Information transmission method and device

Abstract

The embodiment of the invention provides an information transmission method and device. The method comprises the steps that first equipment determines a first encoder and a first decoder; and the first device performs channel estimation based on a second reference signal from the second device to determine a second CSI, compresses the second CSI by using the first encoder to obtain first channel feature information, and sends the first channel feature information to the second device. And the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, and the second device determines the first encoder or the first decoder based on the first CSI and the first channel feature information. Wherein the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal. The method can reduce communication overhead.

Description

Information transmission method and device Technical Field

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

[0002] Multiple-input, multiple-output (MIMO) technology can be used in wireless communication systems to increase system capacity. In this scenario, network devices can precode the signals to be transmitted before sending them to terminal devices, which helps suppress the interference caused by the multiple antennas used in MIMO technology. Specifically, the network device can send a downlink reference signal to the terminal device. The terminal device can perform channel estimation based on the downlink reference signal (RS) and feed back the obtained channel state information (CSI) to the network device. The network device can then precode the signals to be transmitted based on the CSI fed back by the terminal device.

[0003] In one approach, network devices and end devices can implement CSI feedback using an autoencoder. The autoencoder consists of an encoder and a decoder, with the encoder deployed on the end device and the decoder on the network device. The end device can then use the encoder to compress the CSI and send the compressed channel characteristics to the network device. The network device can then use the decoder to decompress the received channel characteristics to obtain the CSI.

[0004] Regarding the process of deploying encoders and decoders on both sides, how to reduce the communication overhead in this process is a technical problem that needs to be solved.

[0005] Summary of the Invention

[0006] The embodiments of the present application provide an information transmission method and apparatus, which can reduce communication overhead.

[0007] In the first aspect, the present application provides an information transmission method, which can be applied to a second device, a chip in the second device, or a logic module or software that can realize all or part of the functions of the second device. The following description takes the second device as an example. The method includes: the second device performs channel estimation based on a first reference signal from the first device to determine a first CSI; the second device sends a second reference signal to the first device, and the second reference signal is used by the first device to perform channel estimation to determine a second CSI; the second device receives first channel characteristic information from the first device, and the first channel characteristic information is obtained by the first device using a first encoder to compress the second CSI; the second device determines the first encoder or the first decoder based on the first CSI and the first channel characteristic information. The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0008] In which, the first encoder and the first decoder belong to the same autoencoder, and the first encoder and the first decoder are adapted to each other. Since the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal, the channel characteristics corresponding to the first reference signal are substantially the same as the channel characteristics corresponding to the second reference channel. The first CSI determined by the second device through channel estimation based on the first reference signal is substantially the same as the second CSI determined by the first device through channel estimation based on the second reference signal. Therefore, the channel characteristic information obtained by the second device using the first encoder to compress the first CSI is substantially the same as the first channel characteristic information obtained by the first device using the first encoder to compress the second CSI. Therefore, the second device can determine the first encoder based on the first CSI and the first channel characteristic information. In addition, since the first CSI determined by the second device through channel estimation based on the first reference signal is substantially the same as the second CSI determined by the first device through channel estimation based on the second reference signal, the CSI obtained by the second device through decompression of the first channel characteristic information using the first decoder is substantially the same as the first CSI. Therefore, the second device can determine the first decoder based on the first channel characteristic information and the first CSI.

[0009] It can be seen that this information transmission method utilizes the reciprocity between the channel corresponding to the first reference signal and the channel corresponding to the second reference signal to enable the second device to determine the first encoder adapted to the first decoder in the first device, or to enable the second device to determine the first decoder adapted to the first encoder in the first device. In this information transmission method, the first device does not need to send the second CSI to the second device, which can reduce communication overhead. In addition, in this information transmission method, the first device does not need to directly send the model parameters of the encoder or decoder obtained by joint training to the second device, which ensures the privatization of the encoder or decoder in the first device, which is conducive to improving the privacy security of the encoder or decoder in the first device when the first device and the second device are from different manufacturers.

[0010] In an optional embodiment, the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the first CSI is the first channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the first channel estimation information.

[0011] Optionally, the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device performs eigendecomposition, such as singular value decomposition (SVD), on the first channel estimation information to obtain the first CSI.

[0012] In an optional embodiment, the second CSI is determined by the first device through channel estimation based on the second reference signal, including: the second channel estimation information is obtained by the first device through channel estimation based on the second reference signal; the second CSI is the second channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the second channel estimation information.

[0013] In an optional embodiment, the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device uses a channel transformation machine learning model to transform the first channel estimation information to obtain third channel estimation information; the first CSI is the third channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information.

[0014] Optionally, the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device uses a channel transformation machine learning model to transform the first channel estimation information to obtain third channel estimation information; the second device performs feature decomposition, such as SVD, on the third channel estimation information to obtain the first CSI.

[0015] In an optional embodiment, the second device performs channel estimation based on a first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI.

[0016] Optionally, the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device performs eigendecomposition, such as SVD, on the first channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information; the second device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI.

[0017] Among them, the channel transformation machine learning model is related to the channel data distribution. The second device uses the channel transformation machine learning model to transform the first channel estimation information. This implementation method, or the second device uses the channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information, is beneficial to make the channel data distribution in the actual deployment scenario of the first encoder and the first decoder different from the channel data distribution corresponding to the first reference signal and the second reference signal. The first CSI can characterize the channel characteristics in the actual deployment scenario, which is beneficial to reducing the performance degradation of the first encoder or the first decoder determined by the second device based on the first CSI in the actual deployment scenario, that is, reducing the deviation between the channel state information recovered by the first decoder in the actual deployment scenario and the actual channel state information input to the first encoder, which is beneficial to improving communication quality.

[0018] In an optional embodiment, the second CSI is determined by the first device performing channel estimation based on the second reference signal, including: the second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; the fourth channel estimation information is obtained by the first device transforming the second channel estimation information using a channel transformation machine learning model; the second CSI is the fourth channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information.

[0019] In an optional embodiment, the second CSI is determined by the first device performing channel estimation based on the second reference signal, including: the second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; the second CSI is obtained by transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information using a channel transformation machine learning model.

[0020] Among them, the channel transformation machine learning model is related to the channel data distribution, and the fourth channel estimation information is obtained by the first device using the channel transformation machine learning model to transform the second channel estimation information. Alternatively, the second CSI is obtained by using the channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information. This is beneficial for making the channel data distribution in the actual deployment scenario of the first encoder and the first decoder different from the channel data distribution corresponding to the first reference signal and the second reference signal. The second CSI can characterize the channel characteristics in the actual deployment scenario, which is beneficial to reducing the performance degradation of the first encoder or the first decoder determined by the second device based on the first channel characteristic information in the actual deployment scenario, that is, reducing the deviation between the channel state information recovered by the first decoder in the actual deployment scenario and the actual channel state information input to the first encoder, which is beneficial to improving communication quality.

[0021] In an optional embodiment, after the second device determines the first encoder based on the first CSI and the first channel characteristic information, the method further includes: the second device uses a parameter sharing machine learning model to transform the first encoder to obtain a second encoder.

[0022] In an optional embodiment, after the second device determines the first decoder based on the first CSI and the first channel characteristic information, the method further includes: the second device uses a parameter sharing machine learning model to transform the first decoder to obtain a second decoder.

[0023] The parameter sharing machine learning model is related to the communication system parameter configuration, which may include one or more of the following: the number of antenna elements, the number of antenna ports, the number of spatial streams, etc. Transforming the first encoder and the first decoder using the parameter sharing machine learning model facilitates adapting the transformed second encoder and second decoder to the actual deployment scenario when the communication system parameter configuration of the first encoder and the first decoder differs from the communication system parameter configuration in the scenario where the second device determines the first encoder, thereby improving communication quality.

[0024] In an optional embodiment, the first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; the second device is a terminal device or a chip in the terminal device, and the first device is a network device or a chip in the network device; the second device determines the first encoder based on the first CSI and the first channel characteristic information.

[0025] Optionally, the method also includes: the second device receives one or more of the following from the first device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0026] In an optional embodiment, the first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; the second device is a network device or a chip in the network device, and the first device is a terminal device or a chip of the terminal device; the second device is determined to be a first decoder based on the first CSI and the first channel characteristic information.

[0027] Optionally, the method also includes: the second device sends one or more of the following to the first device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0028] On the second aspect, the present application provides an information transmission method, which can be applied to a first device, a chip in the first device, or a logic module or software that can realize all or part of the functions of the first device. The following description takes the first device as an example. The method includes: the first device determines a first encoder and a first decoder; the first device sends a first reference signal to the second device, and the first reference signal is used by the second device to perform channel estimation to determine a first CSI; the first device performs channel estimation based on a second reference signal from the second device to determine a second CSI; the first device compresses the second CSI using the first encoder to obtain first channel characteristic information; the first device sends the first channel characteristic information to the second device. The first CSI and the first channel characteristic information are used by the second device to determine the first encoder or the first decoder; the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0029] In which, the first encoder and the first decoder belong to the same autoencoder, and the first encoder and the first decoder are adapted to each other. Since the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal, the channel characteristics corresponding to the first reference signal are substantially the same as the channel characteristics corresponding to the second reference channel. The second CSI determined by the first device through channel estimation based on the second reference signal is substantially the same as the first CSI determined by the second device through channel estimation based on the first reference signal. Therefore, the first channel characteristic information obtained by the first device using the first encoder to compress the second CSI is substantially the same as the channel characteristic information obtained by the second device using the first encoder to compress the first CSI. Therefore, this method facilitates the second device to determine the first encoder based on the first CSI and the first channel characteristic information. In addition, since the second CSI determined by the first device through channel estimation based on the second reference signal is substantially the same as the first CSI determined by the second device through channel estimation based on the first reference signal, the CSI obtained by the second device through decompression of the first channel characteristic information using the first decoder is substantially the same as the first CSI. Therefore, this method facilitates the second device to determine the first decoder based on the first channel characteristic information and the first CSI.

[0030] It can be seen that this information transmission method utilizes the reciprocity between the channel corresponding to the first reference signal and the channel corresponding to the second reference signal, which is conducive to the second device determining the first encoder adapted to the first decoder in the first device, or the second device determining the first decoder adapted to the first encoder in the first device. In this information transmission method, the first device does not need to send the second CSI to the second device, which can reduce communication overhead. In addition, in this information transmission method, the first device does not need to directly send the model parameters of the encoder or decoder obtained by joint training to the second device, which ensures the privatization of the encoder or decoder in the first device, which is conducive to improving the privacy security of the encoder or decoder in the first device when the first device and the second device are from different manufacturers.

[0031] In an optional embodiment, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the second CSI is the second channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the second channel estimation information.

[0032] Optionally, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device performs eigendecomposition, such as SVD, on the second channel estimation information to obtain the second CSI.

[0033] In an optional embodiment, the first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: the first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; the first CSI is the first channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the first channel estimation information.

[0034] In an optional embodiment, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device uses a channel transformation machine learning model to transform the second channel estimation information to obtain fourth channel estimation information; the second CSI is the fourth channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information.

[0035] Optionally, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device uses a channel transformation machine learning model to transform the second channel estimation information to obtain fourth channel estimation information; the first device performs feature decomposition, such as SVD, on the fourth channel estimation information to obtain the second CSI.

[0036] In an optional embodiment, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the second CSI.

[0037] Optionally, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device performs eigendecomposition, such as SVD, on the second channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information; the first device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the second CSI.

[0038] Among them, the channel transformation machine learning model is related to the channel data distribution. The first device uses the channel transformation machine learning model to transform the second channel estimation information. This implementation method, or the first device uses the channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information, is beneficial to make the channel data distribution in the actual deployment scenario of the first encoder and the first decoder different from the channel data distribution corresponding to the first reference signal and the second reference signal. The second CSI can characterize the channel characteristics in the actual deployment scenario, which is beneficial to reducing the performance degradation of the first encoder or the first decoder determined by the second device based on the first channel characteristic information in the actual deployment scenario, that is, reducing the deviation between the channel state information recovered by the first decoder in the actual deployment scenario and the actual channel state information input to the first encoder, which is beneficial to improving communication quality.

[0039] In an optional embodiment, the first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: the first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; the third channel estimation information is obtained by transforming the first channel estimation information using a channel transformation machine learning model; the first CSI is the third channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information.

[0040] In an optional embodiment, the first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: the first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; the first CSI is obtained by transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information using a channel transformation machine learning model.

[0041] Among them, the channel transformation machine learning model is related to the channel data distribution, and the third channel estimation information is obtained by the second device using the channel transformation machine learning model to transform the first channel estimation information. Alternatively, the first CSI is obtained by using the channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information. This embodiment is beneficial for making the channel data distribution in the actual deployment scenario of the first encoder and the first decoder different from the channel data distribution corresponding to the first reference signal and the second reference signal. The first CSI can characterize the channel characteristics in the actual deployment scenario, which is beneficial for reducing the performance degradation of the first encoder or the first decoder determined by the second device based on the first CSI in the actual deployment scenario, that is, reducing the deviation between the channel state information recovered by the first decoder in the actual deployment scenario and the actual channel state information input to the first encoder, which is beneficial for improving communication quality.

