LDPC code construction method based on progressive extension and ring classification sorting evaluation criterion

Abstract

The invention discloses an LDPC code construction method and device based on progressive extension and a ring classification sorting evaluation criterion, and the method proposes the ring classification sorting evaluation criterion and an actual application-oriented performance evaluation criterion according to the high correlation between signal-to-noise ratio threshold performance indexes of a QC-SC-LDPC code set before and after progressive extension. A progressive extension technology is utilized to construct a QC-SC-LDPC code with a specified code length and a specified code rate, and construction of a multi-code-length QC-SC-LDPC code is considered at the same time. Therefore, the evaluation condition setting is more flexible, and better compromise can be obtained between the search complexity and the search effect; and compared with a traditional method, the good code probability under the same search space size is greatly improved, and efficient search of the multi-code-length compatible QC-SC-LDPC codes is achieved.

Description

technical field [0001] The invention relates to the technical field of digital information transmission, in particular to an LDPC code construction method and device based on progressive expansion and ring classification and sorting evaluation criteria. Background technique [0002] R.G. Gallager invented the Low-Density Parity-Check (LDPC) code in the 1960s. In digital communication systems, LDPC codes are used as an error control (Error Control) because of their decoding performance close to the channel capacity and the low-complexity Belief Propagation (BP) iterative decoding algorithm that is linearly related to the code length. ) codes are widely used to realize efficient transmission of digital information. [0003] Spatially-Coupled (Spatially-Coupled, SC-) LDPC codes are a special kind of LDPC codes with spatial coupling characteristics. Thanks to the threshold saturation effect, SC-LDPC codes are proved to have universally optimal properties in the asymptotic sens...

AI Technical Summary

Problems solved by technology

However, traditional evaluation criteria or EMD / ACE evaluation criteria cannot distinguish rings with the same ring length and the same EMD / ACE value. Therefore, compared with traditional LDPC codes, the above evaluation criteria deteriorate the search effect of QC-SC-LDPC codes

[0011] 2) The evaluation conditions set by the traditional evaluation criteria or EMD / ACE evaluation criteria are not flexible enough, if the evaluation conditions are set too strict, the set of candidate codes after the progressive search of the PEG / cPEG algorithm is too small, and the codes with better performance evaluation may be eliminated in advance ; If the evaluation conditions are set too loosely, the set of candidate codes after the progressive search of the PEG / cPEG algorithm will be too large, and the calculation amount of performance evaluation will increase sharply, and at the same time, codes with better performance evaluation may still be missed

Therefore, according to the above evaluation conditions, the compromise effect between the performance evaluation complexity and the good code probability of the candidate code set after progressive search is poor;

[0012] 3) The above-mentioned PEG / cPEG algorithm using traditional evaluation criteria or EMD / ACE evaluation criteria is suitable for the construction of LDPC medium and long codes, because the code space of medium and long codes is relatively large, and the threshold performance of typical construction results is better. The PEG / cPEG algorithm of the traditional evaluation criteria or the EMD / ACE evaluation criteria mainly optimizes the bit error platform performance of long codes in LDPC, and the bit error platform performance is directly related to the distribution of short rings; LDPC short codes for practical application requirements, especially QC -SC-LDPC short code, because the code space is small, the above PEG / cPEG algorithm cannot even guarantee the threshold performance of the construction result

[0013] 4) The above PEG / cPEG algorithm does not consider the compatibility of LDPC codes to multi-code lengths, but practical applications require LDPC codes with scalable code lengths, so it is necessary to design a construction method for multi-code length LDPC codes

Furthermore, for QC-SC-LDPC codes, if the rings with smaller subcode intervals have been completely eliminated, the rings with larger subcode intervals will rise to be the main cause of performance impact

Therefore, the assumption that the extended parity check matrix is ​​the same for different subcodes will severely limit the elimination of rings with large subcode intervals, thereby reducing the probability of good codes in the code space

Images