Multi-step decoding judgment accelerated decoding method and device and storage medium

Abstract

The invention relates to the technical field of communication, and discloses a multi-step decoding judgment acceleration decoding method and device and a storage medium, and the method comprises the steps: for linear block code decoding based on a neural network learning algorithm, calculating a component addition of a multi-iteration decoding result as a selection basis of a learning rate; based on different code lengths, code rates and channel conditions, optimal adjustment of the weight of a neural network learning algorithm is realized by setting iterative decoding result component times and addition coefficients, so that the learning rate is improved. According to the method, component addition of multiple iterative decoding results is used as the basis of learning rate selection, optimal adjustment of neural network weights under different code lengths, code rates and different channel conditions is realized by setting the selection of the iterative decoding result component times and addition coefficients, and thus the training learning rate is improved.

Description

Technical Field

[0001] The present invention relates to the field of communication technologies, and in particular, to a multi-step decoding decision acceleration decoding method, device, and storage medium. Background Art

[0002] With the rapid development of communication, which involves supporting high speed, low latency, high reliable transmission, etc., for example, autonomous driving, etc. The as-correct-as-possible transmission of information has always been the main purpose of the development of communication technologies. As an indispensable anti-interference technology, channel error correction code technology, linear block codes are widely used in communication systems to ensure the correct transmission of information due to their low-complexity performance. The learning algorithms of neural networks have proven their powerful classification and fitting capabilities in application scenarios such as speech, image, and natural language processing. Combining the neural network learning algorithm with linear block code decoding has proven that the performance can be improved compared to the original decoding algorithm. Especially for some linear block codes that were difficult to use soft decision iterative decoding in the past, good decoding performance has been achieved after using the neural network learning algorithm for decoding.

[0003] The existing technologies focus on using neural networks to construct encoding and decoding systems, and basically perform decoding processing according to the conventional processing method, lacking targeted research on improving the efficiency of decision processing in the decoding process, thus forming related technologies. For the linear block code decoding using the neural network learning algorithm, as the length of the linear block code increases, its training convergence speed becomes very slow, which makes the initial model training and subsequent application enhancement training time costs very large. Summary of the Invention

[0004] To solve the above problems, the present invention proposes a multi-step decoding decision acceleration decoding method, device, and storage medium, which can improve the training learning rate and is applicable to a wide range of linear block codes.

[0005] The technical solution adopted by the present invention is as follows: A multi-step decoding decision acceleration decoding method, comprising: For the linear block code decoding based on the neural network learning algorithm, calculate the component addition of the decoding results of multiple iterations as the basis for selecting the learning rate; Based on different code lengths, code rates, and channel conditions, by setting the number of component times and addition coefficients of the iterative decoding results, realize the optimal adjustment of the weights of the neural network learning algorithm, thereby improving the learning rate.

[0006] Further, it includes the following steps: Step 1, set the initial step size of the weights of the neural network learning algorithm ; Step 2, calculate the output after the t-th iteration of the linear block code decoding based on the neural network learning algorithm , and determine whether the decoding is successful. If it is successful, end; otherwise, execute the next step; Step 3: Calculate the loss function of cross-entropy in the t-th iteration , where is a binary block code; Step 4: Based on the loss function of cross-entropy , calculate the accumulated value in the t-th iteration ; Step 5: Based on the accumulated value perform step size adjustment to obtain the updated step size ; Step 6: Based on the updated step size adjust the weights of the neural network learning algorithm, execute the next iteration and go to Step 2.

[0007] Furthermore, the weights of the neural network learning algorithm include:

[0008] Among them, is the weight of the j -th variable in the t-th iteration, is the weight between the check node and the variable node in the t-th iteration, is the weight between the check node and the variable node in the t-th iteration, is a positive integer, is the value in the check set participated by the variable node .

