Research On The Hight Efficient Encoding/Decoding Technique For LDPC Codes And Serially Concatenated LDPC/Convolutional Codes

The low-density parity-check code was originally proposed by Dr. Gallager in 1962, which cannot attract a large amount of interest at that time. However, after decades of years, it is now rediscovered by Mackay and Neal, along with the enhancement of computer ability and the development of some related theories, such as graph theory, BP algorithm, turbo codes principle, etc. It is also shown that the performance of LDPC codes is very close to the Shannon limits with sufficient long block length when decode using the BP decoding algorithm.Ecoding method which we use to detect and correct errors is always a hot topic in comumication theory. Turbo codes have achieved a great success in the third generation wireless communication system. But many experts dout turbo codes can be efficiently uses in next generation wireless communication system in view of their high complicated decoding algorithms.The LDPC codes and turbo codes are similar in many aspects, both can extremely approach to the Shannon limits by their unique ways. The LDPC codes are relatively easy defined and can outperform the turbo codes with sufficiently long block lengths. Its decoding complexity is also lower than turbo code. Recently, LDPC code has drawn the worldwide attentions in channel coding community due to its high performance and strong potential applications in practice.This thesis offers a comprehensive study on the performance of LDPC codes. The author first presents a new method for construction of irregular LDPC codes based on generator matrix in order to significantly reduce encoding complexity. The simulation results are obtained over an AWGN channel, and the obtained results are also compared with those acquired by the traditional approaches. In order to further reduce the decoding complexity, the author researches the decoding methord of LDPC codes in serial way. Finally, in order to further improve the performance, we propose a new kind of code, which is LDPC/convolutional code, simply denoted by"LDPC/Conv."code. The factors that can impact the performance of the coded systems are investigated and concluded through the simulations.