Residual Belief-propagation For LDPC Codes With Dynamic Correction Factors

The rapid development of modern communication technology is imperceptibly changing all aspects of people's life and work,bringing us great convenience.Modern communication has become an important part of modern life style.The purpose of communication is to transmit information effectively and reliably.Wireless channel exists various kinds of interference,which brought a lot of challenges to the effective transmission of digital signals.The LDPC(Low Density Parity Check)codes is a linear block code that can approximate the Shannon limit.LDPC codes have strong error-correcting ability and low latency,and have become a research hotspot in the field of communication coding in recent years.The decoding methods of LDPC codes include: BP decoding algorithm(Belief-Propagation),MS decoding algorithm(Min-Sum),RBP decoding algorithm(Residual Belief-Propagation)and NWRBP decoding algorithm(Node-Wise RBP).BP algorithm has better decoding performance,but the complexity of the algorithm is relatively high.MS algorithm is an approximate algorithm of BP decoding algorithm,which complexity is the lowest,but the decoding performance of MS algorithm is also the worst.RBP and NWRBP are LDPC decoding algorithms with dynamic scheduling mechanism.Compared with BP and MS algorithms,RBP and NWRBP have lower decoding error rate and faster convergence rate.RBP and NWRBP decoding algorithms determine the order of updating edges by calculating the residual value,which includes a large number of exponential and logarithmic multiplication operations.So,RBP and NWRBP decoding algorithms have very high decoding complexity and are difficult to implement in hardware.This paper mainly studies the reduction of complexity and improvement of decoding performance of RBP and NWRBP decoding algorithms for LDPC codes.The main work is as follows:(1)Firstly,the background of LDPC codes,the knowledge of linear block codes and the representation and construction methods of LDPC codes are introduced.The ideas and decoding processes of BP,MS,RBP and NWRBP decoding algorithms for LDPC codes aresystematically introduced.The performance differences of different decoding algorithms are compared through simulation experiments,and the differences in decoding complexity of different algorithms are analyzed in depth.(2)In order to reduce the complexity of RBP and NWRBP decoding algorithms,this paper proposes improved ERBP and ENWRBP decoding algorithms.In the iteration process of RBP and NWRBP decoding algorithms,the idea of approximate calculation of MS decoding algorithm is used to simplify the calculation steps of residual degree,thus greatly reducing the computational complexity of the decoding algorithm and achieving the purpose of reducing the decoding complexity.In order to compensate for the loss of decoding performance caused by the approximation algorithm,double correction factors are introduced to modify the confidence message value of the checking node,so as to improve the accuracy and decoding performance of the approximation algorithm.In order to further improve the decoding performance of the algorithm,this paper proposes a residual confidence propagation decoding algorithm based on dynamic correction factor,namely ENRBP and ENNWRBP.By using dynamic adjustment factor,the dynamic correction factor adjusts the value slowly after each iteration to ensure that the value of dynamic correction factor is small in the whole iteration process.In order to achieve better correction accuracy and improve the performance of the decoding algorithm,the amplitude adjustment is carried out.Finally,the simulation results show that the improved method has excellent decoding performance.