to a set of parameters (W i ,b i ), W∈C 72×14×1×1 , b∈C 72×14×1×1 The affine transformation operation corresponding to each branch adaptive layer can be expressed as: i (f(x))=W i f(x)+b i i represents the index of the K first target branch adaptive layers, and i is a positive integer.

[0300] Of course, the processing performed by each first target branch adaptive layer may also be other processing, and this solution does not impose a strict limitation on this.

[0301] The weights of the K first target branch adaptive layers are used to perform weighted sum processing on the data processed respectively by the K first target branch adaptive layers to obtain the processed data.

[0302] For example, the weighted summation process can be expressed as α i Represents the weights of the K first target branch adaptive layers.

[0303] 204. The base station sends first data;

[0304] 205. The terminal receives first data;

[0305] The terminal receives first data from the base station.

[0306] 206. The terminal inputs the first data into the K first target branch adaptive layers for processing to obtain processed data, wherein a maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value.

[0307] For example, the weights of the K first target branch adaptive layers are sorted to find the largest weight value. If the largest weight value is not less than a preset value, it indicates that the obtained K first target branch adaptive layers have high reliability. Therefore, the received first data is input into the K first target branch adaptive layers for processing.

[0308] The preset value can be any positive number not greater than 1. Of course, it can also be 0, which is not limited in this solution. Optionally, the preset value is 0.8.

[0309] Optionally, the processing may be to perform an affine transformation on the input first data. For example, the first data f(x)∈C corresponding to the Orthogonal Frequency Division Multiplexing (OFDM) waveform output on the base station side 72×14×1×1 , the affine transformation operation corresponding to each branch adaptive layer can be expressed as: i (f(x))=W i f(x)+b i .

[0310] Based on the above processing performed by each first target branch adaptive layer and the weight corresponding to each first target branch adaptive layer, by performing weighted sum processing, it can be calculated: For the introduction of this part, please refer to the introduction of step 203, which will not be repeated here.

[0311] In one possible implementation, the target sparse gating module can also include multiple sparse gating submodules. As shown in Figure 2b, by grading the delay spread and speed, branch level A is trained for different delay spreads (e.g., branch A1: 30ns; branch A2: 100ns), and branch level B is trained for different speeds (e.g., branch B1: 3km / s; branch B2: 30km / s). The final weighted summation process can be expressed as: t Bi , t Ai Respectively represent the affine transformation operations corresponding to each branch adaptive layer. The value range of i is an integer from 1 to N, α Ai is the weight of each branch adaptive layer of branch level A, β Bi is the weight of each branch adaptive layer at branch level B.

[0312] In other words, multiple sparse gating submodules are combined to achieve a multi-level dynamic characterization of the channel distribution. This allows for a nonlinear channel distribution characterization through the combination of multiple parameter spaces, improving generalization and transfer capabilities.

[0313] The processed data may be data obtained by decoding the data T. For example, the decoded data may be obtained by inputting the data T into a target decoding network for decoding.

[0314] As shown in Figure 3, a schematic diagram of a communication system provided by an embodiment of the present application is shown. The first target model includes a first target channel feature extraction network, a target sparse gating module, and N first target branch adaptive layers. The terminal inputs the received first reference signal into the first target channel feature extraction network for processing to obtain channel distribution information and channel category information. Then, the channel distribution information and channel category information are input into the target sparse gating module for processing to obtain K first target branch adaptive layers among the N first target branch adaptive layers and the weights of the K first target branch adaptive layers. When the maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value, the terminal inputs the received first data into the K first target branch adaptive layers for processing, performs a weighted sum calculation, and then inputs the obtained data T into the target decoding network for decoding processing, i.e., obtains the above-mentioned processed data.

[0315] In a possible implementation, when the maximum weight among the weights of the K first target branch adaptive layers is less than a preset value, the terminal directly decodes the first data to obtain processed data.

[0316] When the maximum weight value is less than a preset value, it indicates that the reliability of the K first target branch adaptive layers is low. Therefore, in the fallback operation shown in FIG3 , the terminal directly decodes the received first data without processing it through the K first target branch adaptive layers.

[0317] Optionally, the current channel data can be saved and used as training data to retrain the first target model to obtain a model with better performance. The training of the first target model can be found in the subsequent model training section and will not be repeated here.

[0318] It should be noted that this embodiment is described by taking the base station as the transmitting end and the terminal as the receiving end as an example. It can also be the case that the terminal is the transmitting end and the base station is the receiving end. The processing process is the same as above and will not be repeated here.

[0319] In an embodiment of the present application, a terminal determines K first target branch adaptive layers and weights for these K first target branch adaptive layers out of N first target branch adaptive layers of a first target model based on a received reference signal. When the maximum weight among the K first target branch adaptive layers is no less than a preset value, the received first data is input into these K first target branch adaptive layers for processing, thereby obtaining processed data. This approach characterizes the actual channel distribution based on the determined K first target branch adaptive layers and K weights to address dynamic channel changes, thereby improving system performance.

[0320] Referring to Figure 4, it is a flow chart of another communication method provided by an embodiment of the present application. Optionally, the method can be applied to the aforementioned communication system, such as the communication system shown in Figure 1a. The communication method shown in Figure 4 may include steps 401-409. It should be understood that this application is described in the order of 401-409 for the convenience of description, and is not intended to limit execution to the above order. The embodiment of the present application does not limit the order of execution, execution time, number of executions, etc. of the above one or more steps. The following description is based on the example that the execution subject of steps 401, 405, 406 and 407 of the communication method is a base station, and the execution subject of steps 402-404, 408 and 409 is a terminal. This application is also applicable to other execution subjects. Steps 401-409 are as follows:

[0321] 401. A base station sends a first reference signal.

[0322] For the introduction of this step, please refer to the description of step 201 in the embodiment shown in FIG. 2 a , which will not be repeated here.

[0323] 402. The terminal receives a first reference signal.

[0324] For the introduction of this step, please refer to the description of step 202 in the embodiment shown in FIG. 2 a , which will not be repeated here.

[0325] 403. The terminal inputs the first reference signal into a first target model for processing, and obtains K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, where the first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers.

[0326] For the introduction of this step, please refer to the description of step 203 in the embodiment shown in FIG. 2 a , which will not be repeated here.

[0327] 404. When a maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value, the terminal sends first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers. The weights of the K first target branch adaptive layers and the K first target branch adaptive layers are used by the base station to process the coded data to be sent to obtain first data.

[0328] Specifically, when the maximum weight among the K weights of the first target branch adaptive layer is no less than a preset value, the terminal sends a message to the base station, instructing it to process the coded data before sending it. This allows the base station to process the coded data in the optimal manner for the corresponding channel distribution, improving communication performance.

[0329] It should be noted that the present embodiment describes the example of a terminal sending the first information only when the maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value. Alternatively, the terminal may send the first information under any conditions. For example, if the preset value is 0, the terminal will send the first information. This solution does not impose strict limitations on this.

[0330] 405. The base station receives the first information and determines, based on the first information, K second target branch adaptation layers corresponding to the K first target branch adaptation layers and weights of the K second target branch adaptation layers.

[0331] Optionally, a second target model is deployed on the base station side. Figure 5a shows a schematic diagram of another communication system provided in an embodiment of the present application. The second target model includes M second target branch adaptive layers. The first target model includes the aforementioned first target channel feature extraction network, a target sparse gating module, and N first target branch adaptive layers.

[0332] The second target model includes the K second target branch adaptive layers, which correspond to the K first target branch adaptive layers.

[0333] That is, the terminal side characterizes the current channel distribution based on the aforementioned K first target branch adaptive layers. The terminal side feeds back the result to the base station side so that the base station side can determine K second target branch adaptive layers that characterize the current channel distribution in order to process the coded data to be transmitted.

[0334] The following describes a manner in which the base station side determines the K second target branch adaptation layers corresponding to the K first target branch adaptation layers.

[0335] Example 1: The terminal sends a first channel dictionary set to the base station.

[0336] The base station receives the first channel dictionary set, and performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0337] In this way, the base station can determine K second target branch adaptation layers corresponding to the K first target branch adaptation layers based on the corresponding relationship, and further determine the weights of the K second target branch adaptation layers based on the weights of the K first target branch adaptation layers.

[0338] Example 2: The base station sends a second channel dictionary set to the terminal.

[0339] The terminal receives the second channel dictionary set, and performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0340] Then, the terminal sends first indication information to the base station, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0341] Based on the received first indication information, the base station may determine the K second target branch adaptation layers corresponding to the K first target branch adaptation layers and the weights of the K second target branch adaptation layers.

[0342] Based on Example 1 and Example 2, optionally, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label, and the second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

[0343] That is, the correspondence between the branch adaptation layers can be determined based on the same channel label.

[0344] Of course, other methods may also be used to determine the K second target branch adaptation layers corresponding to the K first target branch adaptation layers, and this solution does not impose a strict limitation on this.

[0345] Optionally, the terminal may obtain a first channel dictionary set based on a first target channel feature extraction network in the first target model. The second target model also includes a second target channel feature extraction network. Accordingly, the base station may obtain a second channel dictionary set based on a second target channel feature extraction network in the second target model.

[0346] The above dictionary alignment operation is performed because the channel feature extraction network is trained only on the channel training set. Both the transmitting and receiving ends can train this module independently. Because the channel training sets on both ends may not be exactly the same, the resulting channel dictionary sets may differ. Therefore, dictionary alignment is used to achieve matching of the adaptive branch layers on each end.

[0347] For example, using labeled historical channel information or sampled data {H, Y label} as the channel data set for network training. H represents the sampling of the channel model, Y label Represents the label of the channel model. By inputting channel data H, the classification result is output: the probability that the channel data belongs to different channel label results according to With Y label Calculate cross entropy as the loss function L for training classify , and then update the network. Among them, the loss function L classify It can be expressed as:

[0348]

[0349] N is the total number of different channel labels involved in training.

