CN-122026924-A - Greedy confidence method and greedy confidence device applied to low-density parity check code

CN122026924ACN 122026924 ACN122026924 ACN 122026924ACN-122026924-A

Abstract

The invention provides a greedy confidence method and a greedy confidence device applied to a low-density parity check code, wherein the greedy confidence method comprises the steps of initializing a decoder, and obtaining variable nodes and check nodes after initialization processing; the method comprises the steps of obtaining a message based on a received signal through a variable node, transmitting the message to a check node, carrying out partial node updating and iteration in the message transmission process of the variable node and the check node by adopting a greedy optimization strategy, obtaining an optimization analysis result, obtaining decoding information according to the optimization analysis result, and outputting the decoding result aiming at the decoding information. The invention applies the greedy algorithm to the BP algorithm, optimizes the BP algorithm, updates the messages of partial variable nodes and check nodes according to the greedy optimization strategy, reduces the complexity of the BP algorithm, and simultaneously effectively reduces the iteration time, thereby reducing the decoding delay and improving the LDPC decoding efficiency.

Inventors

CHEN JINHUI
Zhen Yimo
XU ZHAN
ZHI RUXIN

Assignees

北京信息科技大学

Dates

Publication Date: 20260512
Application Date: 20241112

Claims (10)

1. A greedy confidence method applied to low density parity check codes, comprising: Initializing a decoder, and acquiring variable nodes and check nodes after initialization processing; Transmitting the message obtained based on the received signal to a check node through a variable node; adopting a greedy optimization strategy to update and iterate partial nodes in the process of transmitting messages between variable nodes and check nodes, and obtaining an optimization analysis result; and obtaining decoding information according to the optimized analysis result, and outputting the decoding result aiming at the decoding information.
2. The greedy confidence method applied to low-density parity check codes according to claim 1, wherein the variable nodes and the check nodes have a communication relationship, a parity check matrix is determined according to the communication relationship between the variable nodes and the check nodes, rows of the parity check matrix correspond to the check nodes, the number of rows of the parity check matrix is the same as the number of check nodes, columns of the parity check matrix correspond to the variable nodes, the number of columns of the parity check matrix corresponds to the number of variable nodes, when the communication relationship between the check nodes and the variable nodes exists, elements at corresponding positions in the parity check matrix are 1, and when the communication relationship between the check nodes and the variable nodes does not exist, elements at corresponding positions in the parity check matrix are 0.
3. The greedy confidence method applied to low-density parity-check codes according to claim 1, wherein when initializing the decoder, initializing the decoder for all variable nodes and check nodes, analyzing a form of a received signal when initializing the variable nodes, when the form of the received signal is represented by log likelihood ratios, distributing an initial message of the received signal as the variable nodes to the variable nodes according to an initialization rule, when the form of the received signal is represented by hard decision forms, performing log likelihood ratio calculation for the received signal to obtain the LLR values of the received signal, and distributing the LLR values of the received signal as the initial message of the variable nodes to the variable nodes according to the initialization rule for transmission, and when initializing the check nodes, setting the initial message of the check nodes to 0 and waiting for a message transmitted by the variable nodes to be received.
4. The greedy confidence method applied to low-density parity-check codes according to claim 1, wherein the variable nodes and the check nodes have a one-to-one or one-to-many communication relationship, when a message obtained based on a received signal is transmitted to the check nodes through the variable nodes, association relationship analysis is performed on the variable nodes, check nodes having a communication relationship with the variable nodes are determined, adjacent check nodes are obtained, and then the message of the received signal is transmitted to the adjacent check nodes.
5. The greedy confidence method applied to the low-density parity check code according to claim 1, wherein when a greedy optimization strategy is adopted to update and iterate partial nodes in the process of message transmission between a variable node and a check node, greedy algorithm parameter setting is performed in advance for the greedy optimization strategy, greedy selection threshold and greedy attenuation coefficient are determined, and dynamic threshold calculation is performed by combining the greedy selection threshold and greedy attenuation coefficient with iteration times, so that greedy threshold is obtained.
6. The greedy confidence method applied to the low-density parity check code according to claim 5, wherein the greedy optimization strategy is adopted to update and iterate partial nodes in the process of message transmission between a variable node and a check node, the method comprises the steps that the check node receives messages transmitted by the variable node, calculates and updates according to the received messages, feeds back calculation and update results to the variable node according to a communication relation, screens feedback messages according to the greedy optimization strategy by combining greedy selection thresholds, updates the variable node according to the screening results, updates the current iteration times, updates the greedy threshold according to the current iteration times by combining greedy selection thresholds and greedy attenuation coefficients, performs optimization data analysis according to the greedy threshold, and performs loop updating of the check node and the variable node according to optimization analysis data until the optimization analysis data reaches a decision threshold to obtain the optimization analysis results.
7. The greedy confidence method applied to low-density parity-check codes according to claim 6, wherein when the data analysis is optimized according to greedy threshold values, a mathematical model is built according to decoding problems, then the solved problem is divided into a plurality of sub-problems, the greedy threshold values are determined at the same time, then the local optimal solutions of the sub-problems are found according to the greedy threshold values, and then the optimal solutions of the sub-problems are synthesized to optimize the problems, so that an optimal analysis result is obtained.
8. The greedy confidence method applied to low-density parity-check codes according to claim 6, wherein the decision threshold comprises an iteration threshold and data convergence, when an optimal analysis result is determined according to the decision threshold, optimization screening is performed on current optimal analysis data to obtain the optimal analysis result when the current iteration number reaches a preset iteration threshold, and when the current iteration number does not reach the preset iteration threshold but the optimal analysis data convergence occurs, the optimal analysis result is obtained according to the convergence data.
9. The greedy confidence device applied to the low-density parity check code is characterized by comprising a signal acquisition monitoring module, a node message transmission module, a greedy judgment module and an iteration output module; The signal acquisition monitoring module is used for receiving signals to be processed; the node message transfer module is used for initializing the decoder and transferring messages between the check node and the variable node; The greedy decision module is used for carrying out partial node updating and iteration in the process of carrying out message transmission between the variable nodes and the check nodes by adopting the greedy optimization strategy; and the iteration output module is used for obtaining the decoding information according to the optimized analysis result and outputting the decoding result aiming at the decoding information.
10. The greedy confidence device for low-density parity-check codes according to claim 9, wherein the signal acquisition and monitoring module comprises a signal acquisition unit, an analysis unit and a quality determination unit; The signal acquisition unit is used for acquiring LDPC coded real-time operation data in real time, and performing time sequencing on the operation real-time tracking data to obtain a signal to be processed; the analysis unit is used for determining a parity check matrix according to the signal to be processed, and determining greedy algorithm parameters and decision thresholds of the greedy optimization strategy; the quality determining unit is used for judging the quality of the signal to be processed, and initializing the variable node and the check node after the complete signal to be processed is obtained.

