MIMO communication information sending method and receiving method based on self-coding

Abstract

The invention provides an MIMO communication information sending method and receiving method based on self-coding, and the information sending method comprises the steps: obtaining to-be-sent information, inputting the to-be-sent information to a trained coding neural network for coding, and obtaining # imgabs0 # coding signals outputted by the coding neural network; the # imgabs 1 # coded signals are sent through # imgabs 2 # channels respectively, and each channel corresponds to one channel; the information receiving method comprises the following steps: receiving a # imgabs4 # path coding signal through a # imgabs3 # channel; and inputting the # imgabs5 path coding signal into a trained decoding neural network for decoding, and obtaining decoding information output by the decoding neural network. Wherein the coding neural network and the decoding neural network are jointly obtained by training based on multiple groups of training data, and each group of training data comprises sample to-be-sent information and sample decoding information corresponding to the sample to-be-sent information. According to the invention, the calculation complexity of the MIMO communication system can be reduced, and the receiving performance of the MIMO communication system in an actual multi-path environment is improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technology, and in particular to a MIMO communication information sending method and a receiving method based on self-coding. Background Art

[0002] MIMO (multiple input multiple output) multi-antenna technology is widely used in wireless communication systems. MIMO technology allows communication systems to expand from the time and frequency dimensions to the spatial dimension for signal processing, and utilizes the spatial selectivity generated in the multipath random fading environment, which can avoid the problems of multipath and random fading to a certain extent and ensure broadband transmission. However, the performance of the current MIMO communication system depends on the coding, beamforming and decoding technologies obtained through rigorous analysis, and these modules are usually locally optimized, resulting in certain performance losses after cascading. At the same time, in the spatial division multiplexing and spatial diversity system of the MIMO multi-antenna system, there is interference in the signal transmission between different antennas. How to eliminate the interference between antennas and ensure the optimal demodulation performance at the receiving end is a difficult problem in MIMO transmission. In addition, spatial multiplexing technology mainly relies on channel state information estimation, quantization and feedback, which further makes the system computationally complex and difficult to obtain optimal performance. Summary of the invention

[0003] The present invention provides a MIMO communication information sending method and a receiving method based on self-coding, so as to solve the defect of high computational complexity of the MIMO communication system in the prior art and reduce the computational complexity of the MIMO communication system.

[0004] The present invention provides a MIMO communication information sending method based on self-coding, comprising: Obtain information to be sent, input the information to be sent into a trained encoding neural network for encoding, and obtain the output of the encoding neural network Channel coded signal; The The coded signals are passed through Channels are used for transmission, and each signal corresponds to one channel; The encoding neural network and the decoding neural network are trained together based on multiple groups of training data, and each group of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0005] According to a self-encoding-based MIMO communication information sending method provided by the present invention, the encoding neural network includes a first fully connected layer, a second fully connected layer, a first one-dimensional convolutional layer, a second one-dimensional convolutional layer and a first normalization layer connected in sequence.

[0006] According to a method for sending MIMO communication information based on self-encoding provided by the present invention, the training process of the encoding neural network includes: The sample information to be sent is input into the encoding neural network for encoding, and the output of the encoding neural network is obtained. A sample coded signal; Get the a sample receiving signal corresponding to a sample coded signal, inputting the sample receiving signal into the decoding neural network, and obtaining sample decoding information output by the decoding neural network; Determining a training loss based on a difference between the sample decoded information and the sample to-be-sent information; The encoding neural network and the decoding neural network are updated based on the training loss.

[0007] According to a method for sending MIMO communication information based on self-coding provided by the present invention, the obtaining The sample receiving signal corresponding to the sample coded signal includes: Regarding the The sample coded information of each path is faded through the multipath channel and Gaussian white noise is added to obtain the sample received signal.

[0008] The present invention also provides a MIMO communication information receiving method based on self-encoding, comprising: pass Channels received Channel coded signal; The The coded signal is input into a trained decoding neural network for decoding, and decoding information output by the decoding neural network is obtained; The decoding neural network and the encoding neural network are trained together based on multiple groups of training data, and each group of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0009] According to a MIMO communication information receiving method based on self-encoding provided by the present invention, the decoding neural network includes a third one-dimensional convolutional layer, a fourth one-dimensional convolutional layer, a third fully connected layer and a fourth fully connected layer connected in sequence.

