Research And Improvement Of Low Density Check Parity Code Decoding Algorithm

Communication is important in any age.Nowadays the rapid development of science and technology,people demand from text to communication into voice and even video,step by step change request has been become more and so on the function of the communication system requirements are also getting higher and higher,the accuracy of the communication is key.Low Density Check-Parity Code(LDPC)is very important in the field of error correcting codes.LDPC code relative to other error correction code is approaching shannon limit,simple decoding and better hardware implementation and suitable for most of the advantages of the channel,attracts a lot of relevant researchers LDPC code,the decoding of LDPC code is an important direction to improve its performance.This article mainly studies the LDPC code decoding from the following points:1.Analyze the factors that affect the performance of decoding,starting with the different codes of LDPC code,the ring length and the number of iterations.LDPC in is essentially a set of linear block codes,type in the main code and the rules of code analysis rules,relative to character,the use of rules code type code decoding algorithm performance is superior to the rules,but its implementation is difficult,no rules code complexity is low.For the ring length,the annulus is better than the ring,and the short-link long ring affects the decoding performance.In the analysis of the number of iterations,the iterator is introduced first,and then the influence of different iterations on the decoding performance under the same code simulation is analyzed.Simulation data show that when the higher the number of iterations,decoding the lower bit error rate,but when the number of iterations to achieve a certain degree,to increase the number of iterations does not strengthen the decoding performance again,so choose the appropriate number of iterations can enhance the decoding performance and avoid unnecessary calculation.2.The existing decoding algorithm is briefly introduced,and three methods for improving the decoding algorithm are studied,which mainly include combination improvement,mixed improvement and reliability improvement.Combination of the improved method is mainly aimed at limited geometry of low density parity captchas decoding algorithm,based on the existing combination improve decoding algorithm,can make the two kinds of decoding in the process of decoding methods complement each other,so that can let the decoding performance is improved,but also greatly reduce the decoding time,the key is to avoid the generation of infinite loop.The mixed improvement is mainly aimed at the code decoding,then the main algorithm and the auxiliary algorithm are proposed on the basis of the combination improvement,and the performance improvement is realized through the setting of the priority rules of the two algorithms.The improvement of reliability is to improve the performance of decoding algorithm by dynamically determining the reliability of the information node,and the ability to locate and correct the error information node.3.In this paper,the improved points based on the amplitude of the Average information node weighted bits flip(AMWBF)algorithm and the Reliability ratio weighted bits flip(RRWBF)modified algorithm combination of rapid joint weighted bits flip decoding algorithm,the simulation data show that the improved decoding algorithm ber performance than the single decoding algorithm improved the 0.2 dB ~ 0.7 dB,needs less number of iterations 6 ~ 11,but also avoids the phenomenon of single decoding algorithm may produce infinite loop.