Low density parity check (LDPC) code check matrix construction method and corresponding matrix multiply operation device

Abstract

The invention relates to a low density parity check (LDPC) code check matrix construction method and a corresponding matrix multiply operation device. The method comprises the following steps that: determining the length n*L of a code word, the size L of a circulating replacement sub-matrix, a code rate L and number of rows m*L of a check matrix, wherein the size Hb of a basic matrix is m*n, the size H of the check matrix is mL*nL; determining best distribution of the check matrix H; constructing a basic matrix Hb with maximal loop length; exapdning the obtain basic matrix Hb to a quasi-cyclic-low density parity check (QC-LDCP) code with the loop length being maximized by selecting an appropriate movement factor, and determining the entire check matrix H by selecting the movement factor with element of a 1 position in the basic matrix Hb. The device comprises a check parity generating unit, an information sequence generating unit, a segmenting unit and a multiply unit. Due to the adoption of the method and the device, the coding complexity is reduced, the LDPC code with an error check performance being very approximate to a random structure is obtained, and simultaneously the realization complexity of the LDPC code is reduced.

Description

technical field [0001] The invention relates to the field of wireless information transmission, in particular to an LDPC code check matrix construction method and a corresponding matrix multiplication operation device. Background technique [0002] Low-Density Parity-Check codes (LDPC) were first proposed by Gallager in 1962. In 1981, Tanner introduced a bidirectional graph to describe LDPC codes. In the 1990s, Mackey et al. rediscovered LDPC codes and proved that LDPC codes are an approximation Good yardage for Shannon's limit. LDPC codes are widely used in digital communication and data storage. Standards such as 802.16e, 802.11n, 802.3a, DVB-S2, DMB-T, and CMMB all use LDPC as their channel coding scheme. [0003] At present, the construction methods of LDPC codes mainly include random construction and structural construction methods. Random construction methods include random search, PEG algorithm, etc. Among them, the code word constructed by PEG algorithm has good p...

AI Technical Summary

Problems solved by technology

Although this type of code has a good error correction ability in long codes, due to the long code group and the irregularity of the generation matrix and check matrix, the encoding is too complicated and difficult to implement in hardware. This type of codeword is suitable for theoretical Research and simulation comparison; structural construction methods include geometric, algebraic, and combinatorial design methods. Most LDPC structural codes have cyclic or quasi-cyclic structures, and quasi-cyclic codes (QC-LDPC) have strong error correction in short and medium codes ability, the performance is close to the optimal LDPC code with random structure, and because the hardware implementation is extremely simple, the quasi-cyclic code with a specific structure is used in the existing standards

[0004] LDPC code is a kind of sparse linear block code, but different from general linear block code, LDPC code is usually represented by check matrix H, and its corresponding generator matrix G is usually non-sparse, so the coding complexity is high

Images