Research On LDPC Encoding And Decoding Algorithm

The development of communication technology is very rapid,and the channel bandwidth and capacity are also increasing.When information is transmitted,people need more reliability and effectiveness.The LDPC(Low Density Parity Check)codes proposed by Gallgaer in 1963 is a good code that can approximate the Shannon limit.It has been paid more and more attention by researchers and is applied to various communication standards,including the satellite digital TV transmission standard,CCSDS(Consultative Committee for Space Data Systems)deep space communication and the data channel under standard eMBB in 5G.This paper studies LDPC codes from encoding and decoding.Firstly,we proposed a design method of proto-graph LDPC codes.Due to the high complexity of traditional LDPC encoding,the JPL proposed proto-graph LDPC codes.Its main characteristics are good performance,simple structure and easy analysis and it is easily implemented by hardware.It is a structure type code.Usually we only need to analyze the proto-graph to get the performance of the entire LDPC codes.Such as,the decoding threshold.To construct the proto-graph LDPC codes,we can use the PEG algorithm.This algorithm guarantees the girth as large as possible without the existence of 4 rings.Based on this method,a sparse PEG algorithm for proto-graph is proposed.That is,when the PEG algorithm selects check nodes,the constraint conditions are increased.In the candidate check nodes,the average distance from the other known nonzero is calculated for each candidate node,and the maximum value of the check nodes is selected from the average distance.This method not only inherits the advantage of the proto-graph LDPC codes PEG algorithm,but also reduces the short loop number.This enables the constructed codes to have better performance in burst channels.Secondly,in terms of decoding,in order to optimize the decoding algorithm,a dynamic weighted decoding algorithm is proposed in this paper.The principle is that in the process of iterative decoding,the probability of information is different,so that the reliability of information is different.Reliable information can accelerate convergence.At the same time,it is necessary to suppress the transmission of unreliable information value in decoding,so as to avoid affecting the correct decoding.Therefore,according to this property,the information is dynamically weighted.This enables the information to be accelerated when it is transmitted,thereby reducing the number of iterations and increasing the gain.This algorithm is not only suitable for traditional sum-product algorithm,but also for minimal sum algorithm.It is a universal algorithm.Simulation results show that the method of dynamic weighting indeed improves the performance of decoding and reduces the number of decoding iterations.In addition,this algorithm is applied to LDPC codes in 802.16 e standard and simulated.Then,combined with the simulation results of 802.16 e standard LDPC codes and simulation results of the LDPC codes designed by LU encoding method,the reference suggestions for weighting coefficient selection are given.Finally,this paper further discusses how segmentation can make the decoding performance better and make the complexity relatively low.