迭代软译码中几个关键问题的研究

The study of optimal (or semi-optimal) soft-decision decoding with low complexity is presented. Three achievements about it are obtained. First, the asymptotic performance of the iterative soft-in/soft-out decoding algorithm and the influence of the interleaver on the convergent performance are presented. Second, the mathematical element of Turbo codes is exposed. Third, much effort is given to the improvement of the "error-floor" performance, and theoretical foundation is provided on the constructure of high performance concatenated code. Another main focus is the study of the encoding and decoding of LDPC codes. Appropriate encoding and decoding methods which are fixable to the mobile communication system are obtained. Great effort is spent on the reduction of the complexity of encoding and decoding of LDPC codes so as to accommodate with the demand of real-time communication system. These results can help to improve the theoretical scheme of LDPC codes and could be a guidance for the LDPC codes in using as well. The resolving of these problem can be a great significance for the improvement of the coding theory and the application of Turbo codes and LDPC codes.

本项目拟研究并解决迭代软输入／软输出算法的收敛性问题，研究Ｔｕｒｂｏ码与Ｓｈａｎｎｏｎ随机码的关系和交织器对迭代译码算法收敛性的影响。这些问题是编码理论中的一个重要且较为困难的课题。问题的解决直接关系到Ｔｕｒｂｏ码原理在相关领域的应用。通过对此问题的研究，还有望得到更好的算法和适用于工程的迭代终止的判决条件。