Performance of adaptive MLSE equalization on multipath fading channels with imperfect channel estimates

In the literature, the performance of maximum-likelihood sequence estimation (MLSE) equalization on multipath fading channels is analysed under the assumption of perfect channel estimates. In reality, this condition is difficult to attain, especially for fast fading channels. It is therefore of great significance to study the performance degradation due to imperfect channel estimates. In this paper, we consider a practical MLSE equalizer in conjunction with a least mean square (LMS) channel estimator and a simple channel predictor. A new upper-bound on block error rate is derived for multipath Rayleigh fading channels.<>