Fast simulation of turbo codes over Rayleigh fading channels

In this paper we apply importance sampling (IS) for efficient performance evaluation of turbo codes over Rayleigh fading channels. Our approach is based on the most probable error path (MPEP) method which shifts the input probability density function (pdf) to the mode of the optimal biased distribution where most probable errors are located. In particular, we show how to bias the pdf of the AWGN process in order to maximize the computational efficiency gain. It is demonstrated that the proposed method requires much smaller sample sizes to achieve the same accuracy required by a conventional MC estimator.