Description

Greedy confidence method and greedy confidence device applied to low-density parity check code Technical Field The invention relates to the technical field of channel decoding, in particular to a greedy confidence method and a greedy confidence device applied to a low-density parity check code. Background With the rapid development of information technology, the 5G and 6G networks are very convenient for communication. In 5G and 6G networks, the User Equipment (UE) is receiving a data stream from the base station (gNB), but due to the complexity of the radio channel, the data is affected by noise and interference during transmission, and the task of the receiving end (UE) is to recover the originally transmitted data by decoding the received noisy data, as shown in fig. 1, standard information transfer usually involves LDPC (low density parity check) coding and LDPC decoding, using Low Density Parity Check (LDPC) codes as a Forward Error Correction (FEC) coding scheme, and using the Belief Propagation (BP) algorithm as a soft decision decoding algorithm is widely used for LDPC decoding. The conventional belief propagation decoding algorithm generally updates the messages of all variable nodes and check nodes in each iteration, and has large calculation amount and high complexity, so that the decoding delay is large, therefore, the invention provides a greedy belief method and device applied to low-density parity check codes, and applying a greedy algorithm to the BP algorithm, optimizing the BP algorithm, updating the messages of partial variable nodes and check nodes according to the greedy optimization strategy, reducing the complexity of the BP algorithm, and simultaneously effectively reducing the iteration time, thereby reducing the decoding delay. Disclosure of Invention The present invention is directed to a greedy confidence method and apparatus for low density parity check codes, which solves the above-mentioned problems in the prior art. In order to achieve the above purpose, the invention provides a greedy confidence method applied to a low density parity check code, comprising the following steps: Initializing a decoder, and acquiring variable nodes and check nodes after initialization processing; Transmitting the message obtained based on the received signal to a check node through a variable node; adopting a greedy optimization strategy to update and iterate partial nodes in the process of transmitting messages between variable nodes and check nodes, and obtaining an optimization analysis result; and obtaining decoding information according to the optimized analysis result, and outputting the decoding result aiming at the decoding information. Further, the variable nodes and the check nodes have a communication relation, a parity check matrix is determined according to the communication relation between the variable nodes and the check nodes, the rows of the parity check matrix correspond to the check nodes, the numbers of the rows of the parity check matrix are the same as the number of the check nodes, the columns of the parity check matrix correspond to the variable nodes, the columns of the parity check matrix correspond to the number of the variable nodes, when the communication relation between the check nodes and the variable nodes exists, the elements at the corresponding positions in the parity check matrix are 1, and when the communication relation between the check nodes and the variable nodes does not exist, the elements at the corresponding positions in the parity check matrix are 0. Further, when initializing the decoder, initializing all variable nodes and check nodes, analyzing the form of a received signal when initializing the variable nodes, when the form of the received signal is represented by a log likelihood ratio, distributing LLR values of the received signal as initial messages of the variable nodes to the variable nodes according to an initialization rule for transmission, when the form of the received signal is represented by a hard decision form, carrying out log likelihood ratio calculation on the received signal to obtain LLR values of the received signal, distributing the LLR values of the received signal as initial messages of the variable nodes to the variable nodes according to the initialization rule for transmission, and setting the initial messages of the check nodes to 0 and waiting for messages transmitted by the received variable nodes when initializing the check nodes. Further, the variable nodes and the check nodes have one-to-one or one-to-many communication relation, when the message obtained based on the received signal is transmitted to the check nodes through the variable nodes, association relation analysis is carried out on the variable nodes, the check nodes with the communication relation with the variable nodes are determined, adjacent check nodes are obtained, and then the message of the received signal is transmitted to the adjacent check