Variable length channel coding method with feedback, electronic device, and storage medium

Abstract

The application provides a variable-length channel coding method with feedback, an electronic device and a storage medium. The method comprises the following steps: an encoding module encodes a to-be-sent bit stream to obtain first encoding output information, the information is input into a decoding module by superimposing forward channel additive Gaussian noise information, the decoding module determines the confidence information of each element in the received information, and judges whether the confidence information is greater than a preset signal threshold. If yes, the element is decoded, and feedback information of the decoding module is input into the encoding module by superimposing feedback channel additive Gaussian noise information. The encoding module encodes combined information, and inputs output information into the decoding module by superimposing forward channel additive Gaussian noise information again. The encoding module determines the confidence information corresponding to each element of the information. Until the confidence information of all elements is greater than the preset signal threshold, the synthesized output information is output. According to the application, different bit blocks are decoded according to the noise size carried by the bit blocks, so that adaptive code rate adjustment is realized.

Description

Technical Field

[0001] This application relates to the field of wireless communication technology, and more specifically, to a variable-length channel coding method with feedback, an electronic device, and a storage medium. Background Technology

[0002] In the field of wireless communication, channel coding and decoding techniques are crucial for ensuring signal transmission quality. In channel coding methods with feedback, the classic feedback channel model transmits data from the transmitter to the receiver through a memoryless noise channel. The receiver provides real-time feedback, which is then used for coding via the forward channel. This significantly improves coding reliability for finite lengths. However, this method cannot dynamically adapt to the feedback process and optimize performance under changing channel conditions and receiver states. Therefore, a channel coding method based on a deep learning model has been proposed. This method implements parallel input of bit blocks and iteratively utilizes enhanced downlink channels to obtain feedback information for channel coding. This makes the feedback coding unable to dynamically adapt to the feedback process, thereby reducing the bit error rate.

[0003] However, existing deep learning models are limited to fixed-length encodings, which restricts their adaptability and prevents them from fully utilizing feedback information, resulting in low flexibility. Furthermore, this method performs poorly at high bitrates. Summary of the Invention

[0004] The purpose of this application is to address the shortcomings of the prior art by providing a variable-length channel coding method, electronic device, and storage medium with feedback, thereby solving the problems of poor flexibility and adaptability in the prior art.

[0005] To achieve the above objectives, the technical solution adopted in this application is as follows:

[0006] In a first aspect, this application provides a variable-length channel coding method with feedback, the method being applied to a variable-length channel coding system with feedback, the variable-length channel coding system with feedback including an encoding module and a decoding module, wherein the transmission channel from the encoding module to the decoding module is a forward channel, and the transmission channel from the decoding module to the encoding module is a feedback channel, the method comprising:

[0007] The encoding module encodes the bit stream to be transmitted to obtain first encoded output information, and uses the sum of the first encoded output information and the first additive Gaussian noise information of the forward channel as the first decoding input information. The bit stream to be transmitted includes multiple bit blocks.

[0008] The decoding module performs confidence processing on the first decoding input information to determine the confidence information corresponding to each element in the first decoding input information, wherein each element corresponds to a bit block in the bit stream to be sent.

[0009] The decoding module sequentially determines whether the confidence information corresponding to each element is greater than a preset confidence threshold, decodes the elements whose confidence information is greater than the preset confidence threshold, and uses the sum of the first decoding input information and the first feedback channel additive Gaussian noise information of the feedback channel as the second encoding input information.

[0010] The encoding module increments the encoding round by one and encodes the second encoding input information, the first decoded identifier information, the first encoding output information, and the bit stream to be sent to obtain the second encoding output information. The second encoding output information is then summed with the second forward channel additive Gaussian noise information of the forward channel as the second decoding input information. The first decoded identifier information is used to indicate the elements in the first decoding input information that have been successfully decoded.

[0011] The decoding module decodes the information based on the confidence information corresponding to each element in the first decoding input information and the second decoding input information to determine the confidence information corresponding to each element in the second decoding input information. If the confidence information corresponding to all the elements is greater than a preset confidence threshold, the decoded elements are combined into output information and output.

[0012] Optionally, the process of generating the first decoded identification information includes:

[0013] The decoding module constructs initial identification information and initializes the value of each element in the initial identification information to a preset initial value. The length of the initial identification information is the same as the length of the first decoding input information.

[0014] The decoding module iterates through each element in the first decoding input information. For the current element, if the confidence information of the current element is greater than a preset confidence threshold, the value of the element at the position corresponding to the current element in the initial identification information is modified to a first value; otherwise, the value of the element at the position corresponding to the current element in the initial identification information is modified to a second value. The first value is used to indicate that the decoding has been completed, and the second value is used to indicate that the decoding has not been completed.

[0015] Optionally, the method of encoding the second encoded input information, the first decoded identifier information, the first encoded output information, and the bit stream to be transmitted to obtain the second encoded output information includes:

[0016] The difference between the second encoded input information and the first encoded output information is calculated as the first difference information;

[0017] The second combination information is determined based on the position of the successfully decoded element in the first decoded identifier information, the first difference information, the bit stream to be sent, and the first encoded output information.

[0018] The second combination of information is encoded to obtain the second encoded output information.

[0019] Optionally, after determining the confidence information corresponding to each element in the second decoded input information, the method further includes:

[0020] If there are elements with confidence information less than or equal to the preset confidence threshold, it is determined whether the encoding round has reached the preset encoding round threshold. If so, each element in the second decoding input information is decoded and synthesized into output information and then output. If not, the second encoding input information is generated and encoding continues.

[0021] Optionally, the encoding module includes: an encoding input unit, an encoding feature extraction unit, a first encoding self-attention unit, a second encoding self-attention unit, an encoding feedforward network unit, an encoding head unit, and an encoding output unit;

[0022] The step of encoding the second combined information to obtain the second encoded output information includes:

[0023] The encoding input unit inputs the second combined information into the encoding feature extraction unit;

[0024] The encoding feature extraction unit determines the target extraction structure according to the encoding round, uses the target extraction structure to extract features from the second combined information, and inputs the extracted encoding features into the first encoding self-attention unit;

[0025] The first encoding self-attention unit and the second encoding self-attention unit sequentially perform feature fusion on the extracted encoded features to obtain fused encoded features, and input the fused encoded features into the encoding feedforward network unit;

[0026] The encoding feedforward network unit enhances the fused encoding features to obtain enhanced encoding features, and inputs the enhanced encoding features into the encoding head unit;

[0027] The encoding head unit compresses the enhanced encoding features based on bi-objective optimization control to obtain compressed encoding features, and inputs the compressed encoding features into the encoding output unit;

[0028] The encoding output unit converts the compressed encoding features into the second encoding output information.