[0010] The present invention also provides a MIMO communication system based on self-coding, comprising: The encoding module is used to obtain the information to be sent, input the information to be sent into the trained encoding neural network for encoding, and obtain the output of the encoding neural network. Channel coded signal; A sending module is used to send the The coded signals are passed through Channels are used for transmission, and each signal corresponds to one channel; Receive module, used to Channels received Channel coded signal; A decoding module is used to convert the The coded signal is input into a trained decoding neural network for decoding, and decoding information output by the decoding neural network is obtained; The encoding neural network and the decoding neural network are trained together based on multiple groups of training data, and each group of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0011] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the processor implements any one of the above-described methods for sending MIMO communication information based on self-coding and / or methods for receiving MIMO communication information based on self-coding.

[0012] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described MIMO communication information sending methods based on self-coding and / or MIMO communication information receiving methods based on self-coding.

[0013] The present invention also provides a computer program product, comprising a computer program, wherein when the computer program is executed by a processor, the computer program implements any of the above-mentioned MIMO communication information sending methods based on self-coding and / or MIMO communication information receiving methods based on self-coding.

[0014] The MIMO communication information sending method and receiving method based on self-encoding provided by the present invention obtain information to be sent, input the information to be sent into a trained encoding neural network for encoding, and obtain the output of the encoding neural network. The coded signal will Do not pass the road number signal Compared with the traditional MIMO communication which uses spatial multiplexing technology and relies on the mechanism of channel state information estimation, quantization and feedback, the present invention constructs a MIMO communication system through a neural network, which can greatly reduce the computational complexity of the system. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flow chart of the method for sending MIMO communication information based on self-coding provided by the present invention.

[0017] Figure 2 It is a flow chart of the MIMO communication information receiving method based on self-coding provided by the present invention.

[0018] Figure 3 It is a flow chart of a MIMO communication information sending method and a receiving method based on self-coding provided by the present invention.

[0019] Figure 4 This is a schematic diagram of the effect of the MIMO communication information sending method based on self-coding provided by the present invention Figure 1 .

[0020] Figure 5 The effect of the MIMO communication information sending method based on self-coding provided by the present invention is shown in FIG. Figure 2 .

[0021] Figure 6 It is a structural schematic diagram of a MIMO communication information sending device based on self-coding provided by the present invention.

[0022] Figure 7 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0023] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0024] MIMO (multiple input multiple output) multi-antenna technology is widely used in wireless communication systems. MIMO technology enables communication systems to expand from the time and frequency dimensions to the spatial dimension for signal processing, and utilizes the spatial selectivity generated in a multipath random fading environment, which can avoid the problems of multipath and random fading to a certain extent and ensure broadband transmission.

[0025] However, the performance of current MIMO communication systems depends on coding, beamforming, and decoding technologies obtained through rigorous analysis, and these modules are usually locally optimized, resulting in certain performance losses after cascading. At the same time, in the spatial division multiplexing and spatial diversity systems of MIMO multi-antenna systems, there is interference in the signal transmission between different antennas. How to eliminate the interference between antennas while ensuring the optimal demodulation performance at the receiving end is a difficult problem in MIMO transmission. In addition, spatial multiplexing technology mainly relies on channel state information estimation, quantization, and feedback, which further makes the system computationally complex and difficult to obtain optimal performance.

[0026] In view of the defect of high computational complexity of MIMO communication systems in the prior art, the present invention provides a MIMO communication information sending method and a receiving method based on self-encoding, so as to achieve the technical effect of reducing the computational complexity of the MIMO communication system.

[0027] Combine the following Figure 1-Figure 6 The present invention describes a method for sending MIMO communication information based on self-encoding, which is applied to the sending end in MIMO communication. Figure 1 As shown, the MIMO communication information sending method based on self-encoding includes the steps of: S110, obtaining the information to be sent, inputting the information to be sent into the trained coding neural network for encoding, and obtaining the output of the coding neural network. Channel coded signal; S120, The coded signals are passed through Channels are used for transmission, and each signal corresponds to one channel; The encoding neural network and the decoding neural network are trained together based on multiple sets of training data, and each set of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0028] The method provided by the present invention obtains information to be sent, inputs the information to be sent into a trained coding neural network for encoding, and obtains the output of the coding neural network. The coded signal will Do not pass the road number signal Compared with the traditional MIMO communication which uses spatial multiplexing technology and relies on the mechanism of channel state information estimation, quantization and feedback, the present invention constructs a MIMO communication system through a neural network, learns the communication channel environment through the neural network, and encodes the signal to be sent to obtain The coded signals of each channel are sent using the corresponding channels respectively, without the need for channel state information estimation, quantization and feedback, which reduces the complexity of the MIMO communication system and improves its receiving performance in an actual multipath environment.

[0029] Specifically, the encoding neural network includes a first fully connected layer, a second fully connected layer, a first one-dimensional convolutional layer, a second one-dimensional convolutional layer and a normalization layer which are connected in sequence.