[0042] In an optional embodiment, the method further includes: the first device transforms the first decoder using a parameter sharing machine learning model to obtain a second decoder.

[0043] In an optional embodiment, the method further includes: the first device uses a parameter sharing machine learning model to transform the first encoder to obtain a second encoder.

[0044] The parameter sharing machine learning model is related to the communication system parameter configuration, which may include one or more of the following: the number of antenna elements, the number of antenna ports, the number of spatial streams, etc. Transforming the first encoder and the first decoder using the parameter sharing machine learning model facilitates adapting the transformed second encoder and second decoder to the actual deployment scenario when the communication system parameter configuration of the first encoder and the first decoder differs from the communication system parameter configuration in the scenario where the second device determines the first encoder, thereby improving communication quality.

[0045] In an optional embodiment, the first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; the first device is a network device or a chip in a network device, and the second device is a terminal device or a chip in a terminal device; the first CSI and the first channel characteristic information are used by the second device to determine the first encoder.

[0046] Optionally, the method also includes: the first device sends one or more of the following to the second device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0047] In an optional embodiment, the first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; the first device is a terminal device or a chip in the terminal device, and the second device is a network device or a chip in the network device; the first CSI and the first channel characteristic information are used by the second device to determine the first decoder.

[0048] Optionally, the method also includes: the first device receives one or more of the following from the second device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position corresponding to the first reference signal; the second information is used to indicate the time-frequency resource position corresponding to the second reference signal; and the third information is used to indicate the time-frequency resource position corresponding to the first channel characteristic information.

[0049] In a third aspect, the present application further provides a communication device. The communication device has the function of implementing some or all of the embodiments described in the first aspect above, or has the function of implementing some or all of the functional implementations described in the second aspect above. The functions can be implemented by hardware, or by hardware executing corresponding software. The hardware or software includes one or more units or modules corresponding to the above functions.

[0050] In one possible design, the communication device may include a processing unit and a communication unit. The processing unit is configured to support the communication device in performing the corresponding functions in the above method. The communication unit is configured to support communication between the communication device and other communication devices. The communication device may also include a storage unit, which is coupled to the processing unit and the communication unit and stores program instructions and data necessary for the communication device.

[0051] In one embodiment, the communication device includes: a processing unit and a communication unit, and the processing unit is used to control the communication unit to send and receive data / signaling.

[0052] The processing unit is configured to perform channel estimation based on a first reference signal from the first device to determine a first CSI.

[0053] The communication unit is configured to send a second reference signal to the first device, where the second reference signal is used by the first device to perform channel estimation and determine a second CSI.

[0054] The communication unit is further configured to receive first channel characteristic information from the first device, where the first channel characteristic information is obtained by the first device compressing the second CSI using a first encoder.

[0055] The processing unit is further configured to determine a first encoder or a first decoder based on the first CSI and the first channel characteristic information.

[0056] The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0057] In addition, in this aspect, other optional implementations of the communication device can refer to the relevant content of the first aspect above and will not be described in detail here.

[0058] In another embodiment, the communication device includes: a processing unit and a communication unit, and the processing unit is used to control the communication unit to send and receive data / signaling.

[0059] The processing unit is configured to determine a first encoder and a first decoder.

[0060] The communication unit is configured to send a first reference signal to the second device, where the first reference signal is used by the second device to perform channel estimation and determine a first CSI.

[0061] The processing unit is further configured to perform channel estimation based on a second reference signal from the second device to determine a second CSI.

[0062] The processing unit is further configured to compress the second CSI using the first encoder to obtain first channel characteristic information.

[0063] The communication unit is further configured to send the first channel characteristic information to the second device.

[0064] The first CSI and the first channel characteristic information are used by the second device to determine the first encoder or the first decoder; the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0065] In addition, in this aspect, other optional implementations of the communication device can refer to the relevant content of the second aspect above and will not be described in detail here.

[0066] As an example, the communication unit may be a transceiver or a communication interface, the storage unit may be a memory, and the processing unit may be a processor.

[0067] In one embodiment, the communication device includes: a processor and a transceiver. The processor is used to perform channel estimation based on a first reference signal from the first device to determine a first CSI. The transceiver is used to send a second reference signal to the first device, and the second reference signal is used by the first device to perform channel estimation to determine a second CSI. The transceiver is also used to receive first channel characteristic information from the first device, where the first channel characteristic information is obtained by the first device compressing the second CSI using a first encoder. The processor is also used to determine a first encoder or a first decoder based on the first CSI and the first channel characteristic information. The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0068] In addition, in this aspect, other optional implementations of the communication device can refer to the relevant content of the first aspect above and will not be described in detail here.

[0069] In another embodiment, the communication device includes: a processor and a transceiver. The processor is used to determine a first encoder and a first decoder. The transceiver is used to send a first reference signal to a second device, and the first reference signal is used by the second device to perform channel estimation to determine a first CSI. The processor is also used to perform channel estimation based on a second reference signal from the second device to determine a second CSI. The processor is also used to compress the second CSI using the first encoder to obtain first channel characteristic information. The transceiver is also used to send the first channel characteristic information to the second device. The first CSI and the first channel characteristic information are used by the second device to determine the first encoder or the first decoder; the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0070] In addition, in this aspect, other optional implementations of the communication device can refer to the relevant content of the first aspect above and will not be described in detail here.

[0071] In another embodiment, the communication device is a chip or a chip system. The processing unit may also be embodied as a processing circuit or a logic circuit; and the transceiver unit may be an input / output interface, an interface circuit, an output circuit, an input circuit, a pin, or a related circuit on the chip or chip system.

[0072] During implementation, the processor can be used to perform, for example, but not limited to, baseband-related processing, and the transceiver or communication interface can be used to perform, for example, but not limited to, radio frequency transceiver processing. The aforementioned components can be provided on separate chips, or at least partially or entirely on the same chip. For example, the processor can be further divided into an analog baseband processor and a digital baseband processor. The analog baseband processor can be integrated with the transceiver (or communication interface) on the same chip, while the digital baseband processor can be provided on a separate chip. With the continuous development of integrated circuit technology, more and more components can be integrated on the same chip. For example, a digital baseband processor can be integrated with multiple application processors (such as, but not limited to, a graphics processor, a multimedia processor, etc.) on the same chip. Such a chip can be called a system on a chip (SoC). Whether each component is provided independently on different chips or integrated on one or more chips often depends on the needs of the product design. The embodiments of the present application do not limit the implementation form of the aforementioned components.

[0073] In a fourth aspect, the present application also provides a processor for executing the various methods described above. In the process of executing these methods, the process of sending the above-mentioned signal and receiving the above-mentioned signal in the above-mentioned method can be understood as the process of outputting the above-mentioned signal by the processor, and the process of inputting the above-mentioned signal by the processor. When outputting the above-mentioned signal, the processor outputs the above-mentioned signal to the transceiver so that it is transmitted by the transceiver (or communication interface). After being output by the processor, the above-mentioned signal may also need to undergo other processing before reaching the transceiver (or communication interface). Similarly, when the processor receives the above-mentioned input signal, the transceiver (or communication interface) receives the above-mentioned signal and inputs it into the processor. Furthermore, after the transceiver (or communication interface) receives the above-mentioned signal, the above-mentioned signal may need to undergo other processing before being input into the processor.

[0074] For the sending and receiving operations involved in the processor, unless otherwise specified, or unless they conflict with their actual functions or internal logic in the relevant descriptions, they can be more generally understood as processor output, reception, input and other operations, rather than sending and receiving operations directly performed by the RF circuit and antenna.

[0075] During implementation, the processor may be a processor specifically configured to execute these methods, or may be a processor that executes computer instructions in a memory to execute these methods, such as a general-purpose processor. The memory may be a non-transitory memory, such as a read-only memory (ROM), which may be integrated with the processor on the same chip or disposed on separate chips. The embodiments of the present application do not limit the type of memory or the configuration of the memory and the processor.

[0076] In a fifth aspect, the present application further provides a communication system, which includes at least one first device and at least one second device according to the above aspects. In another possible design, the system may also include other devices that interact with the first device and / or the second device in the solution provided by the present application.

[0077] In a sixth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is run, the method described in the first or second aspect above is executed.

[0078] In a seventh aspect, the present application further provides a computer program product comprising instructions, the computer program product comprising: computer program code, which, when the computer program code is run, enables the method described in the first aspect or the second aspect above to be executed.

[0079] In an eighth aspect, the present application provides a chip system, which includes a processor and an interface, wherein the interface is used to obtain a program or instruction, and the processor is used to call the program or instruction to implement the functions involved in the first aspect, or to call the program or instruction to implement the functions involved in the second aspect. In one possible design, the chip system also includes a memory, which is used to store program instructions and data necessary for the terminal. The chip system can be composed of a chip, or it can include a chip and other discrete devices. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] FIG1 is a schematic diagram of a communication system provided in an embodiment of the present application;

[0081] FIG2 is a schematic diagram of another communication system provided in an embodiment of the present application;

[0082] FIG3a is a schematic diagram of a fully connected neural network provided in an embodiment of the present application;

[0083] FIG3 b is a schematic diagram of a loss function optimization provided in an embodiment of the present application;

[0084] FIG3c is a schematic diagram of a gradient back propagation provided in an embodiment of the present application;

[0085] FIG4 is a schematic diagram of an autoencoder provided in an embodiment of the present application;

[0086] FIG5 is a schematic diagram of MIMO channel compression feedback based on an autoencoder provided in an embodiment of the present application;

[0087] FIG6 is a schematic diagram of an antenna connection provided in an embodiment of the present application;

[0088] FIG7 is a schematic diagram of a transmitting antenna array provided in an embodiment of the present application;

[0089] FIG8 is a schematic diagram of a graph model provided in an embodiment of the present application;

[0090] FIG9 is a flow chart of an information transmission method provided in an embodiment of the present application;

[0091] FIG10 is a schematic diagram of an encoder and decoder based on a graph model according to an embodiment of the present application;

[0092] FIG11 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0093] FIG12 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0094] FIG13 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0095] FIG14 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0096] FIG15 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0097] FIG16 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0098] FIG17 is a schematic diagram of another information transmission method provided in an embodiment of the present application;

[0099] FIG18 is a schematic structural diagram of a communication device provided in an embodiment of the present application;

[0100] FIG19 is a schematic structural diagram of another communication device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0101] The technical solutions in the embodiments of the present application are described clearly and completely below in conjunction with the drawings in the embodiments of the present application.

[0102] In order to better understand the information transmission method disclosed in the embodiment of the present application, a communication system applicable to the embodiment of the present application is described.

[0103] The embodiments of the present application can be applied to fourth-generation (4G) communication systems such as long-term evolution (LTE) systems, fifth-generation (5G) communication systems such as new radio (NR) systems, and can also be applied to communication systems such as satellite communication and short-range communication, communication systems that support the integration of multiple wireless technologies, or communication systems that have evolved after 5G, such as sixth-generation (6G) communication systems.

[0104] In the embodiments of the present application, wireless communication systems include but are not limited to: narrowband Internet of Things (NB-IoT) system, 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 low latency communication (URLLC) and massive machine type communication (mMTC), or mobile communication systems after 5G.

[0105] Figure 1 shows the architecture of a wireless communication system. A wireless communication system may include one or more network devices and one or more terminal devices. Network devices and terminal devices can communicate with each other, and different terminal devices can also communicate with each other. Figure 1 uses a wireless communication system including one network device and two terminal devices as an example.

[0106] Please refer to Figure 2, which is a schematic diagram of the structure of another communication system provided in an embodiment of the present application. The communication system includes but is not limited to a first device and a second device. The number and configuration of the devices shown in Figure 2 are for illustrative purposes only and do not constitute a limitation on the embodiment of the present application. In actual applications, two or more first devices and two or more second devices may be included. The first device may be a network device or a terminal device. The second device may be a network device or a terminal device.

[0107] In an embodiment of the present application, the network device is a device with wireless transceiver functions. In one possible scenario, the network device may be a base station, an evolved NodeB (eNodeB), an access point (AP), a transmission reception point (TRP), a next generation NodeB (gNB), a next generation base station in a 6G mobile communication system, a base station in a future mobile communication system, etc. The network device may be a macro base station, a micro base station or an indoor station, a relay node or a donor node, or a wireless controller in a cloud radio access network (C-RAN) scenario. Optionally, the network device may also be a server, a wearable device, a vehicle, an on-board device, a mobile switching center, and a device that performs base station functions in device-to-device (D2D), vehicle-to-everything (V2X), and machine-to-machine (M2M) communications. For example, the access network device in vehicle to everything (V2X) technology may be a road side unit (RSU).

[0108] In another possible scenario, multiple network devices assist the terminal device in achieving wireless access, and different network devices respectively implement part of the functions of the base station. For example, the network device can be a centralized unit (CU), a distributed unit (DU), a CU-control plane (CP), a CU-user plane (UP), or a radio unit (RU). The CU and DU can be set separately, or they can be included in the same network element, such as a baseband unit (BBU). The RU can be included in a radio frequency device or radio frequency unit, such as a remote radio unit (RRU), an active antenna unit (AAU), or a remote radio head (RRH).

