Optimization Of LDPC Codes In Multiple Access Channel

The encoding, decoding, and the construction methods of generator matrix of LDPC codes with reduced complexity have matured, and the theoretical analysis, approximation methods, optimization work of LDPC codes have finished in single user AWGN Channel. The results obtained by the methods used in single user AWGN channel are not ideal in AWGN multiple access channel. The paper is focused on the optimization of LDPC codes in multiple access channel, it includes the following:1?Enhancing the performance of the system in MAC by means of designing irregular LDPC code degree profiles. Interleaver is used as the unique character to distinguish different user, and low cost iterative detection algorithm is applied. Gaussian approximation is used to approximate the PDF of message on FG, and iterative message updating algorithm is derived. Degree profile is obtained based on this algorithm. The relationship between the iteration number and the performance is studied, and to improve the performance, the guideline for setting the inner and outer iteration number is obtained.2?Curve fitting the EXIT characteristics of LDPC code is studied to optimize the LDPC code degree profile in AWGN MAC. EXIT characteristics are used to judge the convergence behavior of LDPC decoder. The output signal of the detector in AWGN MAC is regarded as the output of single user AWGN channel with SINR deteriorated. Differential evolution is utilized to search the optimized LDPC code degree profile. The results show that the performance of LDPC code profile obtained with EXIT curve fitting is comparable to that of obtained by Gaussian approximation based density evolution. And the EXIT curve fitting technique has the merit of reduced complexity. 3?A study to approach multiple access channel capacity by means of power allocation and LDPC code degree profile optimization iss carried out. The power allocation scheme is designed based on the maximization of mutual information and power profile is obtained by using differential evolution. Iterative detection algorithm based on IDMA is applied and the probability density function of message on factor graph is approximated by Gaussian approximation. Optimized LDPC code degree profiles are obtained by using the power profile. In this way the optimization method of LDPC code degree profile is obtained in unequal power cases by using iteration.4?The optimization method of LDPC code degree profile and spreading profile in MAC is studied. Iterleaver is applied to identify user and iterative detection and decoding is used The scheme to obtain spread value of every variable node from the edge spreading profile in unequal spreading case is discussed. The message update algorithm of different kinds of nodes on factor graph is derived. And LDPC code degree profile and spreading profile is obtained by applying the algorithm.