[0030] Among them, the activation function of the first fully connected layer is ReLU, and the output dimension is , the activation function of the second fully connected layer is Linear, and the output dimension is , the activation function of the first one-dimensional convolutional layer is eLU, and the output dimension is , the activation function of the second one-dimensional convolutional layer is eLU, and the output dimension is , the normalization layer is used to normalize the baseband signal power of each batch, and the output dimension is , the number of convolutional layers is 16, and the convolution kernel size is 1×100, where is a constant, B represents the number of symbols per frame, represents the number of transmitting antennas, and n represents the coding bit length.

[0031] Through the encoding neural network, the k-bit length of the information to be sent S is mapped to The transmitted baseband signal is n bits long. .

[0032] Correspondingly, based on the above-mentioned MIMO communication information sending method based on self-encoding, the present invention also provides a MIMO communication information receiving method based on self-encoding, which is applied to the receiving end in MIMO communication, such as Figure 2 As shown, the method includes: S210, through Channels received Channel coded signal; S220, The coded signal is input into the trained decoding neural network for decoding, and the decoding information output by the decoding neural network is obtained; The encoding neural network and the decoding neural network are trained together based on multiple sets of training data, and each set of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0033] After the encoded signal is obtained and sent through the encoding neural network, the receiving end receives the encoded signal, inputs the encoded signal into the decoding neural network for decoding, obtains the decoded information output by the decoding neural network, and restores the information to be sent. Can be equal to , or it may not be equal to .

[0034] The decoding neural network includes a third one-dimensional convolutional layer, a fourth one-dimensional convolutional layer, a third fully connected layer and a fourth fully connected layer.

[0035] Among them, the activation function of the third one-dimensional convolutional layer is eLU, and the output dimension is , the activation function of the fourth one-dimensional convolutional layer is eLU, and the output dimension is , the activation function of the third fully connected layer is ReLU, and the output dimension is , the activation function of the fourth fully connected layer is softmax, and the output dimension is , the number of convolutional layers is 16, the convolution kernel size is 1×100, k represents the source bit length, that is, the bit length of the information to be sent, Indicates the number of receiving antennas.

[0036] The architecture of each network layer of the encoding neural network and the decoding neural network in the method provided by the present invention can be summarized as shown in Table 1.

[0037] Table 1

[0038] The one-dimensional convolutional layer in the encoding neural network and the decoding neural network in the method provided by the present invention is mainly used to convert the size of 1* q The convolution kernel vector In the information vector Perform traversal convolution on it.

[0039] Among them, the processing process of the convolution layer is shown as follows

[0040] .

[0041] In addition, the expressions of the ReLU, Linear, eLU and Softmax activation functions of neurons in the fully connected layer are: , ReLU

[0042] , Linear

[0043] , eLU

[0044] , Softmax

[0045] In the formula, , are the input and output of the neuron, , are the weight and bias of the neuron respectively, L represents the output dimension, j Indicates the output j elements.

[0046] In addition, the normalization expression of the normalization layer in the encoding network is: .

[0047] Before the transmitter and receiver formally communicate, the encoding neural network and the decoding neural network need to be trained so that the encoding neural network and the decoding neural network converge to the global optimal solution. The specific training process includes: Input the sample information to be sent into the encoding neural network for encoding, and obtain the output of the encoding neural network A sample coded signal; Get The sample receiving signal corresponding to the sample coded signal is input into the decoding neural network to obtain the sample decoding information output by the decoding neural network; Determining a training loss based on a difference between sample decoded information and sample to-be-sent information; Update the encoder neural network and decoder neural network based on the training loss.

[0048] During the training process, the sample information sequence to be sent is used , generated by encoding neural network The roadbed signal then passes through the channel to the receiving end, where the decoding neural network uses Roadbed signal recovery training sequence , the receiving end then calculates the loss value L. Among them, Finally, the parameters of the encoding neural network and the decoding neural network are optimized and adjusted by using the loss value L to calculate the back propagation gradient through the Adam optimizer. The above process is repeated until the loss value basically does not change during the training process. At this time, the encoding neural network and the decoding neural network have reached the optimal state and the iterative training is stopped.

[0049] After the transmitter sends out the coded signal, the coded signal received by the receiver may not be exactly the same as the coded signal sent by the transmitter due to the influence of other factors in the transmission path. This influence can be added to the training of the encoding neural network and the decoding neural network through actual signal transmission, so that the performance of the trained encoding neural network and the decoding neural network is more suitable for actual communication. Since a large amount of data needs to be calculated during the training process, the actual sending and receiving may consume a lot of resources. In another possible implementation method, obtaining The sample receiving signal corresponding to the sample coded signal includes: right The sample coded information is faded through the multipath channel and Gaussian white noise is added to obtain the sample received signal.