[0109] In different systems, CU (or CU-CP and CU-UP), DU or RU may also have different names, but those skilled in the art can understand their meanings. For example, in an open wireless access network (ORAN) system, CU may also be called O-CU (open CU), DU may also be called O-DU, CU-CP may also be called O-CU-CP, CU-UP may also be called O-CU-UP, and RU may also be called O-RU. For the convenience of description, this application takes CU, CU-CP, CU-UP, DU and RU as examples for description. Any unit of CU (or CU-CP, CU-UP), DU and RU in this application can be implemented by a software module, a hardware module, or a combination of a software module and a hardware module.

[0110] In the embodiments of the present application, the terminal device may also be referred to as user equipment (UE), terminal, access terminal, subscriber unit, user station, mobile station, mobile station (MS), remote station, remote terminal, mobile terminal, mobile device, user terminal, user agent or user device, and may be applied to 4G, 5G or even 6G systems. The terminal may 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 device in the embodiments of the present application may be a handheld device with wireless communication function, a vehicle-mounted device, a wearable device, a computing device or other processing device connected to a wireless modem. The terminal device may be a terminal with a function of connecting to a cellular base station. For example, the terminal device may be a cellular phone, a smart phone, a tablet computer, a wireless data card, a personal digital assistant (PDA), a tablet computer, a wireless modem, a handheld device, a laptop computer, a machine type communication (MTC) terminal, etc. The terminal device may also be a virtual reality (VR) terminal device, an augmented reality (AR) terminal device, a wireless terminal in industrial control, a wireless terminal in self-driving, a wireless terminal in remote medical care, a wireless terminal in a smart grid, a wireless terminal in transportation safety, a wireless terminal in a smart city, a wireless terminal in a smart home, an in-vehicle terminal, a wireless communication device in a smart factory, etc.

[0111] The embodiments disclosed herein will present various aspects, embodiments, or features of the present invention centered around a system comprising multiple devices, components, modules, etc. It should be understood that each system may include additional devices, components, modules, etc., and / or may not include all of the devices, components, modules, etc. discussed in conjunction with the accompanying drawings. Furthermore, combinations of these solutions may also be used.

[0112] In order to better understand the information transmission method disclosed in the embodiments of the present application, a brief introduction to the relevant concepts involved in the embodiments of the present application is given.

[0113] 1. Fully Connected Neural Network

[0114] A fully connected neural network is also known as a multilayer perceptron (MLP). An MLP consists of an input layer, an output layer, and multiple hidden layers. Each layer contains several nodes, which are called neurons. Neurons in adjacent layers are connected to each other. For example, the MLP shown in Figure 3a consists of an input layer, an output layer, and two hidden layers. The input layer has four neurons, each hidden layer has eight neurons, and the output layer has six neurons.

[0115] For neurons in two adjacent layers, the output h of the neuron in the next layer is obtained by processing the weighted sum of all neurons x in the previous layer connected to the neuron using the activation function. This can be expressed as shown in the matrix formula (1). h = f(wx + b) (1)

[0116] Among them, w is the weight matrix, b is the bias vector, and f() is the activation function.

[0117] Then, the output of the neural network can be recursively expressed as shown in formula (2). n (w n f n-1 (…)+b n ) (2)

[0118] As can be seen, neural networks can represent the mapping relationship between input data sets and output data sets. Neural networks are typically initialized randomly and require training before they can be used. Training a neural network is the process of determining this mapping relationship from random w and b using existing data.

[0119] Referring to Figure 3b, the specific method for training a neural network includes: using a loss function to evaluate the output of the neural network, backpropagating the error, and iteratively optimizing w and b using the gradient descent method until the loss function reaches a minimum. The gradient descent process can be expressed as shown in Formula (3).

[0120] Among them, θ is the parameter to be optimized (such as w and b), L is the loss function; η is the learning efficiency, which is used to control the step size of gradient descent.

[0121] In addition, the chain rule for partial derivatives is used during backpropagation, meaning that the gradient of the parameters in the previous layer can be recursively calculated from the gradient of the parameters in the next layer. For example, the gradient of the weight wij between neurons j and i in Figure 3c can be expressed as shown in formula (4).

[0122] Where si is the weighted sum of the inputs to neuron i. In addition, in formula (4), It can be called the intermediate layer gradient, that is, s i Considered as the middle layer.

[0123] 2. AutoEncoder

[0124] An autoencoder is a machine learning model structure. For example, a neural network model is a type of machine learning model, and an autoencoder can be a neural network model structure. An autoencoder can be used to represent a set of data, or it can be understood as encoding a set of data, and is typically used for data dimensionality reduction. Referring to Figure 4 , an autoencoder includes an encoder and a decoder. The encoder can be used to encode input data through one or more layers of a neural network to obtain a codeword after dimensionality reduction. The decoder can be used to reconstruct the codeword into output data through one or more layers of a neural network, and the output data is required to be as identical as possible to the data input to the encoder. The encoder and decoder need to be trained jointly to ensure that the output data of the decoder is identical or substantially identical to the input data of the encoder. In addition, an autoencoder can also be referred to as an autoencoder machine learning model, an encoder can also be referred to as an encoder machine learning model, and a decoder can also be referred to as a decoder machine learning model.

[0125] 3. MIMO system precoding

[0126] In wireless communication systems, MIMO technology is often used to increase system capacity. This involves using multiple antennas to transmit signals simultaneously at both the transmitter and receiver. In theory, the use of multiple antennas combined with spatial division multiplexing can exponentially increase system capacity. However, in practice, the use of multiple antennas can increase interference, necessitating certain signal processing to mitigate the effects of interference. Interference suppression through signal processing can be implemented at both the receiver and the transmitter. Precoding is one method of signal processing at the transmitter to achieve interference suppression. Before transmitting the signal over the MIMO channel, the transmitter can precode the signal to be transmitted, thereby mitigating the interference caused by the use of multiple antennas.

[0127] Among them, the transmitting end may precode the signal to be transmitted by using a precoding matrix to process the signal to be transmitted. In principle, singular value decomposition (SVD) can be performed according to the actual channel to obtain the corresponding precoding matrix. Among them, performing SVD on the MIMO channel can convert multiple channels into a one-to-one mode similar to a single input single output (SISO), that is, converting multiple MIMO cross channels into multiple parallel one-to-one channels, which is conducive to identifying useful channels in multiple MIMO cross channels. Performing SVD on the channel can be shown as the following formula (5). H=U∑V T (5)

[0128] Where H is the channel matrix. V is the precoding matrix, V T is the transposed matrix of V. U and V are orthogonal matrices. ∑ is a diagonal matrix with singular values ​​as diagonal elements.

[0129] For example, the transmitting end may use the precoding matrix V to process the signal x to be transmitted, and obtain the processed signal s=Vx. The transmitting end may send the processed signal s, and the signal r received by the receiving end may be shown in the following formula (6). r=H×s+n=U∑V T ×s+n (6)

[0130] Then, the receiver can use ∑ -1 U T The received signal r is decoded to obtain multiple one-to-one channels without interference. -1 is the inverse matrix of ∑, U T is the transposed matrix of U.

[0131] However, in practical systems, such as frequency-division duplex (FDD) systems, MIMO channels may not be readily available. In this case, a series of precoding matrices can be predefined, forming a codebook. The transmitter and receiver can coordinate to select an appropriate precoding matrix from the codebook. In the 3rd Generation Partnership Project (3GPP), network devices can send downlink reference signals to terminal devices. The terminal devices can perform channel estimation based on the downlink RS and feed back the estimated CSI to the network devices. Based on the precoding matrix indicator (PMI) in the CSI fed back by the terminal devices, the network devices can select a precoding matrix from the codebook that maximizes the channel capacity and then use that precoding matrix to process the signal to be transmitted.

[0132] 4. MIMO channel compression feedback based on autoencoders

[0133] The network device and the terminal device can use an autoencoder to implement CSI feedback from the terminal device to the network device. Specifically, in conjunction with Figure 5, the encoder in the autoencoder is deployed in the terminal device, and the decoder is deployed in the network device. The terminal device can estimate the MIMO channel based on the downlink RS from the network device to determine the CSI, and input the determined CSI (including MIMO channel data) into the trained encoder. The encoder compresses the CSI (including MIMO channel data), and then sends the codeword compressed by the encoder to the network device via the feedback channel. The network device can input the received codeword into the decoder trained together with the encoder, reconstruct the MIMO channel based on the decoder, and thus recover the CSI determined by the terminal device (including the recovered MIMO channel data). The network device then precodes the signal to be transmitted based on the recovered CSI and then sends the precoded signal. In this way, the terminal device uses the encoder to compress the CSI (including MIMO channel data) to reduce the dimension of the CSI, so that the codeword transmitted in the feedback channel is the reduced-dimensional CSI, which can reduce communication overhead.

[0134] 5. Graph Neural Networks (GNN) Model

[0135] The GNN model can also be called a graph model. The following uses the antenna connection diagram shown in Figure 6 as an example to illustrate the construction of the graph model. In Figure 6, the number of transmitting antennas N at the transmitting end is t Equal to 3, the number of receiving antennas at the receiving end is N rEqual to 2, each receiving antenna and N can be considered separately t The channel between each transmitting antenna in the transmit antenna is used to construct the graph model. In this case, the channel matrix in the antenna domain-frequency domain is As shown in formula (7).

[0136] in, is the number of subcarriers.

[0137] right Perform a two-dimensional discrete Fourier transform (DFT) to obtain the angle domain-time domain channel matrix Preserve in the time domain The first N c rows, and the truncated channel matrix is ​​obtained Each column of the matrix is ​​an attribute of a node in the graph model; if the attributes of two nodes are adjacent in the truncated channel matrix, there is an edge between the two nodes. Based on the nodes, edges, and node attributes, the graph model can be obtained. Then, for N r A total of N receiving antennas can be constructed r A graph model.

[0138] Assume that the transmitting antenna array is as shown in Figure 7, that is, N t The transmitting antennas form an antenna array with N2 rows and N1 columns (N1+N2=N t After Fourier transforming the channel matrix, we can get N1×N2 angle points in the angle domain, which are the nodes in the graph model. Due to the sparsity of the time domain, only the first N angle points of the time domain information need to be retained. c values, so the feature of each node is N c dimensional complex vector H[:,i]. In order to facilitate numerical calculations, the imaginary and real parts of the complex vector H[:,i] are usually processed separately and concatenated into a dimension of 2N. c vector x i , the feature dimension of each node in the final graph is 2N c Considering that adjacent angle points will affect each other, there are edges between adjacent angle points in the constructed graph model. For all N1×N2 nodes, define an adjacency matrix with N1×N2 rows and N1×N2 columns. If the element A in the i-th row and j-th column of the adjacency matrix A i,j =1, indicating that there is an edge between node i and node j; if A i,j = 0, indicating that there is no edge between node i and node j. Thus, we can obtain the graph model shown in Figure 8.

[0139] The embodiments of the present application are described below with reference to the accompanying drawings.

[0140] Please refer to Figure 9, which is a flow chart of an information transmission method provided by an embodiment of the present application. The diagram uses the first device and the second device as examples to illustrate the corresponding method, but the present application does not limit the execution subjects of the interaction diagram. For example, the first device in the figure can also be a chip or chip system or processor that supports the first device to implement the corresponding method, or a logic module or software that can implement all or part of the functions of the first device; the second device in the figure can also be a chip or chip system or processor that supports the second device to implement the corresponding method, or a logic module or software that can implement all or part of the functions of the second device. The information transmission method includes the following steps:

[0141] S101: A first device determines a first encoder and a first decoder.

[0142] It is understood that the first device can train the autoencoder to obtain a trained autoencoder. The first device training the autoencoder includes: the first device jointly training the encoder and decoder. The first encoder determined by the first device in step S101 is the encoder after the first device performs the joint training, and the first decoder is the decoder after the first device performs the joint training. It is understood that the first encoder and the first decoder belong to the same autoencoder and are mutually compatible.

[0143] S102. The first device sends a first reference signal to the second device; correspondingly, the second device receives the first reference signal from the first device.

[0144] S103. The second device sends a second reference signal to the first device; correspondingly, the first device receives the second reference signal from the second device.

[0145] The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal. In this embodiment of the present application, the channel corresponding to the reference signal is the channel that the reference signal travels through, and can also be understood as the channel that carries the reference signal, which will not be described in detail below. Exemplarily, the channel corresponding to the first reference signal is the channel that the first reference signal travels through, and can also be understood as the channel that carries the first reference signal. The channel corresponding to the second reference signal is the channel that the second reference signal travels through, and can also be understood as the channel that carries the second reference signal.

[0146] Exemplarily, the first reference signal and the second reference signal may be transmitted using time division duplexing (TDD), which can also be understood as being transmitted in a TDD system. In this case, the first reference signal and the second reference signal may be transmitted in different time slots of the same frequency resource. Since the interval between the transmission time of the first reference signal and the transmission time of the second reference signal is relatively short, it can be assumed that the channel fading experienced by the first reference signal and the second reference signal is substantially the same, indicating that the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0147] S104. The second device performs channel estimation based on the first reference signal to determine a first CSI.

