Improving faded turbo code performance using biased channel side information

Turbo decoding is based on the concept of maximum likelihood decision on blocks of data. The decision on a bit affects the decoding of subsequent bits. Fading is a type of corruption whose effect can be extended over a period of multiple bits, resulting in significant performance degradation for turbo codes. In this paper, a Markov process is utilized to model a fading channel, and bit error rate (BER) performance using BPSK with turbo code is examined. A channel interleaver is used to improve performance. In addition, a technique of using biased channel side information for turbo decoding is introduced. It is demonstrated that this technique can further improve performance.