[0050] like Figure 3 As shown, the sample coding information can be faded through multipath channels and Gaussian white noise can be added to simulate the impact on the signal in actual transmission, which can effectively reduce the resource consumption during training. The signal received by the receiving end after adding Gaussian white noise can be expressed as:

[0051] in, represents Gaussian white noise, * represents convolution operation, represents the baseband signal sent by the i-th channel, Indicates the jth received signal.

[0052] The method provided by the present invention is a multi-antenna transceiver system based on an autoencoder architecture, which establishes a coding neural network and a decoding neural network at both ends of the transceiver, and learns an intelligent waveform that is suitable for MIMO multi-antenna scenarios after training.

[0053] The effect of the method provided by the present invention was verified by experiments, and the results are as follows Figure 4 and Figure 5 As shown. Among them, Figure 4 It is a comparison diagram of bit error rate curves of communication using other methods and the method provided by the present invention in a 2×2 MIMO scenario under an EVA multipath channel. Figure 5 It is a comparison diagram of bit error rate curves of communication using other methods and the method provided by the present invention in a 2×1 MISO (multiple inputs, one output) scenario under a Rayleigh channel.

[0054] It can be seen from the experimental results that, compared with the traditional MIMO communication waveform, the MIMO intelligent waveform based on the self-encoding architecture generated by the method provided by the present invention can greatly reduce the computational complexity of the system when the number of antennas increases, and achieves better bit error rate performance in non-Gaussian channels.

[0055] The following describes the MIMO communication system based on self-coding provided by the present invention. The MIMO communication system based on self-coding described below and the MIMO communication information sending method based on self-coding and the MIMO communication information receiving method based on self-coding described above can be referred to each other. Figure 6 As shown, the MIMO communication system based on self-encoding provided by the present invention includes: The encoding module 610 is used to obtain the information to be sent, input the information to be sent into the trained encoding neural network for encoding, and obtain the output of the encoding neural network. Channel coded signal; The sending module 620 is used to The coded signals are passed through Channels are used for transmission, and each signal corresponds to one channel; The receiving module 630 is used to Channels received Channel coded signal; The decoding module 640 is used to The coded signal is input into the trained decoding neural network for decoding, and the decoding information output by the decoding neural network is obtained; The encoding neural network and the decoding neural network are trained together based on multiple sets of training data, and each set of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0056] Furthermore, the encoding neural network includes a first fully connected layer, a second fully connected layer, a first one-dimensional convolutional layer, a second one-dimensional convolutional layer and a first normalization layer connected in sequence, and the decoding neural network includes a third one-dimensional convolutional layer, a fourth one-dimensional convolutional layer, a third fully connected layer and a fourth fully connected layer connected in sequence.

[0057] Furthermore, the MIMO communication information sending device based on self-coding also includes a training module, which is specifically used for: Input the sample information to be sent into the encoding neural network for encoding, and obtain the output of the encoding neural network A sample coded signal; Get The sample receiving signal corresponding to the sample coded signal is input into the decoding neural network to obtain the sample decoding information output by the decoding neural network; Determining a training loss based on a difference between sample decoded information and sample to-be-sent information; Update the encoder neural network and decoder neural network based on the training loss.

[0058] The training module is further specifically used for: The sample coded information passes through the multipath fading channel and Gaussian white noise is added to obtain the sample received signal.

[0059] Figure 7 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 7As shown, the electronic device may include: a processor 710, a communications interface 720, a memory 730 and a communication bus 740, wherein the processor 710, the communications interface 720 and the memory 730 communicate with each other via the communication bus 740. The processor 710 may call the logic instructions in the memory 730 to execute a MIMO communication information sending method based on self-coding and / or a MIMO communication information receiving method based on self-coding. The MIMO communication information sending method based on self-coding includes: obtaining information to be sent, inputting the information to be sent into a trained coding neural network for encoding, and obtaining the output of the coding neural network. Channel coded signal; The coded signals are passed through The MIMO communication information receiving method based on self-encoding includes: Channels received Channel coded signal; The coded signal is input into the trained decoding neural network for decoding to obtain decoding information output by the decoding neural network; wherein the encoding neural network and the decoding neural network are trained based on multiple sets of training data, and each set of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0060] In addition, the logic instructions in the above-mentioned memory 730 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on this understanding, the technical solution of the present invention can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a storage medium, including several instructions to enable a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc. Various media that can store program codes.