[0029] Optionally, determining the target extraction structure based on the encoding round includes:

[0030] Determine whether the number of encoding rounds is greater than or equal to a preset structural round threshold;

[0031] If not, the perceptual structure consisting of a preset number of linear layers based on a preset activation function will be used as the target extraction structure.

[0032] If so, then add another linear layer with a preset activation function to the perceptual structure consisting of a preset number of linear layers based on a preset activation function to obtain the target extraction structure.

[0033] Optionally, the decoding module includes: a decoding input unit, a decoding feature extraction unit, a first decoding self-attention unit, a second decoding self-attention unit, a third decoding self-attention unit, a decoding feedforward network unit, a decoding head unit, and a decoding output unit;

[0034] The step of performing confidence processing on the first decoded input information to determine the confidence information corresponding to each element in the first decoded input information includes:

[0035] After receiving the first decoding input information, the decoding input unit inputs the first decoding input information into the decoding feature extraction unit.

[0036] The decoding feature extraction unit uses the target extraction structure to extract features from the first decoding input information, and inputs the extracted decoding features into the first decoding self-attention unit;

[0037] The first decoding self-attention unit, the second decoding self-attention unit, and the third decoding self-attention unit sequentially perform feature fusion on the extracted decoding features to obtain fused decoding features, and input the fused decoding features into the decoding feedforward network unit;

[0038] The decoding feedforward network unit enhances the fused decoding features to obtain enhanced decoding features, and inputs the enhanced decoding features into the decoding head unit;

[0039] The decoding head unit determines the confidence information of each element in the enhanced decoding feature based on dual-objective optimization control, and outputs the confidence information corresponding to each element through the decoding output unit.

[0040] Optionally, the training process of the encoding module includes:

[0041] The initial encoding module performs multiple encoding rounds based on sample information, and at the end of each encoding round, the loss value of the encoding round is determined based on the confidence information corresponding to the encoding round.

[0042] The loss value of the initial encoding module is determined based on the loss values ​​of all encoding rounds;

[0043] The initial encoding module is iteratively updated based on its loss value, and the initial encoding module at the end of the iteration is used as the encoding module.

[0044] Secondly, this application provides a variable-length channel coding system with feedback, the variable-length channel coding system with feedback including an encoding module and a decoding module, the encoding module and the decoding module being used to perform the steps of the variable-length channel coding method with feedback as described in the first aspect.

[0045] Thirdly, this application provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the variable-length channel coding method with feedback as described in the first aspect.

[0046] The beneficial effects of this application are as follows: The encoding module encodes the bit stream to be transmitted to obtain first encoded output information. The sum of the first encoded output information and the first additive Gaussian noise information of the forward channel is used as second encoded input information. The decoding module performs confidence processing on the first encoded input information to determine the confidence information corresponding to each element in the first decoded input information. It sequentially judges whether the confidence information corresponding to each element is greater than a preset confidence threshold. Elements with confidence information greater than the preset confidence threshold are decoded. The sum of the first decoded input information and the first additive Gaussian noise information of the feedback channel is used as second encoded input information. The encoding module increments the encoding round by one and encodes according to the second encoded input information, the first decoded identifier information, the first encoded output information, and the bit stream to be transmitted to obtain the second encoded output information. In this embodiment, by encoding the second encoded input information, the first decoded identifier information, the first encoded output information, and the bit stream to be transmitted, the forward encoding strategy of the transmitting end is adjusted using feedback information. Then, the sum of the second encoded output information and the additive Gaussian noise information of the second forward channel is used as the second decoding input information. The decoding module performs confidence processing based on the confidence information corresponding to each element in the first decoding input information and the second decoding input information to determine the confidence information corresponding to each element in the second decoding input information. The decoding module can decode each bit block according to the noise level carried by the bit block corresponding to different elements, thereby realizing adaptive rate adjustment at the bit block granularity. If the confidence information corresponding to all elements is greater than a preset confidence threshold, the decoded elements are synthesized into output information and output. This embodiment improves noise immunity by iteratively communicating in multiple rounds through the forward channel and downlink enhanced feedback channel between the encoding and decoding modules. Attached Figure Description

[0047] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0048] Figure 1 This is a schematic diagram illustrating an application scenario of a variable-length channel coding method with feedback provided in an embodiment of this application.

[0049] Figure 2 This is a flowchart illustrating a variable-length channel coding method with feedback provided in this embodiment;

[0050] Figure 3 This is a schematic diagram of the architecture of a variable-length channel coding system with feedback provided in an embodiment of this application;

[0051] Figure 4 This is a schematic diagram of a process for obtaining second encoded output information provided in an embodiment of this application;

[0052] Figure 5 This is a schematic diagram of the structure of an encoding module provided in an embodiment of this application;

[0053] Figure 6 This is a schematic diagram of another process for obtaining the second encoded output information provided in an embodiment of this application;

[0054] Figure 7 This is a schematic diagram of a process for determining the target extraction structure provided in an embodiment of this application;

[0055] Figure 8 This is a schematic diagram of the structure of a decoding module provided in an embodiment of this application;

[0056] Figure 9 This is a flowchart illustrating a process for determining the confidence information corresponding to each element in the first decoded input information, provided in an embodiment of this application.

[0057] Figure 10 This is a schematic diagram of a coding module training process provided in an embodiment of this application;

[0058] Figure 11 This is a schematic diagram comparing the bit error rates of a channel coding method with feedback under different channel conditions and a variable-length channel coding method with feedback at different code rates, as provided in the embodiments of this application. Detailed Implementation

[0059] To make the objectives, technical solutions, and advantages of the embodiments of this application clearer, the technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. It should be understood that the accompanying drawings in this application are for illustrative and descriptive purposes only and are not intended to limit the scope of protection of this application. Furthermore, it should be understood that the schematic drawings are not drawn to scale. The flowcharts used in this application illustrate operations implemented according to some embodiments of this application. It should be understood that the operations in the flowcharts may not be implemented in sequence, and steps without logical contextual relationships may be reversed or implemented simultaneously. In addition, those skilled in the art, guided by the content of this application, may add one or more other operations to the flowcharts, or remove one or more operations from the flowcharts.