[0148] The following describes an optional implementation of the operation in which the second device performs channel estimation based on the first reference signal to determine the first CSI, as described in the following optional implementations 1.1 to 1.3.

[0149] Implementation 1.1, the second device performs channel estimation based on the first reference signal to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the first CSI is the first channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the first channel estimation information.

[0150] It is understandable that the first channel estimation information can characterize the characteristics of the channel corresponding to the first reference signal. In addition, in the embodiment of the present application, the channel estimation information can be, for example, time domain information, frequency domain information, time-frequency domain information, or delay-Doppler domain information of the channel, etc., which can characterize the channel characteristics, and is not limited to this. The channel estimation information can be represented, for example, by a channel matrix or a channel vector, etc., and is not limited to this.

[0151] Optionally, the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information may be: the eigenvalues ​​or eigenvectors of the channel matrix obtained by the second device processing the first channel estimation information. Exemplarily, the second device may perform eigendecomposition on the first channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information. The eigendecomposition may be, for example, SVD.

[0152] In addition, optionally, in addition to the manner in which the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the first channel estimation information, the first CSI may also be other characteristic information of the channel matrix corresponding to the first channel estimation information, without limitation.

[0153] In implementation 1.2, the second device performs channel estimation based on the first reference signal to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device transforms the first channel estimation information using a channel transformation machine learning model to obtain third channel estimation information; the first CSI is the third channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information. For a detailed description of the first channel estimation information, please refer to the relevant description in implementation 1.1 and will not be repeated here.

[0154] Optionally, the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the third channel estimation information may be: the eigenvalues ​​or eigenvectors of the channel matrix obtained by the second device processing the third channel estimation information. Exemplarily, the second device may perform eigendecomposition on the third channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the third channel estimation information. The eigendecomposition may be, for example, SVD.

[0155] In addition, optionally, in addition to the manner in which the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information, the first CSI may also be other characteristic information of the channel matrix corresponding to the third channel estimation information, without limitation.

[0156] Implementation 1.3, the second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: the second device performs channel estimation based on the first reference signal to obtain first channel estimation information; the second device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI. For a specific explanation of the first channel estimation information and the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information, please refer to the relevant explanation in Implementation 1.1 and will not be repeated here. In addition, optionally, in addition to transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI, other characteristic information of the channel matrix corresponding to the first channel estimation information may also be transformed to obtain the first CSI, without limitation.

[0157] In addition, in implementation 1.2 and implementation 1.3, the channel transformation machine learning model is related to the channel data distribution. In implementation 1.2, the second device uses the channel transformation machine learning model to transform the first channel estimation information. This operation, or in implementation 1.3, the second device uses the channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information. This operation is beneficial to make the channel data distribution in the actual deployment scenario of the first encoder and the first decoder different from the channel data distribution corresponding to the first reference signal and the second reference signal. The first CSI can characterize the channel characteristics in the actual deployment scenario, thereby helping to reduce the performance degradation of the first encoder or the first decoder determined by the second device based on the first CSI in the actual deployment scenario, that is, to reduce the deviation between the channel state information recovered by the first decoder in the actual deployment scenario and the actual channel state information input to the first encoder, which is beneficial to improve communication quality. In the embodiment of the present application, the actual deployment scenario of the first encoder and the first decoder can be understood as a scenario in which the first encoder and the first decoder are actually used to implement compressed feedback of the channel state information.

[0158] For example, the first reference signal and the second reference signal are transmitted in a TDD system, while the first device and the second device use a first encoder and a first decoder in an FDD system to implement compressed feedback of channel state information. The channel data distribution of the FDD system may be different from that of the TDD system. If the second device directly uses the first channel estimation information obtained in the TDD system or the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information as the first CSI, this may cause the performance of the first encoder or the first decoder determined by the second device based on the first CSI to degrade when used in the FDD system. Specifically, this may manifest as a large deviation between the channel state information recovered based on the first decoder and the actual channel state information input to the first encoder, thereby affecting communication quality. However, the first CSI determined by the second device in implementation mode 1.2 using a channel transformation machine learning model to transform the first channel estimation information, or the first CSI determined by the second device in implementation mode 1.3 using a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information, is conducive to enabling the first encoder or first decoder determined by the second device based on the first CSI to be applicable to the channel data distribution of the FDD system, thereby reducing the performance degradation of the first encoder or first decoder when used in the FDD system, reducing the deviation between the channel state information recovered based on the first decoder and the actual channel state information input to the first encoder, and improving communication quality.

[0159] In addition, optionally, in implementation 1.2 and implementation 1.3, the channel transformation machine learning model used by the second device may be pre-stored in the second device, or may be downloaded from the channel transformation machine learning model storage location when the second device is actually deployed and used, without restriction.

[0160] S105. The first device performs channel estimation based on the second reference signal to determine a second CSI.

[0161] The following describes an optional implementation of the operation in which the second device performs channel estimation based on the first reference signal to determine the first CSI, as described in the following optional implementations 2.1 to 2.3.

[0162] Implementation 2.1, the first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the second CSI is the second channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the second channel estimation information.

[0163] It is understandable that the second channel estimation information can characterize the characteristics of the channel corresponding to the second reference signal. In addition, this embodiment 2.1 can be applied to the scenario where the second device determines the first CSI using embodiment 1.1.

[0164] Optionally, the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information may be: the eigenvalues ​​or eigenvectors of the channel matrix obtained by the second device processing the second channel estimation information. Exemplarily, the second device may perform eigendecomposition on the second channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information. The eigendecomposition may be, for example, SVD.

[0165] In addition, optionally, in addition to the manner in which the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the second channel estimation information, the first CSI may also be other characteristic information of the channel matrix corresponding to the second channel estimation information, without limitation.

[0166] Implementation 2.2, the first device performs channel estimation based on the second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device uses a channel transformation machine learning model to transform the second channel estimation information to obtain fourth channel estimation information; the second CSI is the fourth channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information. For a specific explanation of the second channel estimation information, please refer to the relevant explanation in Implementation 2.1 and will not be repeated here. In addition, this Implementation 2.2 can be applied to the scenario where the second device uses Implementation 1.2 to determine the first CSI.

[0167] Optionally, the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the fourth channel estimation information may be: the eigenvalues ​​or eigenvectors of the channel matrix obtained by the second device processing the fourth channel estimation information. Exemplarily, the second device may perform eigendecomposition on the fourth channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the fourth channel estimation information. The eigendecomposition may be, for example, SVD.

[0168] In addition, optionally, in addition to the manner in which the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information, the first CSI may also be other characteristic information of the channel matrix corresponding to the fourth channel estimation information, without limitation.

[0169] Implementation 2.3, the first device performs channel estimation based on the second reference signal from the second device to determine the second CSI, including: the first device performs channel estimation based on the second reference signal to obtain second channel estimation information; the first device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the second CSI. For a specific explanation of the second channel estimation information and the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information, please refer to the relevant explanation in Implementation 2.1, which will not be repeated here. In addition, this Implementation 2.3 can be applied to the scenario where the second device determines the first CSI using Implementation 1.3. In addition, optionally, in addition to transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the first CSI, other characteristic information of the channel matrix corresponding to the second channel estimation information can also be transformed to obtain the first CSI, without limitation.

[0170] In addition, in implementation 2.2 and implementation 2.3, the channel transformation machine learning model is related to the channel data distribution. In implementation 2.2, the first device uses the channel transformation machine learning model to transform the second channel estimation information. This operation, or in implementation 2.3, the first device uses the channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information. This operation is beneficial to make the channel data distribution in the actual deployment scenario of the first encoder and the first decoder different from the channel data distribution corresponding to the first reference signal and the second reference signal. The second CSI can characterize the channel characteristics in the actual deployment scenario, thereby helping to reduce the performance degradation of the first encoder or the first decoder determined by the second device based on the first channel characteristic information (the first channel characteristic information is obtained by the first device using the first encoder to compress the second CSI) in the actual deployment scenario, that is, reducing the deviation between the channel state information recovered by the first decoder in the actual deployment scenario and the actual channel state information input to the first encoder, which is beneficial to improving communication quality. For the specific description of the channel transformation machine learning model, please refer to the aforementioned related description and will not be repeated here.

[0171] In addition, for the scenario in which the second device adopts implementation 1.2 to determine the first CSI and the first device adopts implementation 2.2 to determine the second CSI, or for the scenario in which the second device adopts implementation 1.3 to determine the first CSI and the first device adopts implementation 2.3 to determine the second CSI, the channel transformation machine learning model adopted by the first device is the same as the channel transformation machine learning model adopted by the second device.

[0172] Optionally, in implementation 2.2 and implementation 2.3, the channel transformation machine learning model used by the first device may be pre-stored in the first device, or may be downloaded from the channel transformation machine learning model storage location when the first device is actually deployed and used, without restriction.

[0173] S106. The first device compresses the second CSI using the first encoder to obtain first channel characteristic information.

[0174] It is understandable that, since the first device determines the first encoder in step S101, in step S106, the first device may compress the second CSI using the first encoder determined in step S101 to obtain the first channel characteristic information.

[0175] S107. The first device sends first channel characteristic information to the second device; correspondingly, the second device receives the first channel characteristic information from the first device.

[0176] S108. The second device determines a first encoder or a first decoder based on the first CSI and the first channel characteristic information.

[0177] It is understandable that, since the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal, the channel characteristics corresponding to the first reference signal are substantially the same as the channel characteristics corresponding to the second reference channel. It can be seen that the first CSI determined by the second device based on the first reference signal for channel estimation is substantially the same as the second CSI determined by the first device based on the second reference signal for channel estimation. Then, the channel characteristic information obtained by the second device using the first encoder to compress the first CSI is substantially the same as the first channel characteristic information obtained by the first device using the first encoder to compress the second CSI. Therefore, the second device can determine the first encoder based on the first CSI and the first channel characteristic information. In addition, since the first CSI determined by the second device based on the first reference signal for channel estimation is substantially the same as the second CSI determined by the first device based on the second reference signal for channel estimation, then the CSI obtained by the second device using the first decoder to decompress the first channel characteristic information is substantially the same as the first CSI. Therefore, the second device can determine the first decoder based on the first channel characteristic information and the first CSI.

[0178] In an optional embodiment, the second device determines the first encoder based on the first CSI and the first channel characteristic information, which may include: step a, the second device constructing a training data set based on the first CSI and the first channel characteristic information, and training the encoder based on the training data set. The first device and the second device may repeatedly perform steps S102 to S107, and step a, multiple times so that the second device determines the first encoder, and the first encoder determined by the second device satisfies the first convergence condition. Optional methods for the first convergence condition are described below, as described in optional embodiments 3.1 to 3.3 below.

[0179] In embodiment 3.1, the first convergence condition may be: the number of repetitions of steps S102 to S107 and step a reaches a first value. The first value may be predefined or determined by negotiation between the first device and the second device, without limitation.

[0180] It is understandable that the first device and the second device can repeat steps S102 to S107 and step a multiple times until the number of times steps S102 to S107 and step a are repeated reaches a first value. Then, the first device and the second device can stop repeating steps S102 to S107 and step a, and the second device can determine the trained encoder obtained in the most recent execution of step a as the first encoder.

[0181] In implementation 3.2, the first convergence condition may be: the difference between the second channel characteristic information and the third channel characteristic information is less than the second value. The second channel characteristic information is obtained by the second device compressing the third CSI using the first encoder determined in step S108, and the third channel characteristic information is obtained by the first device compressing the fourth CSI using the first encoder determined in step S101. The third CSI is determined by the second device performing channel estimation based on the third reference signal from the first device, and the fourth CSI is determined by the first device performing channel estimation based on the fourth reference signal from the second device; the channel corresponding to the third reference signal and the channel corresponding to the fourth reference signal are reciprocal. The second value may be predefined or determined by negotiation between the first device and the second device, without limitation.

[0182] Exemplarily, after the first device and the second device repeatedly perform steps S102 to S107 and step a for a period of time, the second device may send reference signal #1 to the first device, and the first device may send reference signal #2 to the second device, where the channel corresponding to reference signal #1 and the channel corresponding to reference signal #2 are reciprocal. The first device may perform channel estimation based on reference signal #1 to determine CSI #1, compress CSI #1 using the first encoder determined in step S101 to obtain channel characteristic information #1, and send channel characteristic information #1 to the second device. The second device may perform channel estimation based on reference signal #2 to determine CSI #2, and compress CSI #2 using the encoder trained in the most recent execution of step a to obtain channel characteristic information #2.

[0183] If the difference between channel characteristic information #2 and channel characteristic information #1 is less than the second value, the second device may determine the encoder trained in the most recently executed step a as the first encoder, and the first and second devices cease repeatedly executing steps S102 to S107 and step a. In this case, reference signal #1 is the aforementioned fourth reference signal, reference signal #2 is the aforementioned third reference signal, CSI #1 is the aforementioned fourth CSI, CSI #2 is the aforementioned third CSI, channel characteristic information #1 is the aforementioned third channel characteristic information, and channel characteristic information #2 is the aforementioned second channel characteristic information.