[0350] The two ends train the channel feature extraction network based on the loss function to obtain different channel dictionary sets. For example, for the terminal, there is D UE ={(1:TDL-A30v3,TDL-C30v3),(2:TDL-A300v100)}. For the base station, there is D BS={(a:CDL-A30v100),(b:TDL-A30v3,TDL-C30v3),(c:TDL-A300v100,TDL-C300v100)}. TDL-A30v3, TDL-C30v3, and CDL-A30v100 are all channel labels. (1:TDL-A30v3,TDL-C30v3) indicates that the channel category (i.e., the branch adaptation layer) corresponding to the channel labeled TDL-A30v3 and TDL-C30v3 is 1. (a:CDL-A30v100) indicates that the channel category (i.e., the branch adaptation layer) corresponding to the channel labeled CDL-A30v100 is a. a, b, c, and the aforementioned 1 and 2 represent different channel category labels. For example, TDL-A30v3 represents a TDL-A channel model with a delay spread of 30 ns and a speed of 3 km / s.

[0351] For example, the terminal sends its own channel dictionary set D UE If the channel labels in the channel dictionary do not belong to the TDL or CDL channel model in the protocol (for example, a model built by custom historical channel information), the terminal needs to send additional channel data h closest to the cluster center. p , so that the base station can redefine the channel distribution information.

[0352] The base station receives the channel dictionary set D UE , and the channel dictionary set D obtained by itself BS Match and get the channel matching set D m For example, for the above example, the channel matching set D m ={(b:1),(c:2)}. That is, the channel adaptation layer 1 on the terminal side corresponds to the channel adaptation layer b on the base station side. The channel adaptation layer 2 on the terminal side corresponds to the channel adaptation layer c on the base station side.

[0353] For the additional channel data h p The base station can obtain channel distribution information by inputting its own channel feature extraction network, thereby determining the corresponding channel label. If there are some mismatches, it means that the multi-branch adaptive layers at both ends of the channel are not fully matched, and the model needs to be retrained to obtain a matching channel dictionary set.

[0354] The base station can also match the channel set D m Feedback is sent to the terminal for use in simulation.

[0355] Among them, for the above-mentioned matching failure situation, for example, when M is less than K, the base station updates the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, and the third target model includes M second target branch adaptive layers.

[0356] That is, when only the M second target branch adaptive layers in the third target model are deployed on the base station side, it cannot completely include the branch adaptive layers corresponding to the K first target branch adaptive layers. Therefore, by retraining and updating the third target model, a second target model is obtained. This second target model can meet the condition of including the K second target branch adaptive layers.

[0357] In this way, the base station can determine K second target branch adaptation layers corresponding to the K first target branch adaptation layers.

[0358] In a possible implementation, the K second target branch adaptive layers are used to respectively process the coded data to be sent, to obtain data processed by the K second target branch adaptive layers.

[0359] For example, the processing performed by each second target branch adaptive layer is to perform an affine transformation on the coded data to be sent. For example, the coded data to be sent is f(x)∈C 72×14×1×1 , the dimension represents the complex domain of 72 subcarriers, 14 symbols, 1 receiving antenna, and 1 transmitting antenna. Each second target branch adaptation layer corresponds to a set of parameters (W j ,b j ), W∈C 72×14×1×1 , b∈C 72×14×1×1 The affine transformation operation corresponding to each branch adaptive layer can be expressed as: j (f(x))=W j f(x)+b j . j represents the label of the K second target branch adaptive layers, and j is a positive integer.

[0360] Of course, the processing performed by each second target branch adaptive layer may also be other processing, and this solution does not impose a strict limitation on this.

[0361] The weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed respectively by the K second target branch adaptive layers to obtain first data.

[0362] For example, the weighted summation process can be expressed as α j Represents the weights of the K second target branch adaptive layers.

[0363] 406. The base station inputs the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data.

[0364] For an introduction to this step, please refer to the description of step 405 and will not be repeated here.

[0365] 407. The base station sends first data.

[0366] The base station sends the data processed by the K second target branch adaptive layers to the terminal.

[0367] 408. The terminal receives first data;

[0368] 409. The terminal inputs the first data into the K first target branch adaptive layers for processing to obtain processed data.

[0369] For the introduction of this step, please refer to the description of step 206 in the embodiment shown in FIG. 2 a , which will not be repeated here.

[0370] In a possible implementation, when the maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, the terminal sends second information, where the second information instructs the base station to directly send the first data.

[0371] That is, the base station does not need to process the coded data (first data) to be transmitted based on the second target branch adaptive layer. Accordingly, the terminal directly decodes the received first data without going through the first target branch adaptive layer for processing.

[0372] In a possible implementation manner, the base station further sends a second reference signal.

[0373] The terminal receives the second reference signal. Then, the terminal sends third information based on the second reference signal. The third information is the same as the first information. Alternatively, the third information instructs the base station to send the reference signal after a first time interval.

[0374] In other words, the terminal determines, based on the second reference signal, that the current channel distribution is consistent with the channel distribution corresponding to the first reference signal, i.e., the channel distribution remains relatively unchanged during the period from the first reference signal to the second reference signal. The terminal then instructs the base station to retransmit the reference signal after a certain period of time.

[0375] Optionally, if the channel distribution information of the current subframe is unchanged, as shown in FIG5b , the transmission of the reference signal in the next time slot (time slot #1) is reduced. For example, there is no reference signal in time slot #1.

[0376] After receiving the third information, the base station sends data within the first time.

[0377] The sending of data within the first time may be understood as the base station no longer sending a reference signal within the first time.

[0378] Correspondingly, the terminal receives the data sent within the first time. That is, the terminal no longer receives the reference signal within the first time.

[0379] In this way, the pilot transmission overhead can be reduced and the spectrum efficiency of data transmission can be increased.

[0380] In another possible implementation, the base station transmits a second reference signal. The terminal receives the second reference signal. Then, based on the second reference signal, the terminal transmits third information. The third information differs from the first information. In other words, the current channel distribution has changed. Therefore, the terminal needs to feed back the newly determined first target branch adaptive layers and their weights to the base station.

[0381] Optionally, when the channel distribution information changes, for example, the weight changes, as shown in FIG5 b , it is sent in the frame structure of time slot #0.

[0382] When the fallback operation is triggered (the maximum weight is less than the preset value), it falls back to the data transmission frame structure in the original 3GPP standard.

[0383] In a possible implementation, the feedback dimension of the information in the embodiment of the present application is related to the number N of branches in the first target branch adaptation layer, the preset value Γ, and the TopK value of the target sparse gating module.

[0384] For example:

[0385] (1) Number of bits of the branch label of the first target branch adaptation layer in, It represents rounding up log2N, where N is the total number of adaptive layers of the first target branch.

[0386] (2) Number of quantization bits The quantization loss is set to Δ. By default, the maximum weight value is sent first.

[0387] (3) Set the TopK of the target sparse gating module to K, which means selecting the K weights with the largest weight values. For weight value calculation:

[0388] Set the maximum weight calculation formula to B is the conversion factor for converting the above quantized bit number into decimal;

[0389] The remaining weight calculation formula is:

[0390] Among them, since the sum of all weights is 1, the weight of the last branch is calculated from the weights of the first K-1 branches: B1 is the decimal representation of the maximum weight value. i is the decimal representation of the i-th weight value.

[0391] Among them, when the maximum weight is less than the preset value, 1 bit is set to indicate the fallback operation.

[0392] Based on the above settings, the number of feedback bits required for the feedback table can be calculated as: K*N index +(K-1)*N quan . This can reduce the feedback information overhead.

[0393] The above parameters can be configured in the channel dictionary set so that the transmitter and receiver can be synchronized.

[0394] For example, if N = 6, Γ = 0.8, Δ = 0.02, and TopK = 3, the following control channel feedback table can be designed, requiring a total of 3*3 + 3*2 = 15 bits. Each branch is represented in binary. Since the weight of branch 4 can be calculated based on branches 1 and 2, feedback is not required.

[0395] Table 1

[0396] Branch number (3 bits) Weight (3 bits) Branch 1 (001) 0.9 (100) Branch 2 (010) 0.05 (010) Branch 4 (100) 0.05 ( / )

[0397] In time division duplexing (TDD), since weights are related to channel distribution information, the transmitter can obtain channel estimation information through channel reciprocity. Similarly, the transmitter can directly estimate branch weights using a channel feature extraction network and a sparse gating module. This design allows the transmitter to provide synchronization feedback to the receiver, ensuring consistency between branches at both ends.

[0398] Simulations of systems using this communication method yielded the following results: Adding a multi-branch adaptive layer improves the performance of target encoding and decoding networks trained on mixed channels at high signal-to-noise ratios. Furthermore, for unknown channels, combining clustered channels using a target sparse gating module can improve performance.

[0399] Furthermore, this scheme focuses on expressing channel distribution categories rather than single-frame channel feedback. Experiments have shown that this scheme's adaptive channel adjustment is robust to variations in feedback delay and to the output dimensionality of channel distribution information.

[0400] It should be noted that this embodiment is described by taking the base station as the transmitting end and the terminal as the receiving end as an example. It can also be the case that the terminal is the transmitting end and the base station is the receiving end. The processing process is the same as above and will not be repeated here.

[0401] This embodiment is described by taking an example in which a reference signal receiving end sends first information, where the first information indicates K first target branch adaptive layers and weights of the K first target branch adaptive layers.

[0402] It should be noted that the reference signal receiving end may also transmit the fourth information. The fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers. The K second target branch adaptive layers and weights of the K second target branch adaptive layers are used by the transmitting end to process the coded data to be transmitted.

[0403] In this way, the sending end can directly process the encoded data to be sent according to the information.

