Bayesian blind equalization for coded waveforms

The authors present a novel blind equalization algorithm for coded waveforms received in the presence of intersymbol interference (ISI). The algorithm is an approximation to the optimum maximum a posteriori (MAP), symbol-by-symbol detector for a priori unknown channels, and consists of a bank of Kalman filter channel estimators, each conditioned on a subsequence of code symbols. For convolutional codes, a separate channel estimate is computed for every effective state in the convolution of the encoder and channel. It is shown that the MAP decoding metrics can be computed using the Kalman filter innovations, which are combined in a suboptimum Bayesian formula. The performance of the algorithm is evaluated using simulations and comparisons of channel estimation error