Channel estimation for turbo decoding on AWGN channels

We propose a simple channel estimation scheme for turbo decoding on additive white Gaussian noise (AWGN) channels without channel side information. The proposed estimator is based on statistical computations on the received block of data and provides an estimate of the channel reliability factor. It will be shown that the maximum likelihood (ML) algorithm cannot be employed for the estimation of the reliability factor without any side information (pilot symbols or a priori knowledge of the received data which are derived iteratively from the turbo decoder). Simulation results show that the performance of the turbo decoder is not adversely affected by this simple scheme compared with other complex methods. Also it is shown that the performance of the proposed estimator is very close to the performance of the estimators using any side information.