[0404] The weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined according to the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0405] Optionally, the reference signal receiving end sends a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between N first target branch adaptation layers and M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0406] Then, the transmitting end receives the first channel dictionary set, and performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0407] The transmitting end further sends first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0408] The reference signal receiving end determines K second target branch adaptation layers corresponding to the K first target branch adaptation layers and weights of the K second target branch adaptation layers based on the received first indication information.

[0409] Alternatively, the sending end sends a second channel dictionary set.

[0410] The reference signal receiving end receives the second channel dictionary set and then performs dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0411] Then, the reference signal receiving end determines K second target branch adaptive layers corresponding to the K first target branch adaptive layers, and weights of the K second target branch adaptive layers.

[0412] Accordingly, in a possible implementation manner, the base station further sends a second reference signal.

[0413] The terminal receives the second reference signal. Then, the terminal sends fifth information based on the second reference signal. The fifth information is the same as the fourth information. Alternatively, the fifth information instructs the base station to send the reference signal after a first time interval.

[0414] In other words, the terminal determines, based on the second reference signal, that the current channel distribution is consistent with the channel distribution corresponding to the first reference signal, i.e., the channel distribution remains relatively unchanged during the period from the first reference signal to the second reference signal. The terminal then instructs the base station to retransmit the reference signal after a certain period of time.

[0415] In an embodiment of the present application, a terminal determines K first target branch adaptive layers and weights for these K first target branch adaptive layers out of N first target branch adaptive layers of a first target model based on a received reference signal. When the maximum weight among the K first target branch adaptive layers is no less than a preset value, the terminal sends information to a base station, causing the base station to input the coded data to be transmitted into the K second target branch adaptive layers for processing based on the information and transmit the processed first data. The terminal then inputs the received first data into the K first target branch adaptive layers for processing to obtain the processed data. Using this approach, the terminal characterizes the actual channel distribution based on the determined K first target branch adaptive layers and K weights to cope with dynamic channel changes, thereby improving system performance. Furthermore, the terminal feeds back the K first target branch adaptive layers and K weights corresponding to the actual channel distribution to the base station, allowing the base station to select the optimal processing method for the corresponding channel distribution according to the information, thereby improving communication performance.

[0416] The communication method in the embodiment of the present application is introduced above. The following is an introduction to the training method of the target model in the embodiment of the present application. Referring to Figure 6a, it is a flow chart of a target model training method provided by an embodiment of the present application. The training method of the target model shown in Figure 6a may include steps 601-603. It should be understood that this application is described in the order of 601-603 for the convenience of description, and is not intended to be limited to execution in the above order. The embodiment of the present application does not limit the order of execution, execution time, number of executions, etc. of the above one or more steps. Steps 601-603 are as follows:

[0417] 601. Train an initial channel feature extraction network to obtain a first target channel feature extraction network and a first channel dictionary set.

[0418] In one possible implementation, step 601 may include:

[0419] The initial channel feature extraction network is trained multiple times to obtain the first target channel feature extraction network and the first channel dictionary set. The initial channel feature extraction network is pre-trained based on labeled historical channel information or a preset channel model.

[0420] The preset model can be the TDL model or CDL channel model in the 3GPP protocol. For example, samples are taken from the training set (historical channel information with labels), input into the channel feature extraction network, and the probability of the sample belonging to each channel category is output. The network parameters are then trained by calculating the cross-entropy loss function with the label to obtain the initial channel feature extraction network.

[0421] Among them, when the channel feature extraction network U t-1 During the t-th training, both the unlabeled channel information and the labeled historical channel information are input into the channel feature extraction network U t-1 , and obtains channel distribution information corresponding to the unlabeled channel information and the labeled historical channel information.

[0422] According to the channel distribution information, N t-1 Cluster centers.

[0423] For example, the cluster centers are obtained by performing clustering processing using a K-means clustering algorithm (K-means) or a subspace K-means method.

[0424] According to the N t-1 The cluster centers and the channel distribution information are combined to obtain a predicted channel dictionary set.

[0425] For example, the Euclidean distance between each cluster center and each channel distribution information is calculated. Based on this Euclidean distance, the most likely channel type label is inferred using the existing channel labels. By merging the same channel type labels in each cluster center, a predicted channel dictionary set can be obtained.

[0426] A first loss value is calculated based on the predicted channel dictionary set and the marked channel dictionary set, and the channel feature extraction network U is adjusted based on the first loss value. t-1 Parameters;

[0427] Let t = t + 1, and repeat the above steps until the number of iterations reaches the preset number, and then the channel feature extraction network U t-1 As the first target channel feature extraction network, the predicted channel dictionary set is used as the first channel dictionary set.

[0428] Optionally, the channel feature extraction network may include an extractor network and a classifier network. The input of the extractor network is the reference information, and of course it can also be the channel estimation result H est . H est It can be the least squares (LS) channel estimation or minimum mean squared error (MMSE) channel estimation based on the reference information, or it can be the channel estimation result obtained by the receiving neural network. The output of the extractor network is the channel distribution information h, which is used to represent the feature vector of the current channel distribution in the feature space. The input of the classifier network is the channel distribution information h, and the output is the channel category information p corresponding to the channel distribution information. cluster The channel category information is represented based on a channel dictionary set.

[0429] Among them, the loss function L1 can be expressed as:

[0430] 11=λL classify +(1-λ)L cluster ;

[0431] Among them, λ∈[0,1], λ is a hyperparameter used to adjust L classify and L cluster The influence ratio of the two loss functions; L classify is the cross entropy loss function; L cluster is the clustering loss function; Where KL(P||Q) represents the (Kullback-Leibler, KL) divergence between the auxiliary target distribution P and the soft clustering distribution Q of the sample; q ij represents the probability that sample i belongs to cluster j, hi is the channel distribution information output by the extractor network for sample i, μ j is the centroid of the jth cluster, j′ is 1, 2…N, N is the number of cluster centers; p ij represents the auxiliary target distribution; is 1, 2…M, where M is the number of samples.

[0432] By training based on the above loss function, a trained first target channel feature extraction network can be obtained.

[0433] 602. Train N initial branch adaptive layers according to the target encoding network and the target decoding network to obtain N first target branch adaptive layers, where N is the number of first target branch adaptive layers in the first channel dictionary set.

[0434] In one possible implementation, step 602 may include:

[0435] When the mth initial branch adaptive layer is trained for the tth time, the first sample data is input into the target coding network and the preset channel for processing to obtain the second sample data.

[0436] The second sample data is input into the mth initial branch adaptive layer for processing to obtain data processed by the mth initial branch adaptive layer, where m is a positive integer and is not greater than N.

[0437] The data obtained by processing the m-th initial branch adaptive layer is input into the target decoding network for processing to obtain third sample data.

[0438] A second loss value is calculated based on the first sample data and the third sample data, and the mth initial branch adaptive layer parameter is adjusted based on the second loss value.

[0439] Let t=t+1, and repeat the above steps until the stopping condition is reached, and use the mth initial branch adaptive layer as the mth first target branch adaptive layer.

[0440] Let m=m+1, and repeat the above steps until m=N, to obtain the N first target branch adaptive layers.

[0441] In a possible implementation, the method further includes:

[0442] Second indication information is sent, where the second indication information instructs the m-th initial branch adaptive layer to perform training.

[0443] For example, the transmitting end sends the instruction information to the receiving end to instruct it to train the mth branch. Then, the transmitting end sends training data to the receiving end.

[0444] Alternatively, the receiving end sends the indication information to the transmitting end to inform the receiving end that the transmitting end wants to train the mth branch, so that the transmitting end can send training data to the receiving end.

[0445] Optionally, as shown in Figure 6b, training a multi-branch adaptive layer requires a large dataset to characterize the channel distribution. Pilot signals (reference signals) are transmitted across all time slots within a subframe, and then gradient information is received for training. For each branch corresponding to each channel type, the channel dictionary set number can be additionally indicated in the frame header. This designates the adaptive layer being trained in that frame.

[0446] 603. Train an initial sparse gating module according to the target encoding network, the target decoding network, the first target channel feature extraction network, and the N first target branch adaptive layers to obtain a target sparse gating module.

[0447] That is to say, the initial sparse gating module is trained based on the trained target encoding network, target decoding network, first target channel feature extraction network and N first target branch adaptive layers.

[0448] Based on the above training, a target model as shown in Figure 3 can be obtained. The target model includes the target encoding network, the target decoding network and a first target model, and the first target model includes a first target channel feature extraction network, N first target branch adaptive layers and a target sparse gating module.

[0449] In one possible implementation, step 603 may include:

[0450] The initial sparse gating module is trained multiple times to obtain the target sparse gating module.

[0451] Among them, when the sparse gating module X t-1 During the t-th training, the fourth sample data is input into the target coding network and the preset channel for processing to obtain fifth sample data;

[0452] Inputting the fifth sample data into the first target channel feature extraction network for processing to obtain channel distribution information training data and channel category information training data;

[0453] The channel distribution information training data and the channel category information training data are input into the sparse gating module X t-1 , obtaining R first target branch adaptive layer samples and R first target branch adaptive layer weight samples corresponding to the channel distribution information training data, where R is a positive integer and R is not greater than N;

[0454] Processing the fifth sample data according to the R first target branch adaptive layer samples and the R first target branch adaptive layer weight samples to obtain processed training data;

[0455] Inputting the processed training data into the target decoding network for processing to obtain sixth sample data;

[0456] A third loss value is calculated based on the fourth sample data, the sixth sample data and the weight samples of the R first target branch adaptive layers, and the sparse gating module X is adjusted based on the third loss value. t-1 Parameters;

[0457] Let t = t + 1, and repeat the above steps until the stopping condition is reached, and the sparse gating module X t-1 As the target sparse gating module.

