Low Complexity Iterative Decoding And Iterative System Testing In Digital Communications

Channel coding is an indispensable technique which guarantees the reliable information transmission in digital communications. Recently, based on iterative decoding strategy, significant progress has been achieved in channel coding theory, such as Turbo codes, Turbo-like codes and LDPC codes etc. However, the high decoding complexity of these codes is still an obstacle to many real applications. The complexity of iterative decoding algorithm comes from two aspects: one is the iterative process, and the other is component decoding algorithm itself. This thesis aims at investigating both complexity issues. In particular, our attention is paid to the complexity reduction of Turbo decoding by reducing the average number of iterations at the guarantee of the decoding performance, hard decision majority based algorithm and soft-decision sum-product algorithm for iteratively decoding complex rotary (CR) codes with low complexity, as well as the interleaver by-passed method to test the iterative Turbo systems.Firstly, the thesis discusses that the use of stopping criterions to reduce the average number of iterations as much as possible at the guarantee of the bit error performance. A joint early detection and early stopping criterion is proposed, which can reduce the decoding complexity of Turbo code effectively. By employing the build-in cyclic-redundancy-check stopping criterion, the joint early detection and early stopping criterion for short frame length Turbo codes is also proposed. The research results show that our algorithms can reduce much complexity than the existing methods at the same bit error performance.For Turbo codes adopted by W-CDMA, the half-iteration cyclic-redundancy-check criterion and cyclic-redundancy-check criterion for forward recuesive computing of Max-log-MAP component decoding algorithm are proposed, which achieves bigger complexity reduction ratio than known results. (The reduction ratio is 73% higher than best known results)Secondly, the thesis proposes a low complexity iterative sum-product algorithm and an iterative majority based algorithm for soft and hard decision decoding of complex rotary code. It is shown that the complex rotary codes employing iterative hard decision majority based decoding algorithm outperforms...