[0060] Furthermore, the described embodiments are merely some, not all, of the embodiments of this application. The components of the embodiments of this application described and illustrated herein can typically be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of the application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0061] It should be noted that the term "comprising" will be used in the embodiments of this application to indicate the presence of the features declared thereafter, but does not exclude the addition of other features.

[0062] In channel coding methods with feedback, the classical feedback channel model cannot dynamically adapt to the feedback process and optimize performance under changing channel conditions and receiver states. Deep learning-based channel coding methods are limited to fixed-length codes, which restricts their adaptability and prevents them from fully utilizing feedback information, resulting in low flexibility. Furthermore, this method performs poorly at high code rates.

[0063] Based on this, this application proposes a variable-length channel coding method with feedback. This method is applied to the coding module and the decoding module in a variable-length channel coding system with feedback. The coding module encodes the received information and uses the sum of the encoded output information and the additive Gaussian noise information of the forward channel as the decoding input information. The decoding module performs confidence processing on the decoding input information to obtain the confidence information corresponding to each element in the decoding input information, and sequentially judges whether the confidence information of each element is greater than a preset confidence threshold. If so, the element is decoded; otherwise, the decoding input information and the first decoded identifier information are summed with the noise information of the feedback channel as the coding input information. This completes the first round of encoding and decoding, and initiates the second round: the encoding module encodes the input information, the first decoded identifier, and the output information from the previous round to obtain the current round's output information. The current round's output information is then summed with the noise information from the forward channel as the current round's input information. The encoding module performs confidence processing on the input information to determine the confidence level of each element. It then checks if the confidence level is greater than a preset threshold. If the confidence level of all elements is greater than the preset threshold, all decoded elements are combined into output information and output. This application performs confidence processing / decoding on elements based on the comparison between the confidence level of each element in the decoded input information and a preset execution threshold. Based on the element positions corresponding to confidence levels not exceeding the preset execution threshold, the decoded input information is fed back to the encoding module for continued encoding and decoding. This allows for the decoding of each bit block according to the noise level carried by different bit blocks, thereby achieving adaptive rate adjustment.

[0064] Next, refer to Figure 1 The application scenarios of the variable-length channel coding method with feedback in this application are introduced. Figure 1 This is a schematic diagram illustrating an application scenario of a variable-length channel coding method with feedback provided in an embodiment of this application. For example... Figure 1 As shown, the variable-length channel coding method with feedback is applied to a variable-length channel coding system with feedback, which includes an encoding module and a decoding module. The encoding module can be deployed at the signal transmitting terminal, and the decoding module can be deployed at the base station. The base station can realize signal reception and forwarding. The transmission channel from the terminal to the base station is the forward channel, and the transmission channel from the base station to the terminal is the feedback channel. This method can be applied in environments with poor channel conditions, such as mines, to expand the signal coverage by enhancing the downlink feedback method. It can also be applied in places where it is difficult to replace equipment, such as in forests and high-rise buildings, thereby providing low-power communication.

[0065] Next, refer to Figure 2The specific implementation process of the variable-length channel coding method with feedback is introduced. Figure 2 This is a flowchart illustrating a variable-length channel coding method with feedback provided in this embodiment.

[0066] S201, The encoding module encodes the bit stream to be transmitted to obtain the first encoded output information, and uses the sum of the first encoded output information and the first additive Gaussian noise information of the forward channel as the first decoding input information. The bit stream to be transmitted includes multiple bit blocks.

[0067] Optionally, the bit stream to be transmitted may include multiple bit blocks. Encoding the bit stream to be transmitted yields first encoded output information, where the first encoded output information includes information corresponding to multiple bits. It is worth noting that the bit stream to be transmitted is the information received by the encoding module in the first encoding round. An encoding round is also a communication round, and encoding and decoding are performed once in one communication round.

[0068] Specifically, the encoding module encodes the bit stream to be sent using the following formula (1):

[0069] x 1 =φ(b), b∈(0,1) m (1)

[0070] Where, x 1 This represents the first encoded output information obtained in the first encoding round, where the first encoded output information contains elements corresponding to each bit block, b represents the bit stream to be sent, φ represents the mathematical model of the encoding module, and m is the size of a single bit block.

[0071] It is worth mentioning that in formula (1) and the following formulas, the superscript symbol indicates the coding round, and the subscript symbol indicates the position of the element in the information it belongs to.

[0072] Optionally, channel noise varies depending on the signal transmission environment; the worse the transmission environment, the greater the channel noise. Before signal transmission, noise in the forward and feedback channels can be detected to obtain additive Gaussian noise (AGN) information for both the forward and feedback channels. It's worth noting that while the AGN information for the forward and feedback channels may differ across coding rounds, their variances are the same.

[0073] Optionally, the sum of the first encoded output information and the first forward channel additive Gaussian noise information of the forward channel can be used as the first decoded input information, which can be expressed by the following formula (2):

[0074]

[0075] Among them, y 1 Indicates the first decoded input information, x 1 Indicates the first encoded output information, n 1 n represents the additive Gaussian noise information of the first forward channel. 1 obey The normal distribution, σ ff This represents the variance of the additive Gaussian noise in the forward channel.

[0076] S202. The decoding module performs confidence processing on the first decoding input information to determine the confidence information corresponding to each element in the first decoding input information, wherein each element corresponds to a bit block in the bit stream to be sent.

[0077] Optionally, in the first encoding round, an equal initialization strategy can be used to generate the confidence information corresponding to each element.

[0078] S203. The decoding module sequentially determines whether the confidence information corresponding to each element is greater than the preset confidence threshold, decodes the elements whose confidence information is greater than the preset confidence threshold, and uses the sum of the first decoding input information and the first feedback channel additive Gaussian noise information of the feedback channel as the second encoding input information.

[0079] Optionally, the decoding module pre-acquires a preset confidence threshold, and then sequentially determines whether the confidence information corresponding to each element is greater than the preset confidence threshold. Specifically, it determines whether the element with the highest probability among the confidence information corresponding to each element is greater than the preset confidence threshold.

[0080] Optionally, after decoding elements with confidence levels greater than a preset threshold, the decoding module determines whether there are any elements whose confidence levels are not greater than the preset threshold. If so, the sum of the first decoding input information and the additive Gaussian noise information of the first feedback channel is used as the second encoding input information. The first decoding input information can serve as feedback information for the first decoding module.

[0081] Optionally, the decoding module may also generate first decoded identifier information based on the position of the decoded element in the first decoding input information, and send the first decoded identifier information to the encoding module.