[0458] Optionally, the training process adds entropy regularization constraints to make the judgment of channel distribution more clear. The loss function L2 can be expressed as:

[0459] L2=L llr +βEntropy(α);

[0460] Among them, L llr is the loss function for the joint training of the target encoding network and the target decoding network. α is the weight of the branch adaptive layer, Entropy(α) represents the entropy of the branch adaptive layer weight, and β is a hyperparameter that is not less than 0.

[0461] By training based on the above loss function, a trained target sparse gating module can be obtained.

[0462] In a possible implementation, the target model further includes a second target model, and the second target model includes a second target channel feature extraction network and M second target branch adaptive layers.

[0463] The training process of the second target channel feature extraction network can be referred to the record of step 601 and will not be repeated here.

[0464] The training of the M second target branch adaptive layers may be obtained by training the M initial branch adaptive layers and the aforementioned N initial branch adaptive layers together multiple times, wherein:

[0465] When the m'th initial branch adaptive layer is trained for the tth time, the seventh sample data is input into the target coding network for processing to obtain the eighth sample data.

[0466] The eighth sample data is input into the m'th initial branch adaptive layer for processing to obtain data processed by the m'th initial branch adaptive layer, where m' is a positive integer and is not greater than M.

[0467] The data processed by the m'th initial branch adaptive layer is input into the preset channel, the m'th initial branch adaptive layer, and the target decoding network for processing to obtain ninth sample data. The m'th initial branch adaptive layer in the second target model corresponds to the m'th initial branch adaptive layer in the first target model. For example, a first channel dictionary set and a second channel dictionary set can be obtained by training the aforementioned channel feature extraction network. The first channel dictionary set corresponds to the first target model, and the second channel dictionary set corresponds to the second target model. In this way, during training, the two models can align their adaptive layers and then train together.

[0468] A fourth loss value is calculated based on the seventh sample data and the ninth sample data, and the m'th initial branch adaptive layer parameter and the mth initial branch adaptive layer are adjusted based on the fourth loss value.

[0469] Let t=t+1, and repeat the above steps until the stopping condition is reached, and use the mth initial branch adaptive layer as the mth first target branch adaptive layer, and use the m'th initial branch adaptive layer as the m'th second target branch adaptive layer.

[0470] Let m'=m'+1, and repeat the above steps until m'=M, to obtain the N first target branch adaptive layers and the M second target branch adaptive layers.

[0471] Optionally, M = N. Of course, M and N can also be in other relationships, which is not limited in this solution.

[0472] In this embodiment of the present application, the first target channel feature extraction network is obtained by training the initial channel feature extraction network. Then, N first target branch adaptive layers are trained by training the N initial branch adaptive layers. Finally, the initial sparse gating module is trained based on the trained target encoding network, target decoding network, first target channel feature extraction network, and N first target branch adaptive layers to obtain a target sparse gating module. The target model obtained by this method can solve the problems of neural network training and deduction caused by dynamic changes in the channel environment.

[0473] On the other hand, this scheme adjusts the output of the branch adaptive layer by training the weights of the branch adaptive layer instead of training the target encoding network and target decoding network parameters, which can reduce the overhead of network training.

[0474] The target model of the embodiment of the present application is introduced as an example including multiple network modules. It should be noted that the multiple network modules can be independent network modules, or they can be integrated into one, or some network modules can be integrated into one, etc. This solution does not impose strict restrictions on this.

[0475] It should be noted that the embodiments of the present application are described using the example of deploying the first target model (including the first target channel feature extraction network, N first target branch adaptive layers, and a target sparse gating module) at the receiving end and the second target model (including the second target channel feature extraction network and M second target branch adaptive layers) at the transmitting end. Alternatively, the first target model may be deployed at the transmitting end and the second target model at the receiving end. The embodiments of the present application may also deploy only the first target model, etc. This solution does not impose strict limitations on this.

[0476] It should be noted that in the various embodiments of the present application, unless otherwise specified or there is a logical conflict, the terms and / or descriptions between the various embodiments are consistent and can be referenced by each other. The technical features in different embodiments can be combined to form new embodiments according to their internal logical relationships.

[0477] The above describes in detail the method of the embodiment of the present application, and the following provides the device of the embodiment of the present application. It will be understood that in the various device embodiments of the present application, the division of multiple units or modules is only a logical division based on function, and is not intended to limit the specific structure of the device. In a specific implementation, some functional modules may be subdivided into more small functional modules, and some functional modules may be combined into one functional module, but no matter whether these functional modules are subdivided or combined, the general process performed by the device is the same. For example, some devices include a receiving unit and a sending unit. In some designs, the sending unit and the receiving unit can also be integrated into a communication unit, which can implement the functions implemented by the receiving unit and the sending unit. Typically, each unit corresponds to its own program code (or program instructions), and when the program code corresponding to each of these units runs on the processor, the unit is controlled by the processing unit to execute the corresponding process to implement the corresponding function.

[0478] The embodiments of the present application also provide an apparatus for implementing any of the above methods. For example, a communication apparatus is provided that includes a module (or means) for implementing each step performed by a terminal in any of the above methods. For another example, another communication apparatus is provided that includes a module (or means) for implementing each step performed by a base station in any of the above methods.

[0479] For example, referring to FIG7a , which is a schematic diagram of the structure of a communication device provided in an embodiment of the present application, the communication device is used to implement the aforementioned communication method, such as the communication method shown in FIG2a and FIG4 .

[0480] As shown in FIG7a , the apparatus may include a communication module 701 and a processing module 702 , specifically as follows:

[0481] The communication module 701 is configured to receive a first reference signal;

[0482] A processing module 702 is configured to input the first reference signal into a first target model for processing, and obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, where the first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers.

[0483] The communication module 701 is further configured to receive first data;

[0484] The processing module 702 is further configured to input the first data into the K first target branch adaptive layers for processing to obtain processed data, wherein the maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value.

[0485] In one possible implementation, the first target model also includes a first target channel feature extraction network and a target sparse gating module, wherein the first target channel feature extraction network is used to process the first reference signal to obtain channel distribution information and channel category information corresponding to the channel distribution information; the target sparse gating module is used to calculate the K first target branch adaptive layers and the weights of the K first target branch adaptive layers based on the channel distribution information and the channel category information.

[0486] In one possible implementation, the K first target branch adaptive layers are used to process the first data respectively to obtain the data processed by the K first target branch adaptive layers respectively; the weights of the K first target branch adaptive layers are used to perform weighted sum processing on the data processed by the K first target branch adaptive layers respectively to obtain the processed data.

[0487] In a possible implementation, the processing module 702 is further configured to: when a maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, decode the first data to obtain processed data.

[0488] In one possible implementation, the communication module 701 is further used to: send first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data.

[0489] In a possible implementation, the communication module 701 is further configured to: when a maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, send second information instructing the transmitter to directly send the first data.

[0490] In one possible implementation, the communication module 701 is further used to: send a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine the correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0491] In another possible implementation, the communication module 701 is further configured to:

[0492] receiving a second channel dictionary set;

[0493] The processing module 702 is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set;

[0494] The communication module 701 is further configured to send first indication information, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0495] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0496] In a possible implementation, the communication module 701 is further configured to:

[0497] Send fourth information, where the fourth information indicates the weights of the K second target branch adaptive layers and the K second target branch adaptive layers, where the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0498] In a possible implementation, the communication module 701 is further configured to:

[0499] Sending a first channel dictionary set;

[0500] First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0501] In a possible implementation, the communication module 701 is further configured to:

[0502] receiving a second channel dictionary set;

[0503] The processing module 702 is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0504] In a possible implementation, the communication module 701 is further configured to: receive a second reference signal;

[0505] sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send the reference signal after a first time interval, and the third information is obtained based on the second reference signal;

[0506] Receive data sent by the sending end within the first time.

[0507] For the introduction of the above modules, please refer to the description of the above embodiments, which will not be repeated here.

[0508] For example, referring to FIG7b, there is shown a schematic diagram of the structure of another communication device provided in an embodiment of the present application. The communication device is used to implement the aforementioned communication method, such as the communication method shown in FIG2a and FIG4.

[0509] The communication device includes a communication module 703 and a processing module 704, specifically as follows:

[0510] Communication module 703, configured to send a first reference signal;

[0511] A processing module 704 is configured to, upon receiving first information indicating K first target branch adaptation layers and weights of the K first target branch adaptation layers in a first target model of the first reference signal receiving end, determine, based on the first information, K second target branch adaptation layers and weights of the K second target branch adaptation layers corresponding to the K first target branch adaptation layers, wherein the first information is obtained based on the first reference signal, and K is a positive integer;

[0512] The processing module 704 is further configured to input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data;

[0513] The communication module 703 is further configured to send the first data.

[0514] In one possible implementation, when M is less than K, the processing module 704 is also used to update the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

[0515] In one possible implementation, the K second target branch adaptive layers are used to process the encoded data to be sent respectively to obtain the data processed by the K second target branch adaptive layers respectively; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed by the K second target branch adaptive layers respectively to obtain the first data.

[0516] In a possible implementation, the communication module 703 is further used to: when receiving second information, the second information indicates to directly send the coded data to be sent, then send the coded data to be sent, and the second information is obtained based on the first reference signal.

[0517] In a possible implementation, the communication module 703 is further configured to: receive a first channel dictionary set;

[0518] The processing module 704 is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0519] In another possible implementation, the communication module 703 is further configured to:

[0520] Sending a second channel dictionary set;

[0521] First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0522] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0523] In a possible implementation, the communication module 703 is further configured to: send a second reference signal;

[0524] receiving third information, where the third information is the same as the first information; or, the third information is used to indicate that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal;

[0525] Data is sent within the first time.

[0526] For the introduction of the above modules, please refer to the description of the above embodiments, which will not be repeated here.

