Research Of The Encoding And Decoding Method In LDPC

Low Density Parity Check(LDPC)code,a kind of channel code closing to the Shannon Limit,because of its outstanding merits,such as simple description,low decoding complexity,possible parallel implementation,low error floor,has became one of the heated research topics in the field of channel coding in recent years.This paper mainly focuses on the construction method of LDPC code check matrix,the loop number detection algorithm method,and the encoding and decoding algorithm method.The main work is as follows:Firstly,this paper proposes a new LDPC code parity check matrix construction method.On the basis of QC-LDPC code,a method of constructing an LDPC code with quasi-cyclic result is proposed based on the AP sequence and the QC-LDPC code constructed by using this method is at least 6 in girth,which is,excluding "4-ring." Simulation results show that the coding process is simple,and the performance is better.Secondly,an improved LDPC coding algorithm is proposed.This paper studies the shapes of the short loops with different ring lengths and the traditional detection algorithm of the ring number in the LDPC code check matrix.Based on the SPFA algorithm in the graph theory,an improved low complexity ring number decision method is proposed.The method does not need to detect the ring numbers of different ring lengths like the conventional ring analysis detection method,but obtains the ring numbers of all the rings by single detection.The complexity of the improved algorithm is reduced by nearly two orders of magnitude compared with the latter.Thirdly,an improved LDPC decoding algorithm is proposed.This paper studies the process of BP,LLR-BP and Min-Sum decoding algorithm,and then compares the fundamental differences in the derivation of LLR-BP and Min-Sum.Aiming at the source of the difference,an improved algorithm is proposed by combining these two algorithms.The improved algorithm neutralizes the two through the dynamic selection in the decoding process so that the decoding performance and complexity can be compromised in both decoding algorithms.Fourthly,an improved LDPC encoding scheme are proposed.This paper focuses on the construction and coding scheme of LDPC codes used in the IEEE802.16e standard.Based on the two encoding methods of serial and parallel,a bi-directional recursive encoding method is proposed for the LDPC codes used in the standard,and the implementation scheme is given.The complexity analysis of Gaussian elimination algorithm,LU algorithm and Efficient algorithm shows that the new method has low complexity and the operation in hardware is simpler.