[0061] On the other hand, the present invention also provides a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can execute a MIMO communication information sending method based on self-encoding and / or a MIMO communication information receiving method based on self-encoding. The MIMO communication information sending method based on self-encoding includes: obtaining information to be sent, inputting the information to be sent into a trained encoding neural network for encoding, obtaining the output of the encoding neural network Channel coded signal; The coded signals are passed through The MIMO communication information receiving method based on self-encoding includes: Channels received Channel coded signal; The coded signal is input into the trained decoding neural network for decoding to obtain decoding information output by the decoding neural network; wherein the encoding neural network and the decoding neural network are trained based on multiple sets of training data, and each set of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0062] In another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented when the computer program is executed by a processor to execute a MIMO communication information sending method based on self-encoding and / or a MIMO communication information receiving method based on self-encoding. The MIMO communication information sending method based on self-encoding comprises: obtaining information to be sent, inputting the information to be sent into a trained encoding neural network for encoding, obtaining the output of the encoding neural network, Channel coded signal; The coded signals are passed through The MIMO communication information receiving method based on self-encoding includes: Channels received Channel coded signal; The coded signal is input into the trained decoding neural network for decoding to obtain decoding information output by the decoding neural network; wherein the encoding neural network and the decoding neural network are trained based on multiple sets of training data, and each set of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

[0063] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0064] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0065] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for sending MIMO communication information based on self-encoding, characterized in that: include: Obtain information to be sent, input the information to be sent into a trained encoding neural network for encoding, and obtain the output of the encoding neural network Channel coded signal; The The coded signals are respectively transmitted through Channels are used for transmission, and each signal corresponds to one channel; The encoding neural network and the decoding neural network are trained together based on multiple groups of training data, and each group of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

2. The method for sending MIMO communication information based on self-encoding according to claim 1, characterized in that: The encoding neural network includes a first fully connected layer, a second fully connected layer, a first one-dimensional convolutional layer, a second one-dimensional convolutional layer and a first normalization layer which are connected in sequence.

3. The method for sending MIMO communication information based on self-encoding according to claim 1, characterized in that: The training process of the encoding neural network includes: The sample information to be sent is input into the encoding neural network for encoding, and the output of the encoding neural network is obtained. A sample coded signal; Get the a sample receiving signal corresponding to a sample coded signal, inputting the sample receiving signal into the decoding neural network, and obtaining sample decoding information output by the decoding neural network; Determining a training loss based on a difference between the sample decoded information and the sample to-be-sent information; The encoding neural network and the decoding neural network are updated based on the training loss.

4. The method for sending MIMO communication information based on self-encoding according to claim 3, characterized in that: The acquisition of The sample receiving signal corresponding to the sample coded signal includes: Regarding the The sample coded information of each path is faded through the multipath channel and Gaussian white noise is added to obtain the sample received signal.

5. A method for receiving MIMO communication information based on self-encoding, characterized in that: include: pass Channels received Channel coded signal; The The coded signal is input into a trained decoding neural network for decoding, and decoding information output by the decoding neural network is obtained; The encoding neural network and the decoding neural network are trained together based on multiple groups of training data, and each group of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

6. The method for receiving MIMO communication information based on self-encoding according to claim 5, wherein: The decoding neural network includes a third one-dimensional convolutional layer, a fourth one-dimensional convolutional layer, a third fully connected layer and a fourth fully connected layer which are connected in sequence.

7. A MIMO communication system based on self-coding, characterized in that: include: The encoding module is used to obtain the information to be sent, input the information to be sent into the trained encoding neural network for encoding, and obtain the output of the encoding neural network. Channel coded signal; A sending module is used to send the The coded signals are respectively transmitted through Channels are used for transmission, and each signal corresponds to one channel; Receive module, used to Channels received Channel coded signal; A decoding module is used to convert the The coded signal is input into a trained decoding neural network for decoding, and decoding information output by the decoding neural network is obtained; The decoding neural network and the encoding neural network are trained together based on multiple groups of training data, and each group of training data includes sample information to be sent and sample decoding information corresponding to the sample information to be sent.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the method for sending MIMO communication information based on self-coding according to any one of claims 1 to 4 and / or the method for receiving MIMO communication information based on self-coding according to any one of claims 5 to 6 are implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the method for sending MIMO communication information based on self-coding according to any one of claims 1 to 4 and / or the method for receiving MIMO communication information based on self-coding according to any one of claims 5 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the method for sending MIMO communication information based on self-coding according to any one of claims 1 to 4 and / or the method for receiving MIMO communication information based on self-coding according to any one of claims 5 to 6 are implemented.

Images