[0527] The present application also provides a communication device, comprising:

[0528] A communication module, configured to receive a first reference signal;

[0529] A processing module is used to input the first reference signal into a first target model for processing to obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, where the first target model includes N first target branch adaptive layers, K is not greater than N, and K and N are both positive integers; the communication module is also used to send first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the K first target branch adaptive layers and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent.

[0530] In one possible implementation, the communication module is further used to: receive first data, where the first data is obtained by the transmitting end processing the encoded data to be sent; the processing module is further used to input the first data into the K first target branch adaptive layers for processing to obtain processed data.

[0531] In a possible implementation, the communication module is further used to: send a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine the correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

[0532] In another possible implementation, the communication module is further configured to: receive a second channel dictionary set;

[0533] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set;

[0534] The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0535] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0536] In one possible implementation, the communication module is further used to: receive a second reference signal; send third information, where the third information is the same as the first information, or the third information indicates that the transmitter sends the reference signal after a first time interval, and the third information is obtained based on the second reference signal; and receive data sent by the transmitter within the first time.

[0537] For the introduction of the above modules, please refer to the description of the above embodiments, which will not be repeated here.

[0538] On the other hand, the present application also provides a communication device, comprising:

[0539] A communication module, configured to send a first reference signal;

[0540] The communication module is further used to receive first information, where the first information indicates K first target branch adaptation layers and weights of the K first target branch adaptation layers in the first target model of the first reference signal receiving end, the first information being obtained based on the first reference signal, and K is a positive integer.

[0541] In a possible implementation, the apparatus further includes a processing module, configured to: determine, based on the first information, K second target branch adaptive layers corresponding to the K first target branch adaptive layers and weights of the K second target branch adaptive layers;

[0542] The processing module is further configured to input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data;

[0543] The communication module is further configured to send the first data.

[0544] In one possible implementation, when M is less than K, the processing module is also used to update the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

[0545] In a possible implementation, the communication module is further configured to: receive a first channel dictionary set;

[0546] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0547] In another possible implementation, the communication module is further used to: send a second channel dictionary set; receive first indication information, where the first indication information is used to indicate the correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0548] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0549] In one possible implementation, the communication module is further used to: send a second reference signal; receive third information, where the third information is the same as the first information, or the third information indicates that the reference signal is sent after a first time interval, and the third information is obtained based on the second reference signal; and send data within the first time.

[0550] For the introduction of the above modules, please refer to the description of the above embodiments, which will not be repeated here.

[0551] The present application also provides a communication device, including:

[0552] A communication module, configured to receive a first reference signal;

[0553] a processing module, configured to input the first reference signal into a first target model for processing, and obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, K is not greater than N, and both K and N are positive integers;

[0554] The communication module is further used to send fourth information, where the fourth information indicates the weights of the K second target branch adaptive layers and the K second target branch adaptive layers, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

[0555] In a possible implementation, the communication module is further configured to:

[0556] receiving first data, where the first data is obtained by the transmitting end processing the coded data to be sent;

[0557] The processing module is further configured to input the first data into the K first target branch adaptive layers for processing to obtain processed data.

[0558] In a possible implementation, the communication module is further configured to:

[0559] Sending a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer;

[0560] First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0561] In a possible implementation, the communication module is further configured to:

[0562] receiving a second channel dictionary set;

[0563] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0564] In a possible implementation manner, the first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

[0565] In a possible implementation, the communication module is further configured to:

[0566] receiving a second reference signal;

[0567] sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send a reference signal after a first time interval, and the third information is obtained based on the second reference signal;

[0568] Receive data sent by the sending end within the first time.

[0569] The present application also provides a communication device, including:

[0570] A communication module, configured to send a first reference signal;

[0571] The communication module is further configured to receive fourth information, where the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers, and the fourth information is obtained based on the first reference signal, where K is a positive integer.

[0572] In a possible implementation, the apparatus further includes a processing module configured to:

[0573] Inputting the to-be-sent coded data into the K second target branch adaptive layers of the second target model for processing according to the fourth information to obtain first data;

[0574] The communication module is further configured to send the first data.

[0575] In a possible implementation, the communication module is further configured to:

[0576] receiving a first channel dictionary set;

[0577] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set;

[0578] The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

[0579] In a possible implementation, the communication module is further configured to:

[0580] A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

[0581] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0582] In a possible implementation, the communication module is further configured to:

[0583] sending a second reference signal;

[0584] receiving third information, where the third information is the same as the first information, or the third information indicates that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal;

[0585] Data is sent within the first time.

[0586] The present application also provides a communication device, including:

[0587] A communication module, configured to send a first reference signal;

[0588] a processing module, configured to, upon receiving fourth information indicating K second target branch adaptive layers and weights of the K second target branch adaptive layers, input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data, according to the fourth information, wherein the fourth information is obtained based on the first reference signal, and K is a positive integer;

[0589] The communication module is further configured to send the first data.

[0590] In one possible implementation, the K second target branch adaptive layers are used to process the encoded data to be sent respectively to obtain the data processed by the K second target branch adaptive layers respectively; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed by the K second target branch adaptive layers respectively to obtain the first data.

[0591] In a possible implementation, the communication module is further configured to:

[0592] When second information is received, and the second information indicates to directly send the coded data to be sent, the coded data to be sent is sent, and the second information is obtained according to the first reference signal.

[0593] In a possible implementation, the communication module is further configured to:

[0594] receiving a first channel dictionary set;

[0595] The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set;

[0596] The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

[0597] In a possible implementation, the communication module is further configured to:

[0598] A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

[0599] In a possible implementation manner, the second channel dictionary set includes a correspondence between the second target branch adaptation layer and the channel label.

[0600] In a possible implementation, the processing module is further configured to:

[0601] sending a second reference signal;

[0602] receiving third information, where the third information is the same as the first information; or, the third information is used to indicate that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal;

[0603] Data is sent within the first time.

[0604] The present application also provides a target model training device. The target model includes a target encoding network, a target decoding network, and a first target model. The first target model includes a first target channel feature extraction network, N first target branch adaptive layers, and a target sparse gating module. The device includes:

[0605] A first training module is used to train the initial channel feature extraction network to obtain the first target channel feature extraction network and the first channel dictionary set;

[0606] A second training module is configured to train the N initial branch adaptive layers according to the target encoding network and the target decoding network to obtain the N first target branch adaptive layers, where N is the number of first target branch adaptive layers in the first channel dictionary set;

[0607] The third training module is used to train the initial sparse gating module according to the target encoding network, the target decoding network, the first target channel feature extraction network and the N first target branch adaptive layers to obtain the target sparse gating module.

[0608] Optionally, the first channel dictionary set includes a correspondence between the first target branch adaptation layer and the channel label.

[0609] In one possible implementation, the first training module is configured to train the initial channel feature extraction network multiple times to obtain the first target channel feature extraction network and the first channel dictionary set. The initial channel feature extraction network is pre-trained based on labeled historical channel information or a preset channel model.

[0610] Among them, when the channel feature extraction network U t-1 During the t-th training, both the unlabeled channel information and the labeled historical channel information are input into the channel feature extraction network U t-1 The channel distribution information corresponding to the unlabeled channel information and the labeled historical channel information is obtained by processing. t-1 Cluster centers. t-1 The predicted channel dictionary set is obtained by combining the cluster centers and the channel distribution information. Then, a first loss value is calculated based on the predicted channel dictionary set and the marked channel dictionary set, and the channel feature extraction network U is adjusted based on the first loss value. t-1 Parameters. Let t = t + 1, and repeat the above steps until the number of iterations reaches the preset number, and the channel feature extraction network U t-1 As the first target channel feature extraction network, the predicted channel dictionary set is used as the first channel dictionary set.

[0611] In one possible implementation, the second training module is configured to: during the t-th training of the m-th initial branch adaptive layer, input the first sample data into the target encoding network and a preset channel for processing to obtain second sample data. Input the second sample data into the m-th initial branch adaptive layer for processing to obtain data processed by the m-th initial branch adaptive layer, where m is a positive integer and m is not greater than N. Input the data processed by the m-th initial branch adaptive layer into the target decoding network for processing to obtain third sample data. Then, a second loss value is calculated based on the first and third sample data, and parameters of the m-th initial branch adaptive layer are adjusted based on the second loss value. Set t = t + 1, and repeat the above steps until a stopping condition is met, whereupon the m-th initial branch adaptive layer serves as the m-th first target branch adaptive layer.

[0612] Let m=m+1, and repeat the above steps until m=N, to obtain the N first target branch adaptive layers.

[0613] In a possible implementation, the apparatus further includes a communication module configured to: send second indication information, where the second indication information instructs the mth initial branch adaptive layer to perform training.

[0614] In a possible implementation, the third training module is configured to perform multiple training on the initial sparse gating module to obtain the target sparse gating module.

[0615] Among them, when the sparse gating module X t-1 During the t-th training, the fourth sample data is input into the target coding network and the preset channel for processing to obtain the fifth sample data. The fifth sample data is input into the first target channel feature extraction network for processing to obtain channel distribution information training data and channel category information training data. The channel distribution information training data and the channel category information training data are both input into the sparse gating module X. t-1 , and obtain R first target branch adaptive layer samples and R weight samples of the first target branch adaptive layer corresponding to the channel distribution information training data, where R is a positive integer and R is not greater than N. The fifth sample data is processed according to the R first target branch adaptive layer samples and the R weight samples of the first target branch adaptive layer to obtain processed training data. The processed training data is input into the target decoding network for processing to obtain sixth sample data. Then, a third loss value is calculated based on the fourth sample data, the sixth sample data, and the R weight samples of the first target branch adaptive layer, and the sparse gating module X is adjusted according to the third loss value. t-1Let t = t + 1, and repeat the above steps until the stopping condition is reached. t-1 As the target sparse gating module.

[0616] For the introduction of the above modules, please refer to the description of the above embodiments, which will not be repeated here.