[0082] Optionally, the sum of the first decoded input information and the additive Gaussian noise information of the first feedback channel can be used as the second encoded input information, which can be expressed by the following formula (3):

[0083]

[0084] in, This represents the second encoded input information, y1 This represents the first decoded input information. This represents the additive Gaussian noise information of the first feedback channel. obey The normal distribution, σ fb This represents the variance of the additive Gaussian noise in the feedback channel.

[0085] S204. The encoding module increments the encoding round by one and encodes the second encoding input information, the first decoded identifier information, the first encoding output information, and the bit stream to be transmitted to obtain the second encoding output information. The second encoding output information is then summed with the additive Gaussian noise information of the second forward channel as the second decoding input information. The first decoded identifier information is used to indicate the elements in the first decoding input information that have been successfully decoded.

[0086] Optionally, after receiving the second encoded input information, the encoding module increments the encoding round number by one. It is worth noting that the encoding module increments the encoding round number by one each time it receives encoded input information from the decoding module.

[0087] Optionally, in the second encoding round, the encoding module encodes the second encoded input information, the first decoded identifier information, the first encoded output information, and the bit stream to be sent to obtain the second encoded output information. In the T-th encoding round, the encoding module encodes the T-th encoded output information, all encoded output information from previous encoding rounds, the T-th decoded identifier information, all encoded output information from previous encoding rounds, and the bit stream to be sent to obtain the second encoded output information. The encoding process of the τ-th encoding round is described below using formulas (4) and (5). Wherein, τ is a number greater than 1 and less than T.

[0088]

[0089] x τ =φ(q) τ (5)

[0090] in, This indicates that the encoding module receives the element at position k in the τth encoding round, b k For the bit block at position k of the bit stream to be sent, This represents the element at position k in the first encoded output information. This represents the element at position k in the output information of the (τ-1)th encoding round. q represents the element at position k in the encoded input information. τ x represents the τ-th combination of information at all positions of the input of the τ-th encoding module. τ Encode the output information for the τth encoding round.

[0091] Optionally, the first decoded identifier information can indicate the position of the successfully decoded element in the first decoded input information through the values ​​of each element in the information. This position can correspond to the T-th encoded output information, all encoded output information of previous encoding rounds, the T-th decoded identifier information, all encoded output information of previous encoding rounds, and the position of each element in the bit stream to be sent. Specifically, the first decoded identifier information can be used to prevent the element at the position corresponding to the decoded bit block in the information received by the encoding module from participating in forward transmission and feedback, and simultaneously terminate the update of this position in the encoded output information and decoded output information in the next encoding round.

[0092] S205. The decoding module performs confidence processing based on the confidence information corresponding to each element in the first decoding input information and the second decoding input information to determine the confidence information corresponding to each element in the second decoding input information. If the confidence information corresponding to all elements is greater than the preset confidence threshold, the decoded elements are combined into output information and output.

[0093] Specifically, in the second encoding round, confidence processing is performed based on the confidence information corresponding to each element in the first decoding input information and the second decoding input information to determine the confidence information corresponding to each element in the second decoding input information. In the τth encoding round, confidence processing is performed based on the confidence information corresponding to each element in the (τ-1)th decoding input information and the τth decoding input information to determine the confidence information corresponding to each element in the τth decoding input information. For example, the process of determining the confidence information corresponding to each element in the τth decoding input information can be represented by formulas (6), (7), and (8).

[0094]

[0095] in, This represents the confidence information corresponding to the k-th element in the τ-th decoded input information. This represents the confidence level of the class of the i-th element corresponding to the k-th element. ψ represents the object encoded by the decoding module in the second encoding round, and ψ represents the mathematical model of the decoding module.

[0096] Optionally, if the confidence information corresponding to all elements is greater than a preset confidence threshold, the decoded elements are synthesized into output information and output. If the element with the highest probability among the confidence information corresponding to an element is not greater than the preset confidence threshold, the sum of the second decoding input information and the additive Gaussian noise information of the second feedback channel is used as the third encoding input information. The encoding module increments the encoding round by one, enters the third encoding round, and encodes the third encoding input information, the second encoding input information, the second decoded identifier information, the first encoding output information, the second encoding output information, and the bit stream to be transmitted to obtain the third encoding output information. The sum of the third encoding output information and the additive Gaussian noise information of the third forward channel is used as the third decoding input information. The decoding module performs confidence processing based on the confidence information corresponding to each element in the second decoding input information and the third decoding input information to determine the confidence information corresponding to each element in the third decoding input information and to determine whether the confidence information corresponding to each element is greater than the preset confidence threshold. The above is the complete encoding and decoding process of the third encoding round. When the confidence information corresponding to all elements is greater than the preset confidence threshold, the decoded elements are combined into output information and output. If the confidence information corresponding to an element is not greater than the preset confidence threshold, the encoding and decoding of the fourth encoding round will begin. This process continues until the confidence information corresponding to all elements is greater than the preset confidence threshold. Then, the decoded elements are combined into output information and output, thus completing the adaptive bitrate adjustment.

[0097] The following is the encoding and decoding process for the fourth encoding round: If, in the third encoding round, the decoding module determines that the element with the highest probability in the confidence information corresponding to an element is not greater than a preset confidence threshold, then the sum of the third decoding input information and the additive Gaussian noise information of the third feedback channel is used as the fourth encoding input information. The encoding module increments the encoding round by one, enters the fourth encoding round, and encodes the fourth encoding input information, the third encoding input information, the second encoding input information, the third decoded identifier information, the first encoding output information, the second encoding output information, the third encoding output information, and the bit stream to be transmitted to obtain the fourth encoding output information. The sum of the fourth encoding output information and the additive Gaussian noise information of the fourth forward channel is used as the fourth decoding input information. The decoding module performs confidence processing based on the confidence information corresponding to each element in the third decoding input information and the fourth decoding input information to determine the confidence information corresponding to each element in the fourth decoding input information, and determines whether the confidence information corresponding to each element is greater than the preset confidence threshold. The above is the complete encoding and decoding process of the fourth encoding round. When the confidence information corresponding to all elements is greater than the preset confidence threshold, the decoded elements are combined into output information and output. If the confidence information corresponding to an element is not greater than the preset confidence threshold, the fifth encoding round of encoding and decoding is entered until the confidence information corresponding to all elements is greater than the preset confidence threshold.

