Turbo code performance on Rician fading channels

The performance of turbo codes on frequency non-selective slow Rician fading channels is considered. The analytical average upper bounds of turbo code performance over the channels with independent fading are derived. Furthermore, we developed a decoding metric for fading channels and log-likelihood ratios and extrinsic information for maximum a posteriori (MAP) and soft output Viterbi algorithm (SOVA) decoding methods. Finally, the derived upper bounds are evaluated and the bit error rates for turbo codes over independent and correlated fading channels are simulated. The performance of the MAP and SOVA iterative decoding methods, with and without channel state information is discussed. It is shown that the SOVA decoder has almost the same performance as the MAP decoder on correlated fading channels if ideal channel state information is used