[0617] It should be understood that the division of the modules in the above-mentioned devices is merely a division of logical functions. In actual implementation, they may be fully or partially integrated into a single physical entity, or they may be physically separated. Furthermore, the modules in a communication device or a target model training device may be implemented in the form of a processor calling software; for example, the communication device includes a processor connected to a memory storing instructions, and the processor calls the instructions stored in the memory to implement any of the above methods or the functions of the modules of the device. The processor may be, for example, a general-purpose processor such as a central processing unit (CPU) or a microprocessor, and the memory may be a memory within the device or a memory outside the device. Alternatively, the modules in the device can be implemented in the form of hardware circuits, and the functions of some or all units can be realized by designing the hardware circuits. The hardware circuit can be understood as one or more processors. For example, in one implementation, the hardware circuit is an application-specific integrated circuit (ASIC), which realizes the functions of some or all of the above units by designing the logical relationship of the components in the circuit. For another example, in another implementation, the hardware circuit can be implemented by a programmable logic device (PLD). Taking a field programmable gate array (FPGA) as an example, it can include a large number of logic gate circuits, and the connection relationship between the logic gate circuits is configured by configuring the configuration file, thereby realizing the functions of some or all of the above units. All modules of the above devices can be implemented in the form of software called by the processor, or in the form of hardware circuits, or in part by the form of software called by the processor, and the rest by hardware circuits.

[0618] 8 is a schematic diagram of the hardware structure of another communication device provided in an embodiment of the present application. The communication device 800 shown in FIG8 (the device 800 may be a computer device) includes a memory 801, a processor 802, a communication interface 803, and a bus 804. The memory 801, the processor 802, and the communication interface 803 are connected to each other via the bus 804.

[0619] The memory 801 may be a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM).

[0620] The memory 801 can store programs. When the program stored in the memory 801 is executed by the processor 802, the processor 802 and the communication interface 803 are used to execute the various steps of the communication method or the target model training method of the embodiment of the present application.

[0621] The processor 802 is a circuit with signal processing capabilities. In one implementation, the processor 802 can be a circuit with instruction reading and execution capabilities, such as a central processing unit (CPU), a microprocessor, a graphics processing unit (GPU) (which can be understood as a microprocessor), or a digital signal processor (DSP). In another implementation, the processor 802 can implement certain functions through the logical relationship of a hardware circuit. The logical relationship of the hardware circuit is fixed or reconfigurable. For example, the processor 802 is a hardware circuit implemented by an ASIC or a programmable logic device (PLD), such as an FPGA. In a reconfigurable hardware circuit, the process of the processor loading a configuration file and implementing the hardware circuit configuration can be understood as the process of the processor loading instructions to implement the functions of some or all of the above modules. In addition, it can also be a hardware circuit designed for artificial intelligence, which can be understood as an ASIC, such as a neural network processing unit (NPU), a tensor processing unit (TPU), a deep learning processing unit (DPU), etc. The processor 802 is used to execute relevant programs to implement the functions required to be performed by the units in the communication device or target model training device of the embodiment of the present application, or to execute the communication method or target model training method of the method embodiment of the present application.

[0622] It can be seen that each module in the above device can be one or more processors (or processing circuits) configured to implement the above method, such as: CPU, GPU, NPU, TPU, DPU, microprocessor, DSP, ASIC, FPGA, or a combination of at least two of these processor forms.

[0623] In addition, the modules in the above device can be fully or partially integrated together, or can be implemented independently. In one implementation, these modules are integrated together and implemented in the form of a system-on-a-chip (SOC). The SOC may include at least one processor for implementing any of the above methods or implementing the functions of the modules of the device. The type of the at least one processor can be different, for example, including a CPU and FPGA, a CPU and an artificial intelligence processor, a CPU and a GPU, etc.

[0624] The communication interface 803 uses a transceiver device such as, but not limited to, a transceiver to implement communication between the apparatus 800 and other devices or a communication network. For example, data can be obtained through the communication interface 803 .

[0625] The bus 804 may include a path for transmitting information between various components of the device 800 (eg, the memory 801 , the processor 802 , and the communication interface 803 ).

[0626] It should be noted that although the device 800 shown in FIG8 only shows a memory, a processor, and a communication interface, during the specific implementation process, those skilled in the art will understand that the device 800 also includes other components necessary for normal operation. At the same time, according to specific needs, those skilled in the art will understand that the device 800 may also include hardware components that implement other additional functions. In addition, those skilled in the art will understand that the device 800 may also include only the components necessary to implement the embodiments of the present application, and does not necessarily include all the components shown in FIG8.

[0627] An embodiment of the present application further provides a communication system, for example, including the apparatus shown in FIG. 7 a and the apparatus shown in FIG. 7 b .

[0628] An embodiment of the present application also provides a computer-readable storage medium, which stores instructions. When the computer-readable storage medium is executed on a computer or a processor, the computer or processor executes one or more steps in any of the above methods.

[0629] The present application also provides a computer program product comprising instructions, which, when executed on a computer or processor, causes the computer or processor to execute one or more steps in any of the above methods.

[0630] It should be understood that in the description of this application, unless otherwise specified, " / " indicates that the objects associated with each other are in an "or" relationship. For example, A / B can mean A or B; where A and B can be singular or plural. Also, in the description of this application, unless otherwise specified, "multiple" means two or more than two. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b, or c can mean: a, b, c, ab, ac, bc, or abc, where a, b, and c can be single or plural. In addition, to facilitate the clear description of the technical solutions of the embodiments of this application, in the embodiments of this application, words such as "first" and "second" are used to distinguish between identical or similar items with substantially the same functions and effects. Those skilled in the art will understand that words such as "first" and "second" do not limit the quantity or execution order, and words such as "first" and "second" do not necessarily mean different. At the same time, in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or descriptions. Any embodiment or design described as "exemplary" or "for example" in the embodiments of this application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Rather, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner to facilitate understanding.

[0631] In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. The mutual coupling, direct coupling, or communication connection shown or discussed can be through some interface, indirect coupling or communication connection of devices or units, and can be electrical, mechanical or other forms.

[0632] Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, that is, they may be located in one place or distributed across multiple network units. Some or all of these units may be selected to achieve the purpose of this embodiment according to actual needs.

[0633] In the above embodiments, all or part of the embodiments may be implemented by software, hardware, firmware, or any combination thereof. When implemented using software, all or part of the embodiments may be implemented in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the processes or functions according to the embodiments of the present application are generated. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium or transmitted via the computer-readable storage medium. The computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via a wired (e.g., coaxial cable, optical fiber, digital subscriber line (DSL)) or wireless (e.g., infrared, wireless, microwave, etc.) method. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more available media integrated therein. The available medium may be a read-only memory (ROM), a random access memory (RAM), a magnetic medium such as a floppy disk, a hard disk, a tape, a magnetic disk, or an optical medium such as a digital versatile disc (DVD), or a semiconductor medium such as a solid state disk (SSD).

[0634] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed in the present invention should be included in the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

Claims

1. A communication method, characterized in that: include: receiving a first reference signal; Inputting the first reference signal into a first target model for processing, obtaining K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers; Sending first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, where the K first target branch adaptive layers and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent.

2. The method according to claim 1, characterized in that The method further comprises: receiving first data, where the first data is obtained by the transmitting end processing the coded data to be sent; The first data is input into the K first target branch adaptive layers for processing to obtain processed data.

3. The method according to claim 1 or 2, characterized in that The method further comprises: A first channel dictionary set is sent, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

4. The method according to claim 1 or 2, characterized in that The method further comprises: receiving a second channel dictionary set; Performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set; First indication information is sent, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

5. The method according to claim 3 or 4, characterized in that The first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

6. The method according to any one of claims 1 to 5, characterized in that The method further comprises: receiving a second reference signal; sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send a reference signal after a first time interval, and the third information is obtained based on the second reference signal; Receive data sent by the sending end within the first time.

7. A communication method, characterized in that: include: receiving a first reference signal; Inputting the first reference signal into a first target model for processing, obtaining K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers; Send fourth information, where the fourth information indicates the weights of the K second target branch adaptive layers and the K second target branch adaptive layers, where the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

8. The method according to claim 7, characterized in that The method further comprises: receiving first data, where the first data is obtained by the transmitting end processing the coded data to be sent; The first data is input into the K first target branch adaptive layers for processing to obtain processed data.

9. The method according to claim 7 or 8, characterized in that The method further comprises: Sending a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

10. The method according to claim 7 or 8, characterized in that The method further comprises: receiving a second channel dictionary set; Perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

11. The method according to claim 9 or 10, characterized in that The first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

12. The method according to any one of claims 7 to 11, characterized in that The method further comprises: receiving a second reference signal; sending fifth information, where the fifth information is the same as the fourth information, or the fifth information instructs the transmitting end to send a reference signal after a first time interval, and the fifth information is obtained based on the second reference signal; Receive data sent by the sending end within the first time.

13. A communication method, characterized in that: include: sending a first reference signal; Receive first information, where the first information indicates K first target branch adaptation layers and weights of the K first target branch adaptation layers in a first target model of the first reference signal receiving end, where the first information is obtained based on the first reference signal, and K is a positive integer.

14. The method according to claim 13, characterized in that The method further comprises: Determining, according to the first information, K second target branch adaptive layers corresponding to the K first target branch adaptive layers and weights of the K second target branch adaptive layers; Inputting the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data; The first data is sent.

15. The method according to claim 14, characterized in that When M is less than K, the third target model is updated to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

16. The method according to any one of claims 13 to 15, characterized in that The method further comprises: receiving a first channel dictionary set; Perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

17. The method according to any one of claims 13 to 15, characterized in that The method further comprises: Sending a second channel dictionary set; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

18. The method according to claim 16 or 17, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

19. The method according to any one of claims 13 to 18, characterized in that The method further comprises: sending a second reference signal; receiving third information, where the third information is the same as the first information, or the third information indicates that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal; Data is sent within the first time.