[0098] In this embodiment, the encoding module encodes the bitstream to be transmitted to obtain first encoded output information. The sum of the first encoded output information and the additive Gaussian noise information of the first forward channel is used as second encoded input information. The decoding module performs confidence processing on the first encoded input information to determine the confidence information corresponding to each element in the first decoded input information. It sequentially checks whether the confidence information corresponding to each element is greater than a preset confidence threshold. Elements with confidence information greater than the preset confidence threshold are decoded. The sum of the first decoded input information and the additive Gaussian noise information of the first feedback channel is used as second encoded input information. The encoding module increments the encoding round by one and encodes the second encoded input information, the first decoded identifier information, the first encoded output information, and the bitstream to be transmitted to obtain the second encoded output information. In this embodiment, by encoding the second encoded input information, the first decoded identifier information, the first encoded output information, and the bitstream to be transmitted, the forward encoding strategy at the transmitting end is adjusted using feedback information. Then, the sum of the second encoded output information and the additive Gaussian noise information of the second forward channel is used as the second decoding input information. The decoding module performs confidence processing based on the confidence information corresponding to each element in the first decoding input information and the second decoding input information to determine the confidence information corresponding to each element in the second decoding input information. The decoding module can decode each bit block according to the noise level carried by the bit block corresponding to different elements, thereby realizing adaptive rate adjustment at the bit block granularity. If the confidence information corresponding to all elements is greater than a preset confidence threshold, the decoded elements are synthesized into output information and output. This embodiment improves noise immunity by iteratively communicating in multiple rounds through the forward channel and downlink enhanced feedback channel between the encoding and decoding modules.

[0099] Next, the process of generating the first decoded identifier information in step S204 above will be described.

[0100] Optionally, the decoding module constructs initial identification information and initializes the values ​​of each element in the initial identification information to preset initial values. The length of the initial identification information is the same as the length of the first decoding input information.

[0101] Optionally, the preset initial value can be, for example, 0, in which case the value of each element in the initial identification information is 0.

[0102] Optionally, the decoding module iterates through each element in the first decoded input information. For the current element, if the confidence information of the current element is greater than a preset confidence threshold, the value of the element at the position corresponding to the current element in the initial identification information is modified to the first value; otherwise, the value of the element at the position corresponding to the current element in the initial identification information is modified to the second value. The first value is used to indicate that the element has been decoded, and the second value is used to indicate that the element has not been decoded.

[0103] For example, if the first decoded input information includes three elements, and the element with the highest probability in the confidence information corresponding to the third element is greater than a preset confidence threshold, then the value of the third element in the initial identification information is modified to a first value, and the values ​​of the first and second elements are modified to second values. For example, the first value can be 1, and the second value can be 0. Then the first decoded identification information can be (0,0,1).

[0104] Optionally, after the decoding module generates the first decoded identification information, it can send both the first decoded identification information and the first decoded output information to the encoding module through the feedback channel. During this process, the sum of the first decoded input information and the first feedback channel additive Gaussian noise information of the feedback channel is used as the second encoding input information.

[0105] In this embodiment, by constructing initial identification information and modifying the value of the corresponding position of the initial identification information according to the comparison between the confidence information corresponding to each element in the decoded input information and the preset confidence threshold, the element in the information received by the encoding module corresponding to the position of the element with the first value in the first decoded identification information no longer participates in forward transmission and feedback, and at the same time, the updating of the position of the encoded output information and the decoded output information in the next encoding round is terminated.

[0106] As an optional implementation method, Figure 3 This is a schematic diagram of the architecture of a variable-length channel coding system with feedback provided in an embodiment of this application. Figure 3 As shown, the encoding module and the decoding module are connected through a forward channel and a feedback channel. The encoding module processes the τ-th combination information q at all positions of the τ-th encoding module input. τ After encoding, output the encoded output information x. τ Then encode and output information x τ With the additive Gaussian noise information n of the τth forward channel τ The sum of these values ​​is used as the decoding input information y for the τth time step. τ The decoding module is based on the confidence information corresponding to each element in the (τ-1)th decoding input information. and the τth decoding input information y τ Perform confidence processing and analyze the confidence information corresponding to each element in the decoded output. The τth decoded identifier is generated by combining a preset threshold value with a decision. The decoded identifier information of the τth generation and the τth decoding input information y τ The encoder is input through the feedback channel, and during this process, the τ-th decoded input information y is... τ Additive Gaussian noise information in the τth feedback channel The sum is used as the τ+1th encoded input information.

[0107] Furthermore, referring to Figure 4 The specific process of encoding the second encoded output information based on the second encoded input information, the first decoded identifier information, the first encoded output information, and the bit stream to be sent in step S204 above is described below. Figure 4 This is a schematic diagram of a process for obtaining second encoded output information provided in an embodiment of this application.

[0108] S401. Calculate the difference between the second encoded input information and the first encoded output information as the first difference information.

[0109] Specifically, the difference between the encoded input information of the current encoding round and the encoded output information of the previous encoding round is used as the first difference information.

[0110] S402. Determine the second combination information based on the position of the successfully decoded element in the first decoded identifier information, the first difference information, the bit stream to be sent, and the first encoded output information.

[0111] Specifically, the value of each element in the first decoded identifier information determines whether the bit block corresponding to each element in the first difference information, the bit stream to be sent, and the first encoded output information is decoded. If the value of an element in the first decoded identifier information is the first value, then the element at the corresponding position in the first difference information, the bit stream to be sent, and the first encoded output information will not be encoded or decoded.

[0112] S403. Encode the second combination information to obtain the second encoded output information.

[0113] Optionally, taking the third encoding round as an example, the difference between the third encoding input information and the second encoding output information is calculated as the second difference information, and the difference between the second encoding input information and the first encoding output information is calculated as the first difference information. Based on the position of the successfully decoded element in the second decoded identifier information, the first difference information, the second difference information, the bit stream to be sent, the first encoding output information, and the second encoding output information, the third combination information is determined, and the third combination information is encoded to obtain the third encoding output information.

[0114] In this embodiment, by calculating the first difference information, and determining the second combination information based on the position of the successfully decoded element in the first decoded identifier information, the first difference information, the bit stream to be sent, and the first encoded output information, the second combination information is encoded to obtain the second encoded output information. Thus, the forward transmission and feedback of the element corresponding to the decoded bit block in the information are stopped by using the value of each element in the decoded identifier information.

[0115] As an optional implementation, after determining the confidence information corresponding to each element in the second decoding input information in step S205 above, the following steps can also be performed.

[0116] Optionally, if there are elements with confidence information less than or equal to a preset confidence threshold, it is determined whether the encoding round has reached the preset encoding round threshold. If so, each element in the second decoding input information is decoded and synthesized into output information before being output. If not, the second encoding input information is generated and encoding continues.