[0184] If the difference between channel characteristic information #2 and channel characteristic information #1 is greater than or equal to the second value, the first device and the second device can continue to repeat steps S102 to S107 and step a until the difference between the second channel characteristic information and the third channel characteristic information is less than the second value. The second device stops repeating steps S102 to S107 and step a, and determines the trained encoder obtained in the most recent execution of step a as the first encoder.

[0185] In implementation 3.3, the first convergence condition may be: the difference between the fourth CSI and the fifth CSI is less than the third value. The fourth CSI is determined by the first device based on the fourth reference signal from the second device for channel estimation, and the fifth CSI is obtained by the first device using the first decoder determined in step S101 to decompress the second channel characteristic information from the second device. The second channel characteristic information is obtained by the second device using the first encoder determined in step S108 to compress the third CSI, and the third CSI is determined by the second device based on the third reference signal from the first device for channel estimation. The channel corresponding to the third reference signal and the channel corresponding to the fourth reference signal are reciprocal. The third value may be predefined or determined by negotiation between the first device and the second device, without limitation.

[0186] The difference between the fifth CSI and the fourth CSI is less than the third value. Therefore, it can be considered that the difference between the channel state information recovered by the first device using the first decoder determined in step S101 to decompress the second channel characteristic information from the second device is small, and the actual channel state information determined by the first device through channel estimation based on the reference signal from the second device. Therefore, the first encoder determined by the second device in step S108 and the first decoder determined by the first device in step S101 are compatible. This can also be understood as follows: the first encoder determined by the second device in step S108 is the same as the first encoder determined by the first device in step S101.

[0187] Exemplarily, after the first device and the second device repeatedly perform steps S102 to S107 and step a for a period of time, the second device may send reference signal #1 to the first device, and the first device may send reference signal #2 to the second device, where the channel corresponding to reference signal #1 and the channel corresponding to reference signal #2 are reciprocal. The second device may perform channel estimation based on reference signal #2 to determine CSI #2, compress CSI #2 using the encoder trained in the most recent execution of step a to obtain channel characteristic information #2, and send channel characteristic information #2 to the first device. The first device may perform channel estimation based on reference signal #1 to determine CSI #1, and decompress channel characteristic information #2 using the first decoder determined in step S101 to obtain CSI #3.

[0188] If the difference between CSI #3 and CSI #1 is less than the third value, the first device may send an instruction to the second device to instruct the second device to stop training the encoder. Upon or after receiving the instruction, the second device may determine the encoder trained in the most recently executed step a as the first encoder, and the first and second devices may stop repeatedly executing steps S102 to S107 and step a. In this case, reference signal #1 is the aforementioned fourth reference signal, reference signal #2 is the aforementioned third reference signal, CSI #1 is the aforementioned fourth CSI, CSI #2 is the aforementioned third CSI, CSI #3 is the aforementioned fifth CSI, and channel characteristic information #2 is the aforementioned second channel characteristic information.

[0189] If the difference between CSI#3 and CSI#1 is greater than or equal to the third value, the first device may send an instruction message to the second device to instruct the second device to continue training the encoder. The first device and the second device may continue to repeat steps S102 to S107 and step a until the difference between the fifth CSI and the fourth CSI is less than the third value. The second device stops repeating steps S102 to S107 and step a and determines the trained encoder obtained in the most recently executed step a as the first encoder.

[0190] In an optional embodiment, the second device determines the first decoder based on the first CSI and the first channel characteristic information, which may include: step b, the second device constructing a training data set based on the first CSI and the first channel characteristic information, and training the decoder based on the training data set. The first device and the second device may repeatedly perform steps S102 to S107 and step b multiple times so that the second device determines the first decoder, and the first decoder determined by the second device satisfies the second convergence condition. Optional methods for the second convergence condition are described below, as described in the following optional embodiments 4.1 and 4.2.

[0191] In embodiment 4.1, the first convergence condition may be: the number of repetitions of steps S102 to S107 and step b reaches a first value. The first value may be predefined or determined by negotiation between the first device and the second device, without limitation.

[0192] It is understandable that the first device and the second device can repeatedly execute steps S102 to S107 and step b multiple times until the number of times steps S102 to S107 and step b are repeated reaches a first value. Then, the first device and the second device can stop repeating steps S102 to S107 and step b, and the second device can determine the trained decoder obtained in the most recent execution of step b as the first decoder.

[0193] In implementation 4.2, the first convergence condition may be: the difference between the third CSI and the sixth CSI is less than the fourth value. The third CSI is determined by the second device through channel estimation based on the third reference signal from the first device, and the sixth CSI is obtained by the second device decompressing the third channel characteristic information from the first device using the first decoder determined in step S108. The third channel characteristic information is obtained by the first device compressing the fourth CSI using the first encoder determined in step S101, and the fourth CSI is determined by the first device through channel estimation based on the fourth reference signal from the second device. The channel corresponding to the third reference signal and the channel corresponding to the fourth reference signal are reciprocal. In addition, the fourth value may be predefined or determined by negotiation between the first device and the second device, without limitation.

[0194] The difference between the third CSI and the sixth CSI is less than the fourth value. Therefore, it can be considered that the difference between the channel state information recovered by the second device using the first decoder determined in step S108 to decompress the third channel characteristic information from the first device is small, and the actual channel state information determined by the second device through channel estimation based on the reference signal from the first device. Therefore, the first decoder determined by the second device in step S108 is compatible with the first encoder determined by the first device in step S101. This can also be understood as follows: the first decoder determined by the second device in step S108 is the same as the first decoder determined by the first device in step S101.

[0195] Exemplarily, after the first device and the second device repeatedly perform steps S102 to S107 and step b for a period of time, the second device may send reference signal #1 to the first device, and the first device may send reference signal #2 to the second device, where the channel corresponding to reference signal #1 and the channel corresponding to reference signal #2 are reciprocal. The first device may perform channel estimation based on reference signal #1 to determine CSI #1, and compress CSI #1 using the first encoder determined in step S101 to obtain channel characteristic information #1, and send channel characteristic information #1 to the second device. The second device may perform channel estimation based on reference signal #2 to determine CSI #2, and decompress channel characteristic information #1 using the decoder trained in the most recent step a to obtain CSI #4.

[0196] If the difference between CSI #4 and CSI #2 is less than the fourth value, the second device may determine the decoder trained in the most recently executed step b as the first decoder, and the first and second devices cease repeatedly executing steps S102 to S107 and step b. In this case, reference signal #1 is the aforementioned fourth reference signal, reference signal #2 is the aforementioned third reference signal, CSI #1 is the aforementioned fourth CSI, CSI #2 is the aforementioned third CSI, CSI #4 is the aforementioned sixth CSI, and channel characteristic information #1 is the aforementioned third channel characteristic information.

[0197] If the difference between CSI#4 and CSI#2 is greater than or equal to the fourth value, the first device and the second device may continue to repeatedly perform steps S102 to S107 and step b until the difference between the third CSI and the sixth CSI is less than the fourth value. The second device stops repeating steps S102 to S107 and step b and determines the trained decoder obtained in the most recently executed step b as the first decoder.

[0198] In an optional embodiment, for the scenario in which the second device determines in step S108 that the first encoder is the first encoder, the method may also include: the second device uses a parameter sharing machine learning model to transform the first encoder to obtain a second encoder; the first device uses a parameter sharing machine learning model to transform the first decoder to obtain a second decoder.

[0199] The parameter sharing machine learning model is related to the communication system parameter configuration, which may include one or more of the following: the number of antenna elements, the number of antenna ports, the number of spatial streams, etc. Transforming the first encoder and the first decoder using the parameter sharing machine learning model facilitates adapting the transformed second encoder and second decoder to the actual deployment scenario when the communication system parameter configuration of the first encoder and the first decoder differs from the communication system parameter configuration in the scenario where the second device determines the first encoder, thereby improving communication quality.

[0200] Exemplarily, the second device determines a first encoder in a TDD system and deploys the first encoder and first decoder in an FDD system. The parameter configuration of the FDD system may differ from the parameter configuration of the TDD system. The second device may transform the first encoder using a parameter-sharing machine learning model to obtain a second encoder; the first device may transform the first decoder using a parameter-sharing machine learning model to obtain a second decoder, so that the second encoder and second decoder are suitable for the parameter configuration of the FDD system. The first and second devices implement compressed feedback of channel state information using the second encoder and second decoder in the FDD system.

[0201] Optionally, the parameter sharing machine learning model can be a graph model. For a detailed description of the graph model, please refer to the description in the aforementioned related concepts and will not be repeated here. For example, in conjunction with Figure 10, Figure 10 shows the second encoder and the second decoder obtained based on the graph model to implement the compressed feedback of the channel state information. The input of the second encoder is based on the truncated channel matrix The constructed graph model, specifically the adjacency matrix A and node feature matrix of the graph model Input the second encoder. First, the graph data (including the adjacency matrix A and the node feature matrix X) passes through two layers of graph sampling and aggregation (GraphSAGE) graph convolution layers, where the calculation expression of the k-th layer convolution operation can be shown as formula (8).

[0202] In formula (8), is the embedding result of point i in the kth layer, is the parameter to be learned; N(i) represents the set of neighboring points of point i. The purpose of the above graph convolution operation is to update its own information with the information of the neighboring angles to obtain the embedding result. The dimension of the embedding result is smaller than the dimension of the original feature 2N c (The 2N c See the explanation of the graph model in the previous related concepts), achieving preliminary channel state information compression. In addition, the parameters {W1, W2} involved in the above operation apply to all nodes, so this part of the operation is independent of the size of the input graph data, that is, independent of the number of antennas.

[0203] After two layers of graph convolution, the graph data is pooled. The pooling operation is essentially a downsampling process, that is, only some of the points in the graph and the connection relationship between the points are retained to obtain a subgraph. In the embodiment of the present application, the self-attention mechanism pooling operation can be used to select the nodes and topological structures to be retained. This method fully considers the characteristics of the nodes and the topological structure of the graph, and can also learn the hierarchical representation of the graph in an end-to-end manner with relatively few parameters. The key to the self-attention mechanism pooling is to use GNN to calculate the self-attention score based on the embedding result, retain the top s points with the highest score, and then splice the embedding results of these s points for subsequent compression. The calculation expression of the self-attention mechanism pooling operation can be shown as follows: formula (9) to formula (12). score = GNN (H (2) ,A) (9) idx=top s (score) (10)

[0204] The parameter s in the pooling operation of the self-attention mechanism is independent of the number of antennas, so the dimension of the vector y in formula (12) is fixed. The above pooling operation is independent of the size of the input graph data, that is, independent of the number of antennas.

[0205] After the pooling operation, the vector y can be input into the fully connected layer and compressed to the given dimension to obtain the vector z. The compressed vector z is then transmitted to the first device together with the position information vector idx. Corresponding to the compression part of the second device, the recovery part of the first device first includes a fully connected layer. The fully connected layer is to restore the compressed vector z to s×2N c The dimension corresponds to the initial recovered features of the s points in the position information vector idx. For points not included in idx, their features can be filled with zero vectors. Using the fully connected layer recovery and zero padding, the graph data is initially recovered. Then, the initially recovered graph data is passed through two layers of GraphSage graph convolution, and a direct connection channel is added in the middle using the residual network. Repeat the graph convolution and direct connection operation twice to obtain the final recovered and reconstructed graph data and the corresponding channel matrix

[0206] In addition, in this example, a graph model is constructed for each receiving antenna in the second device. Therefore, the graph convolution and graph pooling process in the CSI compression process of the second device is independent of the number of transmitting antennas N of the first device. t For the graph model corresponding to each receiving antenna in the second device, the above operations may be repeated to achieve compression, feedback, and recovery of the CSI.

[0207] In another optional embodiment, for the scenario where the second device determines in step S108 that the first decoder is the first decoder, the method may further include: the second device transforming the first decoder using a parameter-sharing machine learning model to obtain a second decoder; and the first device transforming the first encoder using a parameter-sharing machine learning model to obtain a second encoder. This embodiment is similar to the aforementioned transformation using a parameter-sharing machine learning model for the scenario where the second device determines in step S108 that the first encoder is the first encoder, and reference may be made to the aforementioned related explanations, which will not be repeated here.

[0208] In an optional embodiment, the method may further include: implementing compressed feedback of channel state information between the first device and the second device based on the first encoder and the first decoder. Alternatively, for a scenario in which a parameter-sharing machine learning model is also used for transformation, the method may further include: implementing compressed feedback of channel state information between the first device and the second device based on the second encoder and the second decoder.

[0209] In another optional embodiment, the second device determines that the first encoder or the first decoder can be deployed in other devices different from the first device and the second device. For example, after the second device determines the first encoder, the first encoder can be deployed in the third device and the first decoder can be deployed in the fourth device. The third device and the fourth device can implement compressed feedback of the channel state information based on the first encoder and the first decoder. Optionally, the third device can also use a parameter-sharing machine learning model to transform the first encoder to obtain a second decoder; the fourth device can also use a parameter-sharing machine learning model to transform the first decoder to obtain a second decoder. The second encoder and the second decoder are used between the third device and the fourth device to implement compressed feedback of the channel state information.

