On The Informed Dynamic Scheduling Based Improved Decoding Algorithm For LDPC Codes And Digital Fountain Codes

With the rapid development of wireless communication, high-speed data transmission and multimedia sevices increase rapidly. Therefore, how to ensure the reliability of communication process is particularly important on the basis of information transmitted, and need higher demand on the error-correcting codes. Low-density parity-check (LDPC)codes can approach the Shannon limit error correction performance, which is described conveniently and easy theoretical analysis as well as decoding. Meanwhile, it is quite Suitable for hardware implementation. On the other hand, digital fountain code as an efficient rateless codes, it can adaptive matching encoding scheme, according to channel status from time to time. Belief Propagation(BP) algorithm is decoding basis of these two codes. However, conventional BP can’t pass the latest updated message to adjacent edges. Then based on the informed dynamic scheduling(IDS) for BP decoding, there is more confidence message for every message passing.In this paper, first briefly introduce the basic concepts of LDPC codes and LT codes, including representations, feature and encoding processing. Then detailed iterative decoding algorithm is described, that is principle and process of BP algorithm. On this basis, analyze two key factors affecting the BP decoding algorithm, as well as informed dynamic schedules (IDS). Next, discuss and analyze the specific process of BP decoding algorithm based on residual value of IDS. On this basis, combined to Tanner gragh, mainly introduce three algorithm, including Residual Belief Propagation(RBP) and Node-wise Residual Belief Propagation (NW-RBP), which based on the differences between the value of the check nodes to variable nodes message before and after an update, variable nodes to check nodes Residual Belief Propagation(VCRBP), which based on the he differences between the value of the variable nodes to check nodes message before and after an update. Moreover, discuss the characteristics of each scheduling algorithm, and expound process that the relevant IDS strategies overcome "trapping sets". By simulation study, analyze and compare the decoding performance on different IDS strategies affecting on LDPC codes and LT codes.Simulation study analysis, according to LDPC and LT codes, show that the performance for the BP decoding algorithm on IDS can be improved. However, it depend on coding constructs. Generally speaking, if a codeword is medium and short, and degree distribution relatively uniform in the Tanner graph, and there is no large number of variable nodes with one degree as well as girth, then the BP decoding on IDS can significantly improve decoding performance. Otherwise, if a codeword is medium and short, the performance improvement is not significant. So the papers of analysis results indicate that, in practice, combined with the coding structure, be reasonable to choose improved BP decoding algorithm on IDS strategies, in order to give full play to the role of IDS that can bring potential performance improvements.