[0117] Optionally, if there are elements with confidence information less than or equal to a preset confidence threshold, it indicates that the bit block has not been fully decoded. Then, it is determined whether the encoding round has reached the preset encoding round threshold. If so, although the current bit rate adjustment of the bit block cannot achieve the best effect, the adjusted bit block will still be decoded and output to ensure signal transmission efficiency. If the encoding round has not reached the preset encoding round threshold, the sum of the second decoding input information and the additive Gaussian noise information of the second feedback channel is used as the second encoding input information. The encoding module increments the encoding round by one and enters the next round of encoding and decoding until all elements have confidence information greater than the preset confidence threshold, or until the encoding round has reached the preset encoding round threshold.

[0118] In this embodiment, signal transmission efficiency is ensured by determining whether the number of encoding rounds has reached a preset encoding round threshold.

[0119] Next, the structure of the encoding and decoding modules, as well as the specific encoding and decoding process, will be introduced.

[0120] Figure 5 This is a schematic diagram of the structure of an encoding module provided in an embodiment of this application. (Refer to...) Figure 5 The structure of the encoding module will be introduced.

[0121] Optionally, the encoding module includes an encoding input unit, an encoding feature extraction unit, a first encoding self-attention unit, a second encoding self-attention unit, an encoding feedforward network unit, an encoding head unit, and an encoding output unit.

[0122] Next, refer to Figure 6 The specific steps in step S403 above, where the second combined information is encoded to obtain the second encoded output information, are described below. Figure 6 This is a schematic diagram of another process for obtaining the second encoded output information provided in an embodiment of this application.

[0123] S601, The encoding input unit inputs the second combination information into the encoding feature extraction unit.

[0124] S602, the coding feature extraction unit determines the target extraction structure according to the coding round, uses the target extraction structure to extract features from the second combination information, and inputs the extracted coding features into the first coding self-attention unit.

[0125] As an optional implementation, the coding feature extraction unit determines the target extraction structure based on the current coding round and a preset structure round threshold. The depth of the target extraction structure can be adjusted according to the current coding round.

[0126] Optionally, the encoding feature extraction unit extracts features from the second combined information and projects the feature matrix to a higher dimension to form extracted encoded features.

[0127] S603, the first encoding self-attention unit and the second encoding self-attention unit sequentially perform feature fusion on the extracted encoded features to obtain fused encoded features, and input the fused encoded features into the encoding feedforward network unit.

[0128] Optionally, the first encoding self-attention unit and the second encoding self-attention unit can fuse feature representations between bit blocks and achieve parallel processing.

[0129] Optionally, the first and second encoding self-attention units can map the fused encoded features to three vectors with the same dimension d. model Given different projection spaces, three sets of feature matrices are generated, denoted as Q, K, and V, respectively. The feature matrices then satisfy... These represent the query, key, and value in the attention mechanism, respectively.

[0130] For example, the first encoding self-attention unit and the second encoding self-attention unit can be represented by the following formula (9):

[0131]

[0132] S604. The coding feedforward network unit enhances the fused coding features to obtain enhanced coding features, and then inputs the enhanced coding features into the coding head unit.

[0133] Optionally, the coding feedforward network unit can project the fused coding features into a higher-dimensional nonlinear space, thereby enhancing the robustness of the error-correcting code.

[0134] Optionally, the encoding feedforward network unit can consist of two linear layers with a large intrinsic dimension and an activation function. The activation function can be the Gelu activation function.

[0135] Alternatively, the coding feedforward network unit can be represented by the following formula (10):

[0136]

[0137] Where FFN(X) represents the post-enhanced encoded features, and f denotes an upsampled linear projection. ′ This represents a downsampled linear projection.

[0138] S605, the encoding head unit compresses the enhanced encoding features based on dual-objective optimization control to obtain compressed encoding features, and inputs the compressed encoding features into the encoding output unit.

[0139] Optionally, the coding head unit can consist of two linear layers and a power control layer, thereby compressing the hidden features of the coding feedforward network unit into a compressed coding feature.

[0140] The dual-objective optimization control can be expressed as follows: Equation (11) and Equation (12):

[0141]

[0142] Among them, P e express The ratio, where b is the bit stream to be sent. To output information, The average number of encoding rounds for all bit blocks is represented by formula (12), which is a constraint condition. This indicates that power control is applied to all encoded transmitted information.

[0143] S606, the encoding output unit converts the compressed encoding features into second encoding output information.

[0144] Optionally, the encoding output unit converts the compressed encoding features into second encoded output information, which is then transmitted over the forward channel.

[0145] In this embodiment, variable-length encoding is achieved by encoding the bit stream or combined information to be transmitted through the encoding input unit, encoding feature extraction unit, first encoding self-attention unit, second encoding self-attention unit, encoding feedforward network unit, encoding head unit, and encoding output unit in the encoding module.

[0146] Furthermore, referring to Figure 7 The specific steps in step S602 above, which involve determining the target extraction structure based on the coding rounds, will be described. Figure 7 This is a schematic diagram of a process for determining the target extraction structure provided in an embodiment of this application.

[0147] S701. Determine whether the encoding round is greater than or equal to the preset structure round threshold.

[0148] Optionally, based on the DeepCode experimental results, the following conclusions can be drawn: as the number of encoding rounds increases, the nonlinear relationship between the input and output of the encoding module becomes stronger. Specifically, if the number of encoding rounds is less than the preset structural round threshold, the nonlinear relationship is not strong, so a shallower neural network can be used for fitting; if the number of encoding rounds is greater than or equal to the preset structural round threshold, the nonlinear relationship is strong, so a deeper neural network is used for fitting. Based on this, the following steps are performed.

[0149] S702. If not, then the perceptual structure consisting of a preset number of linear layers based on a preset activation function will be used as the target extraction structure.

[0150] Specifically, if the number of encoding rounds is less than a preset structural rounds threshold, then a perceptual structure consisting of a preset number of linear layers based on a preset activation function is used as the target extraction structure. The activation function can be a ReLU activation function, and the perceptual structure can be a Multilayer Perceptron (MLP). That is, an MLP consisting of a preset number of linear layers with ReLU activation functions is used as the target extraction structure.

[0151] S703. If so, then add a linear layer with a preset activation function to the perceptual structure consisting of a preset number of linear layers based on a preset activation function to obtain the target extraction structure.