[0210] In summary, in this information transmission method, the first device determines a first encoder and a first decoder. The first device performs channel estimation based on a second reference signal from the second device to determine the second CSI, compresses the second CSI using the first encoder to obtain first channel characteristic information, and sends the first channel characteristic information to the second device. The second device performs channel estimation based on the first reference signal from the first device to determine the first CSI, and the second device determines the first encoder or the first decoder based on the first CSI and the first channel characteristic information. The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0211] This information transmission method utilizes the reciprocity between the channel corresponding to the first reference signal and the channel corresponding to the second reference signal to enable the second device to determine the first encoder that is compatible with the first decoder in the first device, or to enable the second device to determine the first decoder that is compatible with the first encoder in the first device. In this information transmission method, the first device does not need to send the second CSI to the second device, which can reduce communication overhead. In addition, in this information transmission method, the first device does not need to directly send the model parameters of the encoder or decoder obtained through joint training to the second device, which ensures the privatization of the encoder or decoder in the first device, which is beneficial to improving the privacy security of the encoder or decoder in the first device when the first device and the second device are from different manufacturers.

[0212] The information transmission method provided in the embodiment of the present application is exemplarily described below in conjunction with specific scenarios.

[0213] The information transmission method described in the embodiments of this application can be applied to communication scenarios between network devices and terminal devices. The terminal device can compress channel state information using an encoder and feed the compressed channel state information back to the network device; the network device can decompress the information using a decoder to recover the channel state information. The information transmission method in this scenario can be described in Examples 1 and 2 below.

[0214] Example 1: The first device is a network device or a chip in a network device, and the second device is a terminal device or a chip in a terminal device. In this case, the first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; the second device determines the first encoder based on the first CSI and the first channel characteristic information. Below, taking the first device as a network device and the second device as a terminal device as an example, in conjunction with Figure 11, this exemplary information transmission method includes the following steps S201 to S208:

[0215] S201: A network device determines a first encoder and a first decoder.

[0216] S202. The network device sends a first reference signal to the terminal device; correspondingly, the terminal device receives the first reference signal from the network device.

[0217] S203: The terminal device sends a second reference signal to the network device; in response, the network device receives the second reference signal from the terminal device, wherein the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0218] In an optional embodiment, the method further includes: the network device sends one or more of the following to the terminal device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0219] Exemplarily, the frequency domain resource position of the configured first reference signal can be the same as the frequency domain resource position of the second reference signal, and the interval between the time domain resource position of the first reference signal and the time domain resource position of the second reference signal can be small, so that the channel corresponding to the first reference signal and the information corresponding to the second reference signal are reciprocal.

[0220] Optionally, the first information is used to indicate the time-frequency resource position of the first reference signal, which may be: the first information indicates the time-frequency resource position for sending the first reference signal, or the first information indicates the time-frequency resource position for receiving the first reference signal. The second information is used to indicate the time-frequency resource position of the second reference signal, which may be: the second information indicates the time-frequency resource position for sending the second reference signal, or the second information indicates the time-frequency resource position for receiving the second reference signal. The third information is used to indicate the time-frequency resource position of the first channel characteristic information, which may be: the third information indicates the time-frequency resource position for sending the first channel characteristic information, or the third information indicates the time-frequency resource position for receiving the first channel characteristic information.

[0221] Optionally, the first information, the second information, and the third information may be carried in the same signaling for transmission, or any two or more of the first information, the second information, and the third information may be carried in different signaling for transmission, without limitation.

[0222] S204. The terminal device performs channel estimation based on the first reference signal and determines a first CSI.

[0223] S205: The network device performs channel estimation based on the second reference signal to determine a second CSI.

[0224] S206. The network device compresses the second CSI using the first encoder to obtain first channel characteristic information.

[0225] S207. The network device sends first channel characteristic information to the terminal device; correspondingly, the terminal device receives the first channel characteristic information from the network device.

[0226] S208. The terminal device determines a first encoder based on the first CSI and the first channel characteristic information.

[0227] The present embodiment does not limit the order of steps S202 and S203, nor does it limit the order of steps S204 and S205. For details on steps S201-S208, please refer to the relevant description in the aforementioned information transmission method, which will not be repeated here.

[0228] In addition, in an optional embodiment, where the first device is a network device or a chip in the network device, the RU, CU, and DU may collaborate to perform operations of the first device. For example, the CU or DU may determine a first encoder and a first decoder, the RU may send a first reference signal to the second device and receive a second reference signal from the second device, the DU may perform channel estimation based on the second reference signal to determine a second CSI, the CU or DU may use the first encoder to compress the second CSI to obtain first channel characteristic information, and the RU may send the first channel characteristic information to the second device.

[0229] Example 2: The first device is a terminal device or a chip in a terminal device, and the second device is a network device or a chip in a network device. In this case, the first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; the second device determines the first decoder based on the first CSI and the first channel characteristic information. Below, taking the first device as a terminal device and the second device as a network device as an example, in conjunction with Figure 12, this exemplary information transmission method includes the following steps S301 to S308:

[0230] S301: The terminal device determines a first encoder and a first decoder.

[0231] S302. The terminal device sends a first reference signal to the network device; correspondingly, the network device receives the first reference signal from the terminal device.

[0232] S303: The network device sends a second reference signal to the terminal device; in response, the terminal device receives the second reference signal from the network device, wherein the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0233] In an optional embodiment, the method further includes: the network device sending one or more of the following to the terminal device: first information, second information, and third information; wherein the first information is used to indicate the time-frequency resource location of the first reference signal; the second information is used to indicate the time-frequency resource location of the second reference signal; and the third information is used to indicate the time-frequency resource location of the first channel characteristic information. For a detailed description of the first information, the second information, and the third information, please refer to the relevant description of the information transmission method shown in Figure 11 and will not be repeated here.

[0234] S304: The network device performs channel estimation based on the first reference signal to determine a first CSI.

[0235] S305. The terminal device performs channel estimation based on the second reference signal and determines a second CSI.

[0236] In an optional embodiment, the method further includes: the network device sends one or more of the following to the terminal device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0237] S306. The terminal device uses the first encoder to compress the second CSI to obtain first channel characteristic information.

[0238] S307. The terminal device sends the first channel characteristic information to the network device; correspondingly, the network device receives the first channel characteristic information from the terminal device.

[0239] S308. The network device determines a first decoder based on the first CSI and the first channel characteristic information.

[0240] The present embodiment does not limit the order of steps S302 and S303, nor does it limit the order of steps S304 and S305. For details on steps S301-S308, please refer to the relevant description in the aforementioned information transmission method, which will not be repeated here.

[0241] In addition, in an optional embodiment, when the second device is a network device or a chip in the network device, the RU, CU, and DU may collaborate to perform operations of the second device. For example, the RU may receive a first reference signal from the first device and send a second reference signal to the first device, the DU may perform channel estimation based on the first reference signal to determine a first CSI, the RU may receive first channel characteristic information from the first device, and the CU or DU may determine a first decoder based on the first CSI and the first channel characteristic information.

[0242] The information transmission method described in the embodiments of the present application can also be applied to sidelink communication scenarios. For example, a second terminal device compresses channel state information using an encoder and feeds the compressed channel state information back to a first terminal device; the first terminal device decompresses the information using a decoder to recover the channel state information. In this scenario, both the first reference signal and the second reference signal are sidelink reference signals. The information transmission method in this scenario can be described in Examples 3 and 4 below.

[0243] Example 3: The first device is a first terminal device or a chip in the first terminal device, and the second device is a second terminal device or a chip in the second terminal device. In this case, the second device determines that it is the first encoder based on the first CSI and the first channel characteristic information. Below, taking the first device being the first terminal device and the second device being the second terminal device as an example, in conjunction with FIG13 , this exemplary information transmission method includes the following steps S401 to S408:

[0244] S401: A first terminal device determines a first encoder and a first decoder.

[0245] S402. The first terminal device sends a first reference signal to the second terminal device; correspondingly, the second terminal device receives the first reference signal from the first terminal device.

[0246] S403: The second terminal device sends a second reference signal to the first terminal device; in response, the first terminal device receives the second reference signal from the second terminal device, wherein the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0247] In an optional embodiment, the method also includes: the first terminal device sends one or more of the following to the second terminal device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0248] In another optional embodiment, the method also includes: the network device sends one or more of the following to the first terminal device and the second terminal device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0249] For detailed description of the first information, the second information, and the third information, please refer to the relevant description in the information transmission method shown in FIG11 , which will not be repeated here.

[0250] S404. The second terminal device performs channel estimation based on the first reference signal and determines a first CSI.

[0251] S405. The first terminal device performs channel estimation based on the second reference signal and determines a second CSI.

[0252] S406. The first terminal device uses the first encoder to compress the second CSI to obtain first channel characteristic information.

[0253] S407. The first terminal device sends first channel characteristic information to the second terminal device; correspondingly, the second terminal device receives the first channel characteristic information from the first terminal device.

[0254] S408. The second terminal device determines the first encoder based on the first CSI and the first channel characteristic information.

[0255] The present embodiment does not restrict the order of steps S402 and S403, nor does it restrict the order of steps S404 and S405. For details on steps S401-S408, please refer to the relevant description in the aforementioned information transmission method, which will not be repeated here.

[0256] Example 4: The first device is a second terminal device or a chip in the second terminal device, and the second device is the first terminal device or a chip in the first terminal device. In this case, the second device determines that it is the first decoder based on the first CSI and the first channel characteristic information. Below, taking the first device being the second terminal device and the second device being the first terminal device as an example, in conjunction with FIG14 , this exemplary information transmission method includes the following steps S501 to S508:

[0257] S501: The second terminal device determines a first encoder and a first decoder.

[0258] S502. The second terminal device sends a first reference signal to the first terminal device; correspondingly, the first terminal device receives the first reference signal from the second terminal device.

[0259] S503: The first terminal device sends a second reference signal to the second terminal device; in response, the second terminal device receives the second reference signal from the first terminal device, wherein the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0260] In an optional embodiment, the method also includes: the first terminal device sends one or more of the following to the second terminal device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0261] In another optional embodiment, the method also includes: the network device sends one or more of the following to the first terminal device and the second terminal device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0262] For detailed description of the first information, the second information, and the third information, please refer to the relevant description in the information transmission method shown in FIG11 , which will not be repeated here.

[0263] S504. The first terminal device performs channel estimation based on the first reference signal and determines a first CSI.

[0264] S505. The second terminal device performs channel estimation based on the second reference signal and determines a second CSI.

[0265] S506. The second terminal device uses the first encoder to compress the second CSI to obtain first channel characteristic information.

[0266] S507. The second terminal device sends the first channel characteristic information to the first terminal device; correspondingly, the first terminal device receives the first channel characteristic information from the second terminal device.

[0267] S508. The first terminal device determines a first decoder based on the first CSI and the first channel characteristic information.

[0268] The present embodiment does not restrict the order of steps S502 and S503, nor does it restrict the order of steps S504 and S505. For details on steps S501-S508, please refer to the relevant description in the aforementioned information transmission method, which will not be repeated here.

[0269] Below, taking the information transmission method in Example 1 where the first device is a network device and the second device is a terminal device as an example, the implementation method of using the channel transformation machine learning model for transformation is explained. The situations of Examples 2 to 4 are similar and will not be repeated.

[0270] In an optional embodiment, as shown in Figure 15, step S204 in the information transmission method shown in Figure 11 may include S204a and S204b. Step S205 in the information transmission method shown in Figure 11 may include S205a and S205b.

[0271] S204a. The terminal device performs channel estimation based on the first reference signal to obtain first channel estimation information.

[0272] S204b. The terminal device uses a channel transformation machine learning model to transform the first channel estimation information to obtain third channel estimation information; the first CSI is the third channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information.

[0273] S205a: The network device performs channel estimation based on the second reference signal to obtain second channel estimation information.

[0274] S205b. The network device uses a channel transformation machine learning model to transform the second channel estimation information to obtain fourth channel estimation information; the second CSI is the fourth channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information.

[0275] For detailed description of the above steps S204a, S204b, S205a and S205b, please refer to the relevant description in the aforementioned information transmission method, which will not be repeated here.

[0276] In another optional embodiment, as shown in Figure 16, step S204 in the information transmission method shown in Figure 11 may include S204a and S204c. Step S205 in the information transmission method shown in Figure 11 may include S205a and S205c.

[0277] S204a. The terminal device performs channel estimation based on the first reference signal to obtain first channel estimation information.

[0278] S204c. The terminal device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI.

[0279] S205a: The network device performs channel estimation based on the second reference signal to obtain second channel estimation information.

[0280] S205c. The network device uses a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain a second CSI.

[0281] For detailed description of the above steps S204a, S204c, S205a and S205c, please refer to the relevant description in the aforementioned information transmission method, which will not be repeated here.

[0282] Below, taking the information transmission method in Example 1 where the first device is a network device and the second device is a terminal device as an example, the implementation method of using a parameter sharing machine learning model for transformation is explained. The situations in Examples 2 to 4 are similar and will not be repeated.

