A new adaptive channel estimator for turbo decoding in Rayleigh fading channels

A new adaptive channel estimation filter based on the variable step-size LMS algorithm is shown to perform close to the optimum Wiener filter without assuming a-priori knowledge of the channel fading rate. It is unlike other fixed characteristics filters such as the equal weight moving average and FIR filters designed using the windowing method. Under a non-stationary Rayleigh fading channel where the fading rate varies, the proposed adaptive filter is also able to track the time-varying channel, achieving better performance than the fixed characteristics filters.