[0009] Furthermore, in Step 2, calculate the output after the end of the t-th iteration of the linear block code decoding based on the neural network learning algorithm, including:

[0010] In the formula, is an operation, and there is ; is the logarithmic likelihood function, is the t -th iteration of the variable calculation value between the check node and the variable node , is the check set participated by the variable node .

[0011] Furthermore, in Step 3, calculate the loss function of cross-entropy , including:

[0012] In the formula, is a binary block code, with a code length of N and an information bit length of N - M, where N and M are positive integers; is the th codeword.

[0013] Furthermore, in step 4, based on the loss function of cross entropy, calculate the cumulative value at the t-th iteration, including:

[0014] In the formula, is the cumulative influence factor at the t-th iteration, is the th cumulative influence factor at the t-th iteration, .

[0015] Furthermore, in step 5, perform step size adjustment based on the cumulative value to obtain the updated step size , including: Adopt the first terms of the cumulative value at the t-th iteration as the current step size:

[0016] In the formula, is the step size for different iteration times, with the subscript being the iteration times; Set a positive deviation value and a minimum value ; If , then:

[0017] If , then:

[0018] If , then: .

[0019] Furthermore, adjust the weights of the neural network learning algorithm based on the updated step size , including: .

[0020] A computer device includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the above multi-step decoding decision accelerated decoding method is implemented.

[0021] A computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the above multi-step decoding decision accelerated decoding method is implemented.

[0022] The beneficial effects of the present invention are as follows: The present invention uses the component addition of the iterative decoding results multiple times as the basis for learning rate selection, and realizes the optimal adjustment of the neural network weights under different code lengths, code rates, and different channel conditions by setting the number of component times of the iterative decoding results and the selection of addition coefficients, thereby improving the training learning rate. The present invention is applicable to a wide range of linear block codes, including low-density parity-check codes (LDPC). BRIEF DESCRIPTION OF THE DRAWINGS

[0023] Figure 1 is a bipartite graph of a linear block code according to Embodiment 1 of the present invention.

[0024] Figure 2 is a flowchart of a multi-step decoding decision accelerated decoding method according to Embodiment 1 of the present invention. DETAILED DESCRIPTION OF THE INVENTION

[0025] In order to have a clearer understanding of the technical features, objectives, and effects of the present invention, the specific embodiments of the present invention are now described. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts belong to the scope of protection of the present invention.

[0026] Embodiment 1 Since the decoding of linear block codes using neural network learning algorithms becomes very slow as the length of the linear block code increases, the initial model training and subsequent application enhancement training time costs are very large.

[0027] The decoding method of linear block codes based on neural network learning algorithms is as follows: Taking binary block code as an example, where N is the code length, N - M is the information bit length, and N and M are positive integers. Any codeword , is transmitted in the form of where , , is a binary field, . Assume that the codeword sent by the transmitter is . After transmission mapping and binary phase shift keying (BPSK) modulation, it passes through a noisy channel and finally reaches the receiver. The receiver demodulates it and outputs a decision signal , and sends it to the channel decoder. Therefore, is the received signal (or input signal) of the channel decoder. The main task of the decoder is to perform error correction decoding on the received sequence , filter out channel errors, and recover the transmitted codeword from the received signal with as small a decoding error probability as possible.

[0028] Let represent the conditional probability distribution function of the channel output, and the log-likelihood ratio (LLR) function of the symbol can be calculated as:

[0029] where .

[0030] Let a binary block code have a parity-check matrix , where h ij is the element in the H th row and i th column of the parity-check matrix j . Figure 1 is its bipartite graph representation, and the codeword is represented as a set of information nodes , is represented as a set of check nodes . There is a directed edge connecting node to to only when . Let the set represent the check set that the variable participates in, represents is a subset that does not contain , represents the local symbol information set constrained by the check node represents is a subset that does not contain . The bipartite graph was used to describe a special class of linear block codes - LDPC codes in the past, but it is now also used to describe general linear block codes. In particular, good results have been obtained through neural network decoding of BCH codes on the bipartite graph, making the bipartite graph more general.

