The Simulation Research On Encoding And Decoding Of Ldpc Code

With the rapid development of today's communication system, the demand of the quality of data transmission has been increasing continuously. As an important mean to improve the security of information transmission, the error-correcting encoding technology has attracted more and more attention. Since it was proposed again in 1996, LDPC code has received great attention and become a research hot spot due to its good coding performance which is approximate Shannon limit.This article introduced the linear grouping code and introduced its encoding rule and analysed the Gauss channel model and the Rayleigh channel model. Analysed two sorts of decoding algorithms of LDPC code and completed the simulations between them. Summarized the pros and cons of the two sorts of algorithms. Analysed the BP (belief propagation) algorithm and proposed a new iteration stopping rule in order to reduce the times of iterations. The proposed algorithm effectively improved the performence of BP algorithm.The new iteration stopping rule decides the convergence of the result sequences via comparing the results of the interval iterations. It compares the result of the i th and (i+j) th iterations in every j iterations. If the results of comarison are very close, it means that the result sequences are already converged and the decoding ends. Otherwise, the result sequences are not converged and the docoding needs more iterations.At the condition of low SNR, the new iteration stopping rule effectively reduced the times of iteration and lowerd the computing complexity and decoding delay without the change of decoding performence. Because this new iteration stopping rule is based on the BP algorithm, it also suits other improved algorithms.