Analysis And Design Of Puncturing Scheme For QC-LDPC Codes

Low-density parity-check（LDPC）codes have been widely used in the past decade due to their error-correcting capability,error-detecting capability and low decoding complexity,which are close to the Shanoon limit.Thanks to the continuous improvement of LDPC codes'encoding and decoding schemes,the applicability and operability of LDPC codes are greatly improved.In order to get higher capacity and frequency utilization in digital communication systems,a method of puncturing is usually adopted to achieve the rate adaptive,that is,to change the encoding method and modulation scheme adaptively according to the real-time performance of the channel.However,puncturing adjusts the code-rate by deleting part of the code element information,and destroys the confidence information's integrity of check matrix at the same time,thereby affecting the error-correcting capability of LDPC codes.The disadvantage limits the application of LDPC codes in mobile communication system^[1].The dissertation is designed to solve the problems and promote the wider application of LDPC codes.Quasi-cyclic low-density parity-check（QC-LDPC）code is a kind of highly structured LDPC code.Its quasi-cyclic characteristic can simplify the encoding process and improve the decoding efficiency.To reduce the BER and FER of QC-LDPC codes in the decoding process,and also improve the performance of the communication system in noisy channels,starting from the structure of QC-LDPC codes and combining the cycle characteristics of tanner graphs,a set of improved puncturing schemes is proposed.Firstly,an appropriate QC-LDPC code based on cyclic permutation matrix is designed,and the encoding procedure is simplified according to the construction of check matrix.Based on the encoding implementation,the corresponding bipartite graph representation is given,and an algorithm for detecting the minimum cycle in QC-LDPC codes is proposed based on the theoretical relationship between the structure of check matrix and the decoding performance.The detailed description and flow chart of the algorithm are given.Based on the algorithm,a set of QC-LDPC codes check matrices with increasing minimum cycle length is constructed.The appropriate maximum iteration time is confirmed,and the BER and FER simulations are performed under the AWGN channel to detect the effect of the minimum cycle length on the decoding performance,so as to verify the correctness and applicability of the algorithm.On the basis of the algorithm,the puncturing scheme of QC-LDPC codes with limited code length is solved.It is known that the selection of puncturing schemes will affect the decoding performance.The dissertation proposed a series of selection schemes for puncturing positions and numbers based on the minimum cycle length.The variables are set to simulate and contrast with the existing puncturing scheme,so as to compare the advantages and disadvantages of each puncturing scheme.Finally,the influence of minimum cycle length on puncturing positions and numbers is determined,and the optimal scheme is retained.