[0031] The iteration of the linear block code decoding method based on the neural network learning algorithm is as follows: There is a positive integer \(t\), and the \(t\)-th iteration is as follows:

[0032]

[0033] where is the coefficient that needs to train the neural network decoder, is 's inverse function. Let:

[0034]

[0035] Then: .

[0036] Based on this, in order to solve the problem that the coefficient training convergence rate of the linear block code decoding method based on the neural network learning algorithm decreases as the code length increases, this embodiment provides a multi-step decoding decision accelerated decoding method. By adding the components of the decoding results obtained through multiple iterations as the basis for selecting the learning rate, and based on different code lengths, code rates, and channel conditions, by setting the number of component times of the iterative decoding results and the addition coefficient, the optimal adjustment of the weights of the neural network learning algorithm is realized, thereby improving the learning rate.

[0037] Specifically, as Figure 2 shown, the method of this embodiment can be implemented by the following steps: Step 1: Set the initial step size of the weights of the neural network learning algorithm ; Step 2: Calculate the output after the \(t\)-th iteration of the linear block code decoding based on the neural network learning algorithm, and determine whether the decoding is successful. If it is successful, end; otherwise, execute the next step; Step 3: Calculate the loss function of the cross entropy in the \(t\)-th iteration , where is the binary block code; Step 4: Based on the loss function of the cross entropy , calculate the accumulated value in the \(t\)-th iteration; Step 5: Adjust the step size based on the accumulated value to obtain the updated step size ; Step 6: Adjust the weights of the neural network learning algorithm based on the updated step size , execute the next iteration and go back to Step 2.

[0038] In this embodiment, the weights of the neural network learning algorithm include:

[0039] Among them, is the weight of the j -th variable in the t-th iteration, is the weight between the check node and the variable node in the t-th iteration, is the weight between the check node and the variable node in the t-th iteration, is a positive integer, is the value in the parity-check set participated by the variable node .

[0040] It should be noted that the initial step size of the weights of the neural network learning algorithm can be the same or different. In this embodiment, the case where the initial step sizes are the same is taken as an example for illustration.

[0041] Preferably, in step 2, calculating the output after the end of the t-th iteration of the linear block code decoding based on the neural network learning algorithm, includes:

[0042] In the formula, is an operation, and there is ; is the log-likelihood function, is the t -th iteration, the variable calculation value between the check node and the variable node , is the variable node participated in the parity-check set.

[0043] Preferably, in step 3, calculating the loss function of the cross entropy in the t-th iteration, includes:

[0044] In the formula, is the binary block code, the code length is N, the information bit length is N - M, and N and M are positive integers; is the -th codeword.

[0045] Preferably, in step 4, based on the loss function of the cross entropy, calculating the accumulated value in the t-th iteration, includes:

[0046] In the formula, is the cumulative influence factor of the t-th iteration, is the -th cumulative influence factor of the t-th iteration, . They can be the same or different.

[0047] Preferably, in step 5, based on the cumulative value perform step size adjustment to obtain the updated step size , including: Use the first terms of the cumulative value of the t-th iteration as the current step size:

[0048] In the formula, is the step size of different iteration times, and the subscript is the iteration time; Set a positive deviation value and a minimum value ; preferably, .

[0049] If , then:

[0050] If , then:

[0051] If , then: .

[0052] Preferably, based on the updated step size adjust the weights of the neural network learning algorithm, including: .

[0053] In summary, the method of this embodiment uses the component addition of the iterative decoding results multiple times as the basis for learning rate selection, and realizes the optimal adjustment of the neural network weights under different code lengths, code rates, and different channel conditions by setting the number of component times and the selection of addition coefficients of the iterative decoding results, thereby improving the training learning rate.

[0054] Embodiment 2 This embodiment is based on Embodiment 1: This embodiment provides a computer device, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, it implements the multi-step decoding decision acceleration decoding method of Embodiment 1. Among them, the computer program can be in the form of source code, object code, executable file, or some intermediate form, etc.