[0283] In an optional implementation, as shown in FIG17 , after step S208 in the information transmission method shown in FIG11 , step S209 and step S210 as shown in the dotted box in the figure may also be included.

[0284] S209. The terminal device uses a parameter-sharing machine learning model to transform the first encoder to obtain a second encoder.

[0285] S210. The network device uses a parameter-sharing machine learning model to transform the first decoder to obtain a second decoder.

[0286] Then, the second encoder and the second decoder can be used between the network device and the terminal device to implement compressed feedback of the channel state information.

[0287] The embodiment of the present application does not limit the order of the above steps S209 and S210. In addition, the specific description of steps S209 and S210 can be found in the relevant description of the above information transmission method, which will not be repeated here.

[0288] To implement the various functions of the methods provided in the embodiments of the present application, the first device or the second device may include hardware structures and / or software modules, and implement the aforementioned functions in the form of hardware structures, software modules, or a combination of hardware structures and software modules. Whether a particular one of the aforementioned functions is implemented in the form of hardware structures, software modules, or a combination of hardware structures and software modules depends on the specific application and design constraints of the technical solution.

[0289] As shown in Figure 18, an embodiment of the present application provides a communication device 1800. The communication device 1800 can be a first device or a second device, or can also be a component of the first device (for example, an integrated circuit, a chip, etc.), or can also be a component of the second device (for example, an integrated circuit, a chip, etc.). The communication device 1800 can also be other communication units for implementing the method in the method embodiment of the present application. The communication device 1800 may include a processing unit 1801. Optionally, the communication device 1800 may also include a communication unit 1802, and the processing unit 1801 is used to control the communication unit 1802 to send and receive data / signaling. The communication unit 1802 may also be referred to as a transceiver unit. Optionally, the communication unit 1802 may include a sending unit and a receiving unit. The sending unit can be used to send data / signaling, and the receiving unit can be used to receive data / signaling. Optionally, the communication device 1800 may further include a storage unit 1803 , which may be used to store information and / or data and / or instructions, etc. The storage unit 1803 may interact with the processing unit 1801 and may also interact with the communication unit 1802 .

[0290] In one possible design, for a case where the communication apparatus 1800 is used to implement the function of the second device in the above method embodiment:

[0291] The processing unit 1801 is configured to perform channel estimation based on a first reference signal from a first device, and determine a first CSI.

[0292] The communication unit 1802 is configured to send a second reference signal to the first device, where the second reference signal is used by the first device to perform channel estimation and determine a second CSI.

[0293] The communication unit 1802 is further configured to receive first channel characteristic information from the first device, where the first channel characteristic information is obtained by the first device compressing the second CSI using a first encoder.

[0294] The processing unit 1801 is further configured to determine a first encoder or a first decoder based on the first CSI and the first channel characteristic information.

[0295] The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0296] In an optional embodiment, the processing unit 1801 performs channel estimation based on a first reference signal from the first device to determine the first CSI, and is specifically used to: perform channel estimation based on the first reference signal to obtain first channel estimation information; the first CSI is the first channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the first channel estimation information.

[0297] Optionally, the processing unit 1801 performs channel estimation based on the first reference signal from the first device to determine the first CSI, and is specifically used to: perform channel estimation based on the first reference signal to obtain first channel estimation information; perform eigendecomposition, such as SVD, on the first channel estimation information to obtain the first CSI.

[0298] In an optional embodiment, the second CSI is determined by the first device through channel estimation based on the second reference signal, including: the second channel estimation information is obtained by the first device through channel estimation based on the second reference signal; the second CSI is the second channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the second channel estimation information.

[0299] In an optional embodiment, the processing unit 1801 performs channel estimation based on a first reference signal from the first device to determine a first CSI, and is specifically used to: perform channel estimation based on the first reference signal to obtain first channel estimation information; transform the first channel estimation information using a channel transformation machine learning model to obtain third channel estimation information; the first CSI is the third channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information.

[0300] Optionally, the processing unit 1801 performs channel estimation based on the first reference signal from the first device to determine the first CSI, and is specifically used to: perform channel estimation based on the first reference signal to obtain first channel estimation information; use a channel transformation machine learning model to transform the first channel estimation information to obtain third channel estimation information; and perform feature decomposition, such as SVD, on the third channel estimation information to obtain the first CSI.

[0301] In an optional embodiment, the second CSI is determined by the first device performing channel estimation based on the second reference signal, including: the second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; the fourth channel estimation information is obtained by the first device transforming the second channel estimation information using a channel transformation machine learning model; the second CSI is the fourth channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information.

[0302] In an optional embodiment, the processing unit 1801 performs channel estimation based on a first reference signal from the first device to determine the first CSI, and is specifically used to: perform channel estimation based on the first reference signal to obtain first channel estimation information; and use a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI.

[0303] Optionally, the processing unit 1801 performs channel estimation based on the first reference signal from the first device to determine the first CSI, including: performing channel estimation based on the first reference signal to obtain first channel estimation information; performing eigendecomposition, such as SVD, on the first channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information; and using a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI.

[0304] In an optional embodiment, the second CSI is determined by the first device performing channel estimation based on the second reference signal, including: the second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; the second CSI is obtained by transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information using a channel transformation machine learning model.

[0305] In an optional embodiment, the processing unit 1801 is further used to, after determining the first encoder based on the first CSI and the first channel characteristic information, transform the first encoder using a parameter sharing machine learning model to obtain a second encoder.

[0306] In an optional embodiment, the processing unit 1801 is further used to, after determining the first decoder based on the first CSI and the first channel characteristic information, transform the first decoder using a parameter sharing machine learning model to obtain a second decoder.

[0307] In an optional embodiment, the first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; the communication device 1800 is a terminal device or a chip in the terminal device, and the first device is a network device or a chip in the network device; the processing unit 1801 determines the first encoder based on the first CSI and the first channel characteristic information.

[0308] Optionally, the communication unit 1802 is also used to receive one or more of the following from the first device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0309] In an optional embodiment, the first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; the communication device 1800 is a network device or a chip in a network device, and the first device is a terminal device or a chip in a terminal device; the processing unit 1801 determines the first decoder based on the first CSI and the first channel characteristic information.

[0310] Optionally, the communication unit 1802 is also used to send one or more of the following to the first device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0311] In another possible design, for a case where the communication apparatus 1800 is used to implement the function of the first device in the above method embodiment:

[0312] The processing unit 1801 is configured to determine a first encoder and a first decoder.

[0313] The communication unit 1802 is configured to send a first reference signal to the second device, where the first reference signal is used by the second device to perform channel estimation and determine a first CSI.

[0314] The processing unit 1801 is further configured to perform channel estimation based on a second reference signal from the second device to determine a second CSI.

[0315] The processing unit 1801 is further configured to compress the second CSI using the first encoder to obtain first channel characteristic information.

[0316] The communication unit 1802 is further configured to send the first channel characteristic information to the second device.

[0317] The first CSI and the first channel characteristic information are used by the second device to determine the first encoder or the first decoder; the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0318] In an optional embodiment, the processing unit 1801 performs channel estimation based on a second reference signal from the second device to determine the second CSI, and is specifically used to: perform channel estimation based on the second reference signal to obtain second channel estimation information; the second CSI is the second channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the second channel estimation information.

[0319] Optionally, the processing unit 1801 performs channel estimation based on a second reference signal from the second device to determine a second CSI, and is specifically used to: perform channel estimation based on the second reference signal to obtain second channel estimation information; and perform eigendecomposition, such as SVD, on the second channel estimation information to obtain a second CSI.

[0320] In an optional embodiment, the first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: the first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; the first CSI is the first channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the first channel estimation information.

[0321] In an optional embodiment, the processing unit 1801 performs channel estimation based on a second reference signal from the second device to determine the second CSI, and is specifically used to: perform channel estimation based on the second reference signal to obtain second channel estimation information; transform the second channel estimation information using a channel transformation machine learning model to obtain fourth channel estimation information; the second CSI is the fourth channel estimation information, or the second CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the fourth channel estimation information.

[0322] Optionally, the processing unit 1801 performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: performing channel estimation based on the second reference signal to obtain second channel estimation information; transforming the second channel estimation information using a channel transformation machine learning model to obtain fourth channel estimation information; and performing feature decomposition, such as SVD, on the fourth channel estimation information to obtain the second CSI.

[0323] In an optional embodiment, the first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: the first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; the third channel estimation information is obtained by transforming the first channel estimation information using a channel transformation machine learning model; the first CSI is the third channel estimation information, or the first CSI is the eigenvalue or eigenvector of the channel matrix corresponding to the third channel estimation information.

[0324] In an optional embodiment, the processing unit 1801 performs channel estimation based on a second reference signal from the second device to determine the second CSI, and is specifically used to: perform channel estimation based on the second reference signal to obtain second channel estimation information; and use a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the second CSI.

[0325] Optionally, the processing unit 1801 performs channel estimation based on a second reference signal from the second device to determine the second CSI, including: performing channel estimation based on the second reference signal to obtain second channel estimation information; performing eigendecomposition, such as SVD, on the second channel estimation information to obtain the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information; and using a channel transformation machine learning model to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the second CSI.

[0326] In an optional embodiment, the first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: the first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; the first CSI is obtained by transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information using a channel transformation machine learning model.

[0327] In an optional embodiment, the processing unit 1801 is further configured to transform the first decoder using a parameter-sharing machine learning model to obtain a second decoder.

[0328] In an optional embodiment, the processing unit 1801 is further used to transform the first encoder using a parameter sharing machine learning model to obtain a second encoder.

[0329] In an optional embodiment, the first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; the communication device 1800 is a network device or a chip in a network device, and the second device is a terminal device or a chip in a terminal device; the first CSI and the first channel characteristic information are used by the second device to determine the first encoder.

[0330] Optionally, the communication unit 1802 is also used to send one or more of the following to the second device: first information, second information and third information; wherein the first information is used to indicate the time-frequency resource position of the first reference signal; the second information is used to indicate the time-frequency resource position of the second reference signal; and the third information is used to indicate the time-frequency resource position of the first channel characteristic information.

[0331] In an optional embodiment, the first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; the communication device 1800 is a terminal device or a chip in the terminal device, and the second device is a network device or a chip in the network device; the first CSI and the first channel characteristic information are used by the second device to determine the first decoder.

[0332] Optionally, the communication unit 1802 is also used to receive one or more of the following from the second device: first information, second information, and third information; wherein the first information is used to indicate the time-frequency resource position corresponding to the first reference signal; the second information is used to indicate the time-frequency resource position corresponding to the second reference signal; and the third information is used to indicate the time-frequency resource position corresponding to the first channel characteristic information.

[0333] The application embodiment and the method embodiment shown above are based on the same concept, and the technical effects they bring are also the same. For the specific principles, please refer to the description of the embodiment shown above, and no further details will be given.

[0334] The present application also provides a communication device 1900, as shown in Figure 19. Communication device 1900 can be a first device or a second device, or a chip, chip system, or processor that supports the first device in implementing the above method, or a chip, chip system, or processor that supports the second device in implementing the above method. This device can be used to implement the method described in the above method embodiment. For details, please refer to the description of the above method embodiment.

[0335] The communication device 1900 may include one or more processors 1901. The processor 1901 may be used to implement some or all of the functions of the first device or the second device through logic circuits or running computer programs. The processor 1901 may be a general-purpose processor or a dedicated processor, etc. For example, it may be a baseband processor, a digital signal processor, an application-specific integrated circuit, a field programmable gate array or other programmable logic device, a discrete gate or transistor logic device, a discrete hardware component or a central processing unit (CPU). The baseband processor may be used to process communication protocols and communication data, and the central processing unit may be used to control the communication device, execute software programs, and process data of the software programs, wherein the communication device is, for example, a base station, a baseband chip, a terminal, a terminal chip, a distributed unit (DU) or a distributed unit (CU), etc.

[0336] Optionally, the communication device 1900 may include one or more memories 1902, on which instructions 1904 may be stored. The instructions may be executed on the processor 1901, causing the communication device 1900 to perform the method described in the above method embodiment. Optionally, the memory 1902 may also store data. The processor 1901 and the memory 1902 may be provided separately or integrated together.

[0337] The memory 1902 may include, but is not limited to, non-volatile memory such as a hard disk drive (HDD) or a solid-state drive (SSD), random access memory (RAM), erasable programmable ROM (EPROM), ROM or compact disc read-only memory (CD-ROM), etc.

[0338] Optionally, the communication device 1900 may further include a transceiver 1905 and an antenna 1906. The transceiver 1905 may be referred to as a transceiver unit, a transceiver, or a transceiver circuit, etc., and is configured to implement transceiver functions. The transceiver 1905 may include a receiver and a transmitter. The receiver may be referred to as a receiver or a receiving circuit, etc., and is configured to implement a receiving function; the transmitter may be referred to as a transmitter or a transmitting circuit, etc., and is configured to implement a transmitting function.