20. A communication method, characterized in that: include: sending a first reference signal; Fourth information is received, where the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers, the fourth information is obtained according to the first reference signal, and K is a positive integer.

21. The method according to claim 20, characterized in that The method further comprises: Inputting the to-be-sent coded data into the K second target branch adaptive layers of the second target model for processing according to the fourth information to obtain first data; The first data is sent.

22. The method according to claim 20 or 21, characterized in that The method further comprises: receiving a first channel dictionary set; Performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set; First indication information is sent, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

23. The method according to claim 20 or 21, characterized in that The method further comprises: A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

24. The method according to claim 22 or 23, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

25. The method according to any one of claims 20 to 24, characterized in that The method further comprises: sending a second reference signal; receiving fifth information, where the fifth information is the same as the fourth information, or the fifth information indicates that a reference signal is sent after a first time interval, and the fifth information is obtained based on the second reference signal; Data is sent within the first time.

26. A communication method, characterized in that: include: receiving a first reference signal; Inputting the first reference signal into a first target model for processing, obtaining K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, where K is not greater than N, and both K and N are positive integers; receiving first data; The first data is input into the K first target branch adaptive layers for processing to obtain processed data, wherein a maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value.

27. The method according to claim 26, characterized in that The first target model also includes a first target channel feature extraction network and a target sparse gating module, The first target channel feature extraction network is used to process the first reference signal to obtain channel distribution information and channel category information corresponding to the channel distribution information; The target sparse gating module is used to calculate the K first target branch adaptive layers and the weights of the K first target branch adaptive layers according to the channel distribution information and the channel category information.

28. The method according to claim 26 or 27, characterized in that The K first target branch adaptive layers are used to process the first data respectively to obtain the data processed by the K first target branch adaptive layers respectively; the weights of the K first target branch adaptive layers are used to perform weighted sum processing on the data processed by the K first target branch adaptive layers respectively to obtain the processed data.

29. The method according to any one of claims 26 to 28, characterized in that The method further comprises: When the maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, decoding processing is performed on the first data to obtain processed data.

30. The method according to any one of claims 26 to 29, characterized in that The method further comprises: Send first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data.

31. The method according to any one of claims 26 to 30, characterized in that The method further comprises: A first channel dictionary set is sent, where the first channel dictionary set is used by a transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

32. The method according to any one of claims 26 to 30, characterized in that The method further comprises: receiving a second channel dictionary set; Performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set; First indication information is sent, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

33. The method according to any one of claims 26 to 29, characterized in that The method further comprises: Send fourth information, where the fourth information indicates the weights of the K second target branch adaptive layers and the K second target branch adaptive layers, where the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

34. The method according to any one of claims 26 to 29 and 33, characterized in that The method further comprises: Sending a first channel dictionary set; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

35. The method according to any one of claims 26 to 29 and 33, characterized in that The method further comprises: receiving a second channel dictionary set; Perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

36. The method according to claim 31, 32, 34 or 35, characterized in that The first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

37. The method according to any one of claims 26 to 36, characterized in that The method further comprises: When the maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, second information is sent, where the second information instructs the sending end to directly send the first data.

38. The method according to claim 30, wherein The method further comprises: receiving a second reference signal; sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send the reference signal after a first time interval, and the third information is obtained based on the second reference signal; Receive data sent by the sending end within the first time.

39. A communication method, characterized in that: include: sending a first reference signal; upon receiving first information indicating K first target branch adaptation layers and weights of the K first target branch adaptation layers in a first target model of the first reference signal receiving end, determining, based on the first information, K second target branch adaptation layers and weights of the K second target branch adaptation layers corresponding to the K first target branch adaptation layers, wherein the first information is obtained based on the first reference signal, and K is a positive integer; Inputting the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data; The first data is sent.

40. The method according to claim 39, wherein When M is less than K, the third target model is updated to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

41. The method according to claim 39 or 40, characterized in that The K second target branch adaptive layers are used to process the encoded data to be sent respectively to obtain the data processed by the K second target branch adaptive layers respectively; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed by the K second target branch adaptive layers respectively to obtain the first data.

42. The method according to any one of claims 39 to 41, characterized in that The method further comprises: When second information is received, and the second information indicates to directly send the coded data to be sent, the coded data to be sent is sent, and the second information is obtained according to the first reference signal.

43. The method according to any one of claims 39 to 42, characterized in that The method further comprises: receiving a first channel dictionary set; Perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

44. The method according to any one of claims 39 to 42, characterized in that The method further comprises: Sending a second channel dictionary set; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

45. The method according to claim 43 or 44, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

46. ​​The method according to any one of claims 39 to 45, characterized in that The method further comprises: sending a second reference signal; receiving third information, where the third information is the same as the first information; or, the third information is used to indicate that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal; Data is sent within the first time.

47. A communication method, characterized in that: include: sending a first reference signal; When fourth information is received, the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers, and according to the fourth information, inputs the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data, where the fourth information is obtained based on the first reference signal, and K is a positive integer; The first data is sent.

48. The method according to claim 47, wherein The K second target branch adaptive layers are used to process the encoded data to be sent respectively to obtain the data processed by the K second target branch adaptive layers respectively; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed by the K second target branch adaptive layers respectively to obtain the first data.

49. The method according to claim 47 or 48, characterized in that The method further comprises: When second information is received, and the second information indicates to directly send the coded data to be sent, the coded data to be sent is sent, and the second information is obtained according to the first reference signal.

50. The method according to any one of claims 47 to 49, characterized in that The method further comprises: receiving a first channel dictionary set; Performing dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set; First indication information is sent, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

51. The method according to any one of claims 47 to 49, characterized in that The method further comprises: A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

52. The method according to claim 50 or 51, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

53. The method according to any one of claims 47 to 52, characterized in that The method further comprises: sending a second reference signal; receiving fifth information, where the fifth information is the same as the fourth information; or, the fifth information is used to indicate that the reference signal is to be sent after a first time interval, and the fifth information is obtained based on the second reference signal; Data is sent within the first time.

54. A communication device, characterized in that include: A communication module, configured to receive a first reference signal; a processing module, configured to input the first reference signal into a first target model for processing, and obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, K is not greater than N, and both K and N are positive integers; The communication module is also used to send first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent.

55. The device according to claim 54, characterized in that The communication module is further used for: receiving first data, where the first data is obtained by the transmitting end processing the coded data to be sent; The processing module is further configured to input the first data into the K first target branch adaptive layers for processing to obtain processed data.

56. The device according to claim 54 or 55, characterized in that The communication module is further used for: A first channel dictionary set is sent, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

57. The device according to claim 54 or 55, characterized in that The communication module is further used for: receiving a second channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

58. The device according to claim 56 or 57, characterized in that The first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

59. The device according to any one of claims 54 to 58, characterized in that The communication module is further used for: receiving a second reference signal; sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send a reference signal after a first time interval, and the third information is obtained based on the second reference signal; Receive data sent by the sending end within the first time.

60. A communication device, characterized in that include: A communication module, configured to receive a first reference signal; a processing module, configured to input the first reference signal into a first target model for processing, and obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, K is not greater than N, and both K and N are positive integers; The communication module is further used to send fourth information, where the fourth information indicates the weights of the K second target branch adaptive layers and the K second target branch adaptive layers, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

61. The device according to claim 60, characterized in that The communication module is further used for: receiving first data, where the first data is obtained by the transmitting end processing the coded data to be sent; The processing module is further configured to input the first data into the K first target branch adaptive layers for processing to obtain processed data.

62. The device according to claim 60 or 61, characterized in that The communication module is further used for: Sending a first channel dictionary set, where the first channel dictionary set is used by the transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

63. The device according to claim 60 or 61, characterized in that The communication module is further used for: receiving a second channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

64. The device according to claim 62 or 63, characterized in that The first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

65. The device according to any one of claims 60 to 64, characterized in that The communication module is further used for: receiving a second reference signal; sending fifth information, where the fifth information is the same as the fourth information, or the fifth information instructs the transmitting end to send a reference signal after a first time interval, and the fifth information is obtained based on the second reference signal; Receive data sent by the sending end within the first time.

66. A communication device, characterized in that include: A communication module, configured to send a first reference signal; The communication module is further used to receive first information, where the first information indicates K first target branch adaptation layers and weights of the K first target branch adaptation layers in the first target model of the first reference signal receiving end, the first information is obtained based on the first reference signal, and K is a positive integer.

67. The device according to claim 66, characterized in that The device further comprises a processing module, configured to: Determining, according to the first information, K second target branch adaptive layers corresponding to the K first target branch adaptive layers and weights of the K second target branch adaptive layers; The processing module is further configured to input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data; The communication module is further configured to send the first data.

68. The device according to claim 67, characterized in that When M is less than K, the processing module is further used to update the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

69. The device according to any one of claims 66 to 68, characterized in that The communication module is further used for: receiving a first channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

70. The device according to any one of claims 66 to 68, characterized in that The communication module is further used for: Sending a second channel dictionary set; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

71. The device according to claim 69 or 70, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

72. The device according to any one of claims 66 to 71, characterized in that The communication module is further used for: sending a second reference signal; receiving third information, where the third information is the same as the first information, or the third information indicates that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal; Data is sent within the first time.

73. A communication device, characterized in that include: A communication module, configured to send a first reference signal; The communication module is further configured to receive fourth information, where the fourth information indicates K second target branch adaptive layers and weights of the K second target branch adaptive layers, and the fourth information is obtained based on the first reference signal, where K is a positive integer.

74. The device according to claim 73, characterized in that The device further comprises a processing module, configured to: Inputting the to-be-sent coded data into the K second target branch adaptive layers of the second target model for processing according to the fourth information to obtain first data; The communication module is further configured to send the first data.