[0055] Embodiment 3 Based on Embodiment 1, this embodiment: This embodiment provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, it implements the multi-step decoding decision acceleration decoding method of Embodiment 1. Among them, the computer program can be in the form of source code, object code, executable file, or some intermediate form, etc. The storage medium includes: any entity or device capable of carrying computer program code, recording medium, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal, and software distribution medium, etc. It should be noted that the content included in the storage medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, the storage medium does not include electrical carrier signals and telecommunication signals.

[0056] It should be noted that for the foregoing method embodiments, for the sake of simple description, they are expressed as a series of action combinations. However, those skilled in the art should know that this application is not limited by the described action sequence, because according to this application, some steps can be performed in other sequences or simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

Claims

1. A multi-step decoding decision accelerated decoding method, characterized in that: include: For linear block code decoding based on neural network learning algorithm, the component addition of multiple iterative decoding results is calculated as the basis for selecting the learning rate; Based on different code lengths, code rates and channel conditions, the optimal adjustment of the weights of the neural network learning algorithm is achieved by setting the number of iterative decoding result components and the addition coefficient, thereby improving the learning rate.

2. The multi-step decoding decision accelerated decoding method according to claim 1, characterized in that: The following steps are involved: Step 1: Set the initial step size of the neural network learning algorithm weights ; Step 2: Calculate the output of the linear block code decoding based on the neural network learning algorithm after the tth iteration , and determine whether the decoding is successful, if successful, end, otherwise proceed to the next step; Step 3: Calculate the cross entropy loss function in the tth iteration ,in is a binary block code; Step 4: Loss function based on cross entropy , calculate the cumulative value of the tth iteration ; Step 5: Based on the accumulated value Adjust the step size to get the updated step size ; Step 6: Based on the updated step size Adjust the neural network learning algorithm weights, perform the next iteration and go to step 2.

3. The multi-step decoding decision accelerated decoding method according to claim 2, characterized in that: The neural network learning algorithm weights include: in, is the tth iteration j The weight of the variable, is the check node for the tth iteration With variable nodes The weight between is the check node for the tth iteration With variable nodes The weight between is a positive integer, For variable nodes The value in the validation set to participate in.

4. The multi-step decoding decision accelerated decoding method according to claim 3, characterized in that: In step 2, the output of the linear block code decoding based on the neural network learning algorithm after the tth iteration is calculated ,include: In the formula, is an operation, and there is ; is the log-likelihood function, For the t Check the node at the next iteration With variable nodes Calculate the value of the variable between For variable nodes The validation set to participate in.

5. The multi-step decoding decision accelerated decoding method according to claim 4, characterized in that: In step 3, the cross entropy loss function is calculated in the tth iteration ,include: In the formula, is a binary block code, the code length is N, the information bit length is NM, N and M are positive integers; For the A code word.

6. The multi-step decoding decision accelerated decoding method according to claim 5, characterized in that: In step 4, the loss function based on cross entropy is , calculate the cumulative value of the tth iteration ,include: In the formula, is the cumulative impact factor of the tth iteration, is the tth iteration Cumulative impact factor, .

7. The multi-step decoding decision accelerated decoding method according to claim 6, characterized in that: In step 5, based on the accumulated value Adjust the step size to get the updated step size ,include: Use the accumulated value of the tth iteration Before The term is the current step size: In the formula, is the step size of different iteration numbers, and the subscript is the number of iterations; Set a positive deviation and minimum ; like ,but: like ,but: like ,but: 。 8. The multi-step decoding decision accelerated decoding method according to claim 7, characterized in that: In step 6, based on the updated step size Adjust the weights of the neural network learning algorithm, including: 。 9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the multi-step decoding decision accelerated decoding method described in any one of claims 1-8 is implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the multi-step decoding decision accelerated decoding method described in any one of claims 1-8 is implemented.

Images