[0339] In one possible design, processor 1901 is configured to perform channel estimation based on a first reference signal from a first device to determine a first CSI. Transceiver 1905 is configured to send a second reference signal to the first device, where the second reference signal is used by the first device to perform channel estimation and determine a second CSI. Transceiver 1905 is further configured to receive first channel characteristic information from the first device, where the first channel characteristic information is obtained by the first device compressing the second CSI using a first encoder. Processor 1901 is further configured to determine a first encoder or a first decoder based on the first CSI and the first channel characteristic information. The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0340] Exemplarily, the communication device 1900 may be configured to perform the related operations of the second device in the information transmission method shown in FIG9 . The processor 1901 may be configured to perform the related operations of steps S104 and S108 shown in FIG9 , and the transceiver 1905 may be configured to perform the related operations of receiving the first reference signal in step S102, the related operations of sending the second reference signal in step S103, and the related operations of receiving the first channel characteristic information in step S107 shown in FIG9 .

[0341] In another possible design, processor 1901 is used to determine a first encoder and a first decoder. Transceiver 1905 is used to send a first reference signal to a second device, where the first reference signal is used by the second device to perform channel estimation and determine a first CSI. Processor 1901 is also used to perform channel estimation based on a second reference signal from the second device to determine a second CSI. Processor 1901 is also used to compress the second CSI using the first encoder to obtain first channel characteristic information. Transceiver 1905 is also used to send the first channel characteristic information to the second device. The first CSI and the first channel characteristic information are used by the second device to determine the first encoder or the first decoder; the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

[0342] Exemplarily, the communication device 1900 may be used to perform the related operations of the first device in the information transmission method shown in Figure 9. The processor 1901 may be used to perform the related operations of steps S101, S105, and S106 shown in Figure 9, and the transceiver 1905 may be used to perform the related operations of sending the first reference signal in step S102, the related operations of receiving the second reference signal in step S103, and the related operations of sending the first channel characteristic information in step S107 shown in Figure 9.

[0343] In another possible design, processor 1901 may include a transceiver for implementing receiving and transmitting functions. For example, the transceiver may be a transceiver circuit, an interface, or an interface circuit. The transceiver circuit, interface, or interface circuit for implementing the receiving and transmitting functions may be separate or integrated. The transceiver circuit, interface, or interface circuit may be used for reading and writing code / data, or the transceiver circuit, interface, or interface circuit may be used for transmitting or delivering signals.

[0344] In another possible design, processor 1901 may optionally store instructions 1903. Instructions 1903, when executed on processor 1901, may cause communication device 1900 to perform the method described in the above method embodiment. Instructions 1903 may be fixed in processor 1901. In this case, processor 1901 may be implemented by hardware.

[0345] In another possible design, the communication device 1900 may include a circuit that can implement the functions of sending, receiving, or communicating in the aforementioned method embodiments. The processor and transceiver described in the embodiments of the present application can be implemented in an integrated circuit (IC), an analog IC, a radio frequency integrated circuit (RFIC), a mixed signal IC, an application specific integrated circuit (ASIC), a printed circuit board (PCB), an electronic device, etc. The processor and transceiver can also be manufactured using various IC process technologies, such as complementary metal oxide semiconductor (CMOS), N-type metal oxide semiconductor (nMetal-oxide-semiconductor, NMOS), P-type metal oxide semiconductor (positive channel metal oxide semiconductor, PMOS), bipolar junction transistor (bipolar junction transistor, BJT), bipolar CMOS (BiCMOS), silicon germanium (SiGe), gallium arsenide (GaAs), etc.

[0346] Those skilled in the art will also appreciate that the various illustrative logical blocks and steps listed in the embodiments of the present application can be implemented by electronic hardware, computer software, or a combination of both. Whether such functions are implemented by hardware or software depends on the specific application and the design requirements of the entire system. Those skilled in the art may use various methods to implement the described functions for each specific application, but such implementation should not be construed as exceeding the scope of protection of the embodiments of the present application.

[0347] The embodiments of the present application and the above-mentioned method embodiments are based on the same concept, and the technical effects they bring are also the same. For the specific principles, please refer to the description in the above-mentioned method embodiments, and no further details will be given.

[0348] The present application also provides a computer-readable storage medium for storing computer software instructions, which, when executed by a communication device, implements the functions of any of the above method embodiments.

[0349] The present application also provides a computer program product for storing computer software instructions, which, when executed by a communication device, implements the functions of any of the above method embodiments.

[0350] The present application also provides a computer program that, when executed on a computer, implements the functions of any of the above method embodiments.

[0351] 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 instructions are loaded and executed on a computer, all or part of the processes or functions described in 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 from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server, or data center to another website, computer, server, or data center 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 magnetic medium (eg, a floppy disk, a hard disk, a magnetic tape), an optical medium (eg, a high-density digital video disc (DVD)), or a semiconductor medium (eg, an SSD).

[0352] The above description is merely a specific embodiment of the present application, but the scope of protection of the present application is not limited thereto. Any changes or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in this application should be included in the scope of protection of this application. Therefore, the scope of protection of this application should be based on the scope of protection of the claims.

Claims

1. An information transmission method, characterized in that: The method comprises: Perform channel estimation based on a first reference signal from the first device to determine first channel state information CSI; Sending a second reference signal to the first device, where the second reference signal is used by the first device to perform channel estimation to determine a second CSI; receiving first channel characteristic information from the first device, where the first channel characteristic information is obtained by the first device compressing the second CSI using a first encoder; Determine the first encoder or the first decoder based on the first CSI and the first channel characteristic information; The channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

2. The method according to claim 1, characterized in that The determining the first CSI by performing channel estimation based on a first reference signal from the first device includes: Perform channel estimation based on the first reference signal to obtain first channel estimation information; The first CSI is the first channel estimation information; or, the first CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the first channel estimation information.

3. The method according to claim 1 or 2, characterized in that: The second CSI is determined by the first device through channel estimation based on the second reference signal, including: The second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; The second CSI is the second channel estimation information; or, the second CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the second channel estimation information.

4. The method according to claim 1, characterized in that: The determining the first CSI by performing channel estimation based on a first reference signal from the first device includes: Perform channel estimation based on the first reference signal to obtain first channel estimation information; transforming the first channel estimation information using a channel transformation machine learning model to obtain third channel estimation information; The first CSI is the third channel estimation information; or, the first CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the third channel estimation information.

5. The method according to claim 1 or 4, characterized in that: The second CSI is determined by the first device through channel estimation based on the second reference signal, including: The second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; The fourth channel estimation information is obtained by transforming the second channel estimation information by the first device using a channel transformation machine learning model; The second CSI is the fourth channel estimation information; or, the second CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the fourth channel estimation information.

6. The method according to claim 1, characterized in that The determining the first CSI by performing channel estimation based on a first reference signal from the first device includes: Perform channel estimation based on the first reference signal to obtain first channel estimation information; A channel transformation machine learning model is used to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information to obtain the first CSI.

7. The method according to claim 1 or 6, characterized in that: The second CSI is determined by the first device through channel estimation based on the second reference signal, including: The second channel estimation information is obtained by the first device performing channel estimation based on the second reference signal; The second CSI is obtained by transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information using a channel transformation machine learning model.

8. The method according to any one of claims 1 to 7, characterized in that: After determining the first encoder based on the first CSI and the first channel characteristic information, the method further includes: The first encoder is transformed using a parameter sharing machine learning model to obtain a second encoder.

9. The method according to any one of claims 1 to 7, characterized in that: After determining the first decoder based on the first CSI and the first channel characteristic information, the method further includes: The first decoder is transformed using a parameter sharing machine learning model to obtain a second decoder.

10. The method according to any one of claims 1 to 8, characterized in that: The first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; The execution subject of the method is a terminal device or a chip in the terminal device, and the first device is a network device or a chip in the network device; The first encoder is determined based on the first CSI and the first channel characteristic information.

11. The method according to claim 10, characterized in that The method further comprises: receiving one or more of the following from the first device: first information, second information, and third information; The first information is used to indicate the time-frequency resource position of the first reference signal; The second information is used to indicate a time-frequency resource position of the second reference signal; The third information is used to indicate the time-frequency resource position of the first channel characteristic information.

12. The method according to any one of claims 1 to 7 and 9, characterized in that: The first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; The execution subject of the method is a network device or a chip in the network device, and the first device is a terminal device or a chip in the terminal device; The first decoder is determined based on the first CSI and the first channel characteristic information.

13. The method according to claim 12, characterized in that The method further comprises: Sending one or more of the following to the first device: first information, second information, and third information; The first information is used to indicate the time-frequency resource position of the first reference signal; The second information is used to indicate a time-frequency resource position of the second reference signal; The third information is used to indicate the time-frequency resource position of the first channel characteristic information.

14. An information transmission method, characterized in that: The method comprises: determining a first encoder and a first decoder; Sending a first reference signal to a second device, where the first reference signal is used by the second device to perform channel estimation and determine first channel state information CSI; Perform channel estimation based on a second reference signal from the second device to determine a second CSI; compressing the second CSI using the first encoder to obtain first channel characteristic information; sending the first channel characteristic information to the second device; The first CSI and the first channel characteristic information are used by the second device to determine the first encoder or the first decoder; and the channel corresponding to the first reference signal and the channel corresponding to the second reference signal are reciprocal.

15. The method according to claim 14, characterized in that The determining a second CSI by performing channel estimation based on a second reference signal from the second device includes: Perform channel estimation based on the second reference signal to obtain second channel estimation information; The second CSI is the second channel estimation information; or, the second CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the second channel estimation information.

16. The method according to claim 14 or 15, characterized in that The first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: The first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; The first CSI is the first channel estimation information; or, the first CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the first channel estimation information.

17. The method according to claim 14, characterized in that The determining a second CSI by performing channel estimation based on a second reference signal from the second device includes: Perform channel estimation based on the second reference signal to obtain second channel estimation information; transforming the second channel estimation information using a channel transformation machine learning model to obtain fourth channel estimation information; The second CSI is the fourth channel estimation information; or, the second CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the fourth channel estimation information.

18. The method according to claim 14 or 17, characterized in that The first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: The first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; The third channel estimation information is obtained by transforming the first channel estimation information using a channel transformation machine learning model; The first CSI is the third channel estimation information; or, the first CSI is an eigenvalue or an eigenvector of a channel matrix corresponding to the third channel estimation information.

19. The method according to claim 14, characterized in that The determining a second CSI by performing channel estimation based on a second reference signal from the second device includes: Perform channel estimation based on the second reference signal to obtain second channel estimation information; A channel transformation machine learning model is used to transform the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the second channel estimation information to obtain the second CSI.

20. The method according to claim 14 or 19, characterized in that The first reference signal is used by the second device to perform channel estimation to determine the first CSI, including: The first reference signal is used by the second device to perform channel estimation to obtain first channel estimation information; The first CSI is obtained by transforming the eigenvalues ​​or eigenvectors of the channel matrix corresponding to the first channel estimation information using a channel transformation machine learning model.

21. The method according to any one of claims 14 to 20, characterized in that The method further comprises: The first decoder is transformed using a parameter sharing machine learning model to obtain a second decoder.

22. The method according to any one of claims 14 to 20, characterized in that The method further comprises: The first encoder is transformed using a parameter sharing machine learning model to obtain a second encoder.

23. The method according to any one of claims 14 to 21, characterized in that The first reference signal is a downlink reference signal, and the second reference signal is an uplink reference signal; The execution subject of the method is a network device or a chip in the network device, and the second device is a terminal device or a chip in the terminal device; The first CSI and the first channel characteristic information are used by the second device to determine the first encoder.

24. The method according to claim 23, characterized in that The method further comprises: Sending one or more of the following to the second device: first information, second information, and third information; The first information is used to indicate the time-frequency resource position of the first reference signal; The second information is used to indicate a time-frequency resource position of the second reference signal; The third information is used to indicate the time-frequency resource position of the first channel characteristic information.

25. The method according to any one of claims 14 to 20 and 22, characterized in that: The first reference signal is an uplink reference signal, and the second reference signal is a downlink reference signal; The execution subject of the method is a terminal device or a chip in the terminal device, and the second device is a network device or a chip in the network device; The first CSI and the first channel characteristic information are used by the second device to determine the first decoder.

26. The method according to claim 25, characterized in that The method further comprises: receiving one or more of the following from the second device: first information, second information, and third information; The first information is used to indicate the time-frequency resource position corresponding to the first reference signal; The second information is used to indicate a time-frequency resource position corresponding to the second reference signal; The third information is used to indicate the time-frequency resource position corresponding to the first channel characteristic information.

27. A communication device, characterized in that: The device includes a module or unit for implementing the method according to any one of claims 1 to 13, or the device includes a module or unit for implementing the method according to any one of claims 14 to 26.

28. A communication device, characterized in that: Including processors; The processor is used to execute a computer program or instruction in a memory so that the communication device executes the method described in any one of claims 1 to 13, or so that the communication device executes the method described in any one of claims 14 to 26.

29. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed, the method according to any one of claims 1 to 13 is implemented, or the method according to any one of claims 14 to 26 is implemented.

30. A computer program product, characterized in that The computer program product comprises: a computer program code, and when the computer program code is run, it implements the method according to any one of claims 1 to 13, or implements the method according to any one of claims 14 to 26.

Images