75. The device according to claim 73 or 74, characterized in that The communication module is further used for: receiving a first channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

76. The device according to claim 73 or 74, characterized in that The communication module is further used for: A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

77. The device according to claim 75 or 76, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

78. The device according to any one of claims 73 to 77, characterized in that The communication module is further used for: sending a second reference signal; receiving fifth information, where the fifth information is the same as the fourth information, or the fifth information indicates that a reference signal is sent after a first time interval, and the fifth information is obtained based on the second reference signal; Data is sent within the first time.

79. A communication device, characterized in that include: A communication module, configured to receive a first reference signal; a processing module, configured to input the first reference signal into a first target model for processing, and obtain K first target branch adaptive layers and weights of the K first target branch adaptive layers in the first target model, wherein the first target model includes N first target branch adaptive layers, K is not greater than N, and both K and N are positive integers; The communication module is further configured to receive first data; The processing module is further used to input the first data into the K first target branch adaptive layers for processing to obtain processed data, wherein the maximum weight among the weights of the K first target branch adaptive layers is not less than a preset value.

80. The device according to claim 79, characterized in that The first target model also includes a first target channel feature extraction network and a target sparse gating module, The first target channel feature extraction network is used to process the first reference signal to obtain channel distribution information and channel category information corresponding to the channel distribution information; The target sparse gating module is used to calculate the K first target branch adaptive layers and the weights of the K first target branch adaptive layers according to the channel distribution information and the channel category information.

81. The device according to claim 79 or 80, characterized in that The K first target branch adaptive layers are used to process the first data respectively to obtain the data processed by the K first target branch adaptive layers respectively; the weights of the K first target branch adaptive layers are used to perform weighted sum processing on the data processed by the K first target branch adaptive layers respectively to obtain the processed data.

82. The device according to any one of claims 79 to 81, characterized in that The processing module is further configured to: When the maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, decoding processing is performed on the first data to obtain processed data.

83. The device according to any one of claims 79 to 82, characterized in that The communication module is further used for: Send first information, where the first information indicates the K first target branch adaptive layers and the weights of the K first target branch adaptive layers, and the weights of the K first target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data.

84. The device according to any one of claims 79 to 83, characterized in that The communication module is further used for: A first channel dictionary set is sent, where the first channel dictionary set is used by a transmitting end to determine a correspondence between the N first target branch adaptation layers and the M second target branch adaptation layers of the transmitting end, where M is a positive integer.

85. The device according to any one of claims 79 to 83, characterized in that The communication module is further used for: receiving a second channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

86. The device according to any one of claims 79 to 82, characterized in that The communication module is further used for: Send fourth information, where the fourth information indicates the weights of the K second target branch adaptive layers and the K second target branch adaptive layers, where the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are used by the transmitting end to process the coded data to be sent to obtain the first data, and the weights of the K second target branch adaptive layers and the K second target branch adaptive layers are determined based on the weights of the K first target branch adaptive layers and the K first target branch adaptive layers.

87. The device according to any one of claims 79 to 82 and 86, characterized in that The communication module is further used for: Sending a first channel dictionary set; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

88. The device according to any one of claims 79 to 82 and 86, characterized in that The communication module is further used for: receiving a second channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

89. The device according to claim 84, 85, 87 or 88, characterized in that The first channel dictionary set includes a correspondence between a first target branch adaptation layer and a channel label.

90. The device according to any one of claims 79 to 89, characterized in that The communication module is further used for: When the maximum weight among the weights of the K first target branch adaptive layers is less than the preset value, second information is sent, where the second information instructs the sending end to directly send the first data.

91. The device according to claim 83, characterized in that The communication module is further used for: receiving a second reference signal; sending third information, where the third information is the same as the first information, or the third information instructs the transmitting end to send the reference signal after a first time interval, and the third information is obtained based on the second reference signal; Receive data sent by the sending end within the first time.

92. A communication device, characterized in that include: A communication module, configured to send a first reference signal; a processing module, configured to, upon receiving first information indicating K first target branch adaptation layers and weights of the K first target branch adaptation layers in a first target model of a first reference signal receiving end, determine, based on the first information, K second target branch adaptation layers and weights of the K second target branch adaptation layers corresponding to the K first target branch adaptation layers, wherein the first information is obtained based on the first reference signal, and K is a positive integer; The processing module is further configured to input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data; The communication module is further configured to send the first data.

93. The device according to claim 92, characterized in that When M is less than K, the processing module is further used to update the third target model to obtain the second target model, so that the second target model includes the K second target branch adaptive layers, wherein the third target model includes M second target branch adaptive layers, and M is a positive integer.

94. The device according to claim 92 or 93, characterized in that The K second target branch adaptive layers are used to process the encoded data to be sent respectively to obtain the data processed by the K second target branch adaptive layers respectively; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed by the K second target branch adaptive layers respectively to obtain the first data.

95. The device according to any one of claims 92 to 94, characterized in that The communication module is further used for: When second information is received, and the second information indicates to directly send the coded data to be sent, the coded data to be sent is sent, and the second information is obtained according to the first reference signal.

96. The device according to any one of claims 92 to 95, characterized in that The communication module is further used for: receiving a first channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

97. The device according to any one of claims 92 to 95, characterized in that The communication module is further used for: Sending a second channel dictionary set; First indication information is received, where the first indication information is used to indicate a correspondence between a second target branch adaptation layer in the second channel dictionary set and a first target branch adaptation layer in the first channel dictionary set.

98. The device according to claim 96 or 97, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

99. The device according to any one of claims 92 to 98, characterized in that The communication module is further used for: sending a second reference signal; receiving third information, where the third information is the same as the first information; or, the third information is used to indicate that a reference signal is to be sent after a first time interval, and the third information is obtained based on the second reference signal; Data is sent within the first time.

100. A communication device, characterized in that: include: A communication module, configured to send a first reference signal; a processing module, configured to, upon receiving fourth information indicating K second target branch adaptive layers and weights of the K second target branch adaptive layers, input the coded data to be sent into the K second target branch adaptive layers of the second target model for processing to obtain first data, according to the fourth information, wherein the fourth information is obtained based on the first reference signal, and K is a positive integer; The communication module is further configured to send the first data.

101. The device according to claim 100, characterized in that The K second target branch adaptive layers are used to process the encoded data to be sent respectively to obtain the data processed by the K second target branch adaptive layers respectively; the weights of the K second target branch adaptive layers are used to perform weighted sum processing on the data processed by the K second target branch adaptive layers respectively to obtain the first data.

102. The device according to claim 100 or 101, characterized in that The communication module is further used for: When second information is received, and the second information indicates to directly send the coded data to be sent, the coded data to be sent is sent, and the second information is obtained according to the first reference signal.

103. The device according to any one of claims 100 to 102, characterized in that The communication module is further used for: receiving a first channel dictionary set; The processing module is further configured to perform dictionary alignment on the first channel dictionary set and the second channel dictionary set to obtain a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set; The communication module is further configured to send first indication information, where the first indication information is used to indicate a correspondence between the second target branch adaptation layer in the second channel dictionary set and the first target branch adaptation layer in the first channel dictionary set.

104. The device according to any one of claims 100 to 102, characterized in that The communication module is further used for: A second channel dictionary set is sent, where the second channel dictionary set is used by a reference signal receiving end to determine a correspondence between N first target branch adaptation layers of the reference signal receiving end and M second target branch adaptation layers of a transmitting end, where M and N are both positive integers.

105. The device according to claim 103 or 104, characterized in that The second channel dictionary set includes a correspondence between a second target branch adaptation layer and a channel label.

106. The method according to any one of claims 100 to 105, characterized in that The processing module is further configured to: sending a second reference signal; receiving fifth information, where the fifth information is the same as the fourth information; or, the fifth information is used to indicate that the reference signal is to be sent after a first time interval, and the fifth information is obtained based on the second reference signal; Data is sent within the first time.

107. A communication device, characterized in that The communication device includes one or more processors; wherein the one or more processors are used to execute computer programs stored in one or more memories, so that the communication device implements the method according to any one of claims 1 to 6, or implements the method according to any one of claims 7 to 12, or implements the method according to any one of claims 13 to 19, or implements the method according to any one of claims 20 to 25, or implements the method according to any one of claims 26 to 38, or implements the method according to any one of claims 39 to 46, or implements the method according to any one of claims 47 to 53.

108. The communication device according to claim 107, characterized in that The communication device also includes the one or more memories.

109. The communication device according to claim 107 or 108, characterized in that The communication device is a chip or a chip system.

110. A computer-readable storage medium, characterized in that The computer-readable storage medium stores instructions, which, when executed by a processor, implement the method according to any one of claims 1 to 6, or implement the method according to any one of claims 7 to 12, or implement the method according to any one of claims 13 to 19, or implement the method according to any one of claims 20 to 25, or implement the method according to any one of claims 26 to 38, or implement the method according to any one of claims 39 to 46, or implement the method according to any one of claims 47 to 53.

111. A computer program product, characterized in that Comprising a computer program which, when executed, implements the method according to any one of claims 1 to 6, or implements the method according to any one of claims 7 to 12, or implements the method according to any one of claims 13 to 19, or implements the method according to any one of claims 20 to 25, or implements the method according to any one of claims 26 to 38, or implements the method according to any one of claims 39 to 46, or implements the method according to any one of claims 47 to 53.

112. A communication system, characterized in that The method comprises the apparatus as described in any one of claims 54 to 59, and the apparatus as described in any one of claims 66 to 72, or the communication system comprises the apparatus as described in any one of claims 60 to 65, and the apparatus as described in any one of claims 73 to 78, or the communication system comprises the apparatus as described in any one of claims 79 to 91, and the apparatus as described in any one of claims 92 to 99, or the apparatus as described in any one of claims 100 to 106.

Images