[0152] Specifically, if the number of encoding rounds is greater than or equal to a preset structural round threshold, then an additional linear layer with a preset activation function is added to the perceptual structure consisting of a preset number of linear layers based on a preset activation function. That is, an additional linear layer with a ReLU activation function is added to the MLP consisting of a preset number of linear layers with a ReLU activation function, and this structure is used as the target extraction structure.

[0153] For example, the target extraction structure can be determined by the following formulas (13) and (14):

[0154]

[0155] in, Let l represent the l-th linear layer, and σ represent the activation function. This represents the number of intermediate neural network layers at the τth encoding round. The number of layers is n when the encoding round is less than a preset structural round threshold, and n+1 when the encoding round is greater than or equal to the preset structural round threshold. ξ represents upsampling projection. This represents the downsampling projection. The target extraction structure represents the structure.

[0156] In this embodiment, the target extraction structure is determined based on the encoding round and the preset structure round threshold, thereby adjusting the nonlinearity of the target extraction structure according to the encoding round, so as to capture the features of the combined information more accurately.

[0157] Next, refer to Figure 8 The structure of the decoding module will be introduced. Figure 8 This is a schematic diagram of the structure of a decoding module provided in an embodiment of this application. Figure 8 As shown, the decoding module includes: a decoding input unit, a decoding feature extraction unit, a first decoding self-attention unit, a second decoding self-attention unit, a third decoding self-attention unit, a decoding feedforward network unit, a decoding head unit, and a decoding output unit.

[0158] based on Figure 8 The structure of the decoding module shown below will be referred to next. Figure 9 The specific process of performing confidence processing on the first decoded input information and determining the confidence information corresponding to each element in the first decoded input information in step S202 above will be described. Among them, Figure 9 This is a flowchart illustrating a process for determining the confidence information corresponding to each element in the first decoded input information, as provided in an embodiment of this application.

[0159] S901. After receiving the first decoding input information, the decoding input unit inputs the first decoding input information into the decoding feature extraction unit.

[0160] S902, the decoding feature extraction unit uses the target extraction structure to extract features from the first decoding input information and inputs the extracted decoding features into the first decoding self-attention unit.

[0161] Optionally, the decoding feature extraction unit uses the target extraction structure to extract features from the first decoding input information and projects them to a higher dimension to generate extracted decoding features.

[0162] S903, the first decoding self-attention unit, the second decoding self-attention unit and the third decoding self-attention unit sequentially perform feature fusion on the extracted decoding features to obtain fused decoding features, and input the fused decoding features into the decoding feedforward network unit.

[0163] Optionally, the first decoding self-attention unit, the second decoding self-attention unit, and the third decoding self-attention unit aggregate the features between the elements corresponding to each bit block to generate the extracted decoding features.

[0164] S904. The decoding feedforward network unit enhances the fused decoding features to obtain enhanced decoding features, and inputs the enhanced decoding features into the decoding head unit.

[0165] S905, the decoding head unit determines the confidence information of each element in the enhanced decoding feature based on dual-objective optimization control, and outputs the confidence information corresponding to each element through the decoding output unit.

[0166] Optionally, the decoding header unit can be a confidence classification header unit, which can output confidence information for each element corresponding to each bit block. When the confidence information corresponding to an element is greater than a preset confidence threshold, the element is decoded.

[0167] In this embodiment, the decoding input unit, decoding feature extraction unit, first decoding self-attention unit, second decoding self-attention unit, third decoding self-attention unit, decoding feedforward network unit, decoding head unit, and decoding output unit in the decoding module are used to achieve decoding of the decoding input information and confidence judgment.

[0168] Next, refer to Figure 10 The training process of the encoding module is described. Among other things, Figure 10 This is a schematic diagram of the training process of an encoding module provided in an embodiment of this application.

[0169] S1001. The initial encoding module performs multiple encoding rounds based on the sample information, and at the end of each encoding round, the loss value of the encoding round is determined according to the confidence information corresponding to the encoding round.

[0170] Optionally, during training, the forward channel environment of additive white Gaussian noise components under different signal-to-noise ratios can be simulated, while the feedback channel adopts a noise-free channel environment.

[0171] Optionally, m*K integer values ​​can be uniformly and randomly sampled from 0 and 1, and divided into K bit blocks to form a training sample. Here, m and K can be random numbers.

[0172] S1002. Determine the loss value of the initial encoding module based on the loss values ​​of all encoding rounds.

[0173] Specifically, the sum of the loss values ​​from all encoding rounds is used as the loss value of the initial encoding module.

[0174] S1003. Iteratively update the initial encoding module based on the loss value of the initial encoding module, and use the initial encoding module at the end of the iteration as the encoding module.

[0175] Specifically, the loss function can be expressed by the following formula (15):

[0176]

[0177] Where L is the loss function, T is the preset encoding round threshold, and γ τ-1This is the exponential weighting coefficient, which increases exponentially with the number of coding rounds.

[0178] Optionally, the training process of the decoding module is the same as that of the encoding module, and will not be described again here.

[0179] In this embodiment, the loss function is defined by the exponential weighted sum of the confidence cross-entropy loss of multiple coding rounds, which is more suitable for channel coding optimization.

[0180] in addition, Figure 11 This diagram illustrates a comparison of the bit error rates (BER) of a general block attention feedback (BFF) channel coding method and a variable-length BFF channel coding method under different channel conditions at different code rates, as provided in embodiments of this application. Figure 11 As shown, under channel conditions with a signal-to-noise ratio of 0dB and 1dB, the variable-length channel coding method with feedback in this embodiment achieves a significantly lower bit error rate than the traditional coding method. This reflects that the method improves the short-code performance of the feedback code.

[0181] This application also provides a variable-length channel coding system with feedback. The variable-length channel coding system with feedback includes an encoding module and a decoding module, which are used to perform the steps of the variable-length channel coding method with feedback described above.

[0182] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, performs the steps of the above-described variable-length channel coding method with feedback.

[0183] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and devices described above can be referred to the corresponding processes in the method embodiments, and will not be repeated here. In the several embodiments provided in this application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. Furthermore, multiple modules or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed mutual coupling or direct coupling or communication connection can be through some communication interfaces; the indirect coupling or communication connection of devices or modules can be electrical, mechanical, or other forms.

