NB-LDPC code decoding method with high performance and low complexity

Abstract

The invention discloses a non-binary low density parity check (NB-LDPC) code decoding method with high performance and low complexity, which belongs to the technical field of communication. The methodachieves the target of high performance and low complexity from two aspects: 1) with the combination of advantages of simplifying minimum sum algorithm (SMSA) having high performance and grid-based expanded mini-sum algorithm (T-EMSA), an improved grid-based mini-sum algorithm (IT-MSA) is brought forward, wherein under the same condition, IT-MSA obtains better performance and lower complexity compared with the conventional method; 2) a double threshold-based shrinking method (DTBS) is applied in the IT-MSA, and the calculating complexity is further and greatly reduced, while the decoding performance basically maintains the same. The invention is advantageous in that through the processing of the two steps, the performance of the decoding method is equal to the optimized expanding minimumsum algorithm (EMSA), and the complexity is greatly reduced compared with the EMSA, SMSA and T-EMSA.

Description

technical field [0001] The invention relates to a decoding method in the technical field of communication; in particular, it is aimed at a decoding method based on NB-LDPC codes in a communication system. Background technique [0002] In a digital communication system, when a signal is transmitted in a channel, it will inevitably be interfered by various noises. If the signal is sent directly, the wrong information will generally be received. Therefore, in the process of signal transmission, a certain To correct the error information in the transmission process, forward error correction (FEC) is one of the commonly used error correction methods at present. After error correction coding, it can be decoded after passing through the noise channel, and a certain degree of error correction can be performed. Function. Common error correction codes include: Low Density Parity Check (LDPC) codes, Turbo codes, BCH codes, RS codes, Polar codes, etc. [0003] The LDPC code was first ...

AI Technical Summary

Problems solved by technology

The biggest problem of the former is that there are too many serial calculations, which seriously affect the decoding throughput. The latter mainly include the simplified minimum sum method (SMSA), the grid-based extended minimum sum method (T-EMSA), and the grid-based The minimum maximum value method (T-MMA), when the maximum deviation number is selected properly, the decoding performance of SMSA and T-EMSA is equivalent, slightly inferior to EMSA, and significantly better than T-MMA

[0005] In addition, note that in the existing decoding methods, the decoding process is performed on the complete Galois field, that is, in order to find a correct codeword, it is necessary to convert all received symbols The consideration of information also means that when q is larger, more redundant information will be included. Previous scholars have also noticed the existence of this problem, but everyone has focused on CNP. The information input by the secondary verification node is compressed to reduce the complexity, such as: improved EMSA, X-EMSA, μ-EMSA, etc. These methods have two problems: 1) they are only simplified for the CNP of EMSA; 2) In each iterative input, additional input information processing operations (such as truncation, compression, etc.) are required, and the subsequent output information also needs to be inversely processed, which increases the amount of calculation

Images