Method and device for constructing check matrix of low-density parity-check code

Abstract

The invention discloses a constructing method and a device of a low density parity check (LDPC) code check matrix. The method includes steps of dividing bits to groups; searching an initial check address of each group; judging whether the length of a ring which exists in a check matrix formed by the searched initial check address and the initial check address of the group is no greater than that of a ring of a first ruled threshold value, and determining the searched initial check address to be the initial check address of the group if the judging result is negative; judging whether the quantity of the initial check address of the group reaches a preset second ruled threshold value, and continuing searching the initial check address if the judging result is negative; and determining the initial check address of each group, and constructing the corresponding LDPC code check matrix according to the determined initial check address, wherein the quantity of the initial check address of the group is the second ruled threshold value. The invention further provides a specific implementation scheme for constructing the LDPC code check matrix without short rings.

Description

technical field [0001] The invention relates to the technical field of low-density parity-check codes, in particular to a method and device for constructing a check matrix of a low-density parity-check code. Background technique [0002] Low Density Parity Check (LDPC, Low Density Parity Check) codes are a class of linear block codes defined by sparse matrices. Its iterative decoding algorithm is relatively simple and has performance close to the Shannon limit. At present, many communication system standards LDPC codes are used. [0003] There are many encoding methods of LDPC codes, which can be roughly divided into random construction method and algebraic construction method. Although the random construction method conforms to the random coding idea in Shannon's information theory, it has the performance close to the Shannon limit when the code length is long, but because the randomly constructed LDPC code does not have a certain code structure, the complexity of encoding...

AI Technical Summary

Problems solved by technology

However, the check matrix of the LDPC code generally has a ring, and the length of the ring is an even number not less than 4. If there is a ring in the check matrix of the LDPC code, the probability calculated by the sum-product algorithm is not a true posterior probability. The optimal symbol-by-symbol maximum a posteriori probability cannot be achieved, so the existence of the loop makes the optimal performance of decoding unsatisfactory, especially the loop with a short loop length (also called a short loop, for example, the loop length is 4) It will seriously affect the decoding performance of LDPC codes. Therefore, when constructing the parity check matrix of LDPC codes, short loops should be avoided as much as possible to reduce the impact of short loops on the decoding performance of LDPC codes

[0005] In the prior art, no specific implementation scheme for constructing a check matrix of an LDPC code without short loops has been proposed

Images