[0184] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. If the functions are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0185] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

Claims

1. A variable length channel coding method with feedback, characterized by, The method is applied to a variable-length channel coding system with feedback, the variable-length channel coding system with feedback comprising an encoding module and a decoding module, a transmission channel of the encoding module to the decoding module being a forward channel, a transmission channel of the decoding module to the encoding module being a feedback channel, and the method comprising: The encoding module encodes a to-be-sent bit stream to obtain first encoding output information, and adds the first encoding output information and first forward channel additive Gaussian noise information of the forward channel to obtain first decoding input information, the to-be-sent bit stream comprising a plurality of bit blocks; The decoding module performs confidence processing on the first decoding input information to determine confidence information corresponding to each element in the first decoding input information, wherein each element corresponds to a bit block in the to-be-sent bit stream. The decoding module sequentially determines whether the confidence information corresponding to each element is greater than a preset confidence threshold, decodes an element whose confidence information is greater than the preset confidence threshold, and adds the first decoding input information and first feedback channel additive Gaussian noise information of the feedback channel to obtain second encoding input information. The encoding module increases an encoding round by one, and encodes the second encoding input information, first decoded identification information, the first encoding output information, and the to-be-sent bit stream to obtain second encoding output information, and adds the second encoding output information and second forward channel additive Gaussian noise information of the forward channel to obtain second decoding input information, wherein the first decoded identification information is used to indicate an element in the first decoding input information that has been successfully decoded. The decoding module performs confidence processing on the second decoding input information according to the confidence information corresponding to each element in the first decoding input information to determine confidence information corresponding to each element in the second decoding input information, and if the confidence information corresponding to all the elements is greater than the preset confidence threshold, combines the decoded elements into output information and outputs the output information.

2. The variable-length channel coding method with feedback according to claim 1, characterized by, The generation process of the first decoded identification information comprises: The decoding module constructs initial identification information, and initializes values of elements in the initial identification information to preset initial values, a length of the initial identification information being the same as a length of the first decoding input information. The decoding module traverses each element in the first decoding input information, and for a current element that is traversed, if confidence information of the current element is greater than a preset confidence threshold, modifies a value of an element at a position corresponding to the current element in the initial identification information to a first value, otherwise, modifies the value of the element at the position corresponding to the current element in the initial identification information to a second value, wherein the first value is used to indicate that decoding has been performed, and the second value is used to indicate that decoding has not been performed.

3. The variable-length channel coding method with feedback according to claim 1, characterized by, The encoding of the second encoding input information, the first decoded identification information, the first encoding output information, and the to-be-sent bit stream to obtain the second encoding output information comprises: calculating a difference between the second encoding input information and the first encoding output information as first difference information; determining second combination information according to positions of elements successfully decoded in the first decoded identification information, the first difference information, the bit stream to be sent, and the first encoding output information; encoding the second combination information to obtain second encoding output information.

4. The variable-length channel coding method with feedback according to claim 1, characterized by, After the confidence information corresponding to each element in the second decoding input information is determined, the method further includes: If there is an element with confidence information less than or equal to the preset confidence threshold, it is determined whether the encoding round reaches a preset encoding round threshold, if yes, the elements in the second decoding input information are decoded and combined into output information and then output; if no, the second encoding input information is generated and the encoding is continued.

5. The variable-length channel coding method with feedback according to claim 3, characterized by, The encoding module includes an encoding input unit, an encoding feature extraction unit, a first encoding self-attention unit, a second encoding self-attention unit, an encoding feedforward network unit, an encoding head unit, and an encoding output unit. The encoding of the second combination information to obtain the second encoding output information includes: The encoding input unit inputs the second combination information to the encoding feature extraction unit; The encoding feature extraction unit determines a target extraction structure according to the encoding round, and extracts features of the second combination information by using the target extraction structure, and inputs the extracted encoding features to the first encoding self-attention unit; The first encoding self-attention unit and the second encoding self-attention unit sequentially perform feature fusion on the extracted encoding features to obtain fused encoding features, and input the fused encoding features to the encoding feedforward network unit; The encoding feedforward network unit performs feature enhancement on the fused encoding features to obtain enhanced encoding features, and inputs the enhanced encoding features to the encoding head unit; The encoding head unit compresses the enhanced encoding features based on double-target optimization control to obtain compressed encoding features, and inputs the compressed encoding features to the encoding output unit; The encoding output unit converts the compressed encoding features into the second encoding output information.

6. The variable-length channel coding method with feedback according to claim 5, characterized by, The determination of the target extraction structure according to the encoding round includes: It is determined whether the encoding round is greater than or equal to a preset structure round threshold; If no, a perception structure of a preset number of layers composed of linear layers based on a preset activation function is used as the target extraction structure; If yes, a layer of linear layers based on a preset activation function is added to the perception structure of the preset number of layers composed of linear layers based on a preset activation function to obtain the target extraction structure.

7. The variable-length channel coding method with feedback according to claim 5, characterized by, The decoding module includes a decoding input unit, a decoding feature extraction unit, a first decoding self-attention unit, a second decoding self-attention unit, a third decoding self-attention unit, a decoding feedforward network unit, a decoding head unit, and a decoding output unit. The confidence processing on the first decoding input information to determine the confidence information corresponding to each element in the first decoding input information includes: The decoding input unit receives the first decoding input information and inputs the first decoding input information to the decoding feature extraction unit; The decoding feature extraction unit extracts features from the first decoding input information by using the target extraction structure, and inputs the extracted decoding features to the first decoding self-attention unit; The first decoding self-attention unit, the second decoding self-attention unit and the third decoding self-attention unit sequentially perform feature fusion on the extracted decoding features to obtain fused decoding features, and input the fused decoding features to the decoding feedforward network unit; The decoding feedforward network unit performs feature enhancement on the fused decoding features to obtain enhanced decoding features, and inputs the enhanced decoding features to the decoding head unit; The decoding head unit determines confidence information of each element in the enhanced decoding features based on double-target optimization control, and outputs the confidence information corresponding to each element through the decoding output unit.

8. The variable-length channel coding method with feedback according to claim 1, characterized by, The training process of the encoding module includes: The initial encoding module performs multiple encoding rounds based on sample information, and determines the loss value of each encoding round according to the confidence information corresponding to the encoding round at the end of each encoding round; Determine the loss value of the initial encoding module according to the loss values of all encoding rounds; Iteratively update the initial encoding module according to the loss value of the initial encoding module, and take the initial encoding module at the end of iteration as the encoding module.

9. A variable length channel coding system with feedback, characterized by The variable-length channel coding system with feedback includes an encoding module and a decoding module, and the encoding module and the decoding module are used to perform the steps of the variable-length channel coding method with feedback as claimed in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer program is run by the processor to perform the steps of the variable-length channel coding method with feedback as claimed in any one of claims 1-8.

Images