Blind Channel Estimation for Systems Employing Alamouti Coding

This paper presents blind channel estimation for Alamouti coding operating in time-correlated Rayleigh channels. A decision algorithm first obtains primary data estimates of the transmitted symbols based on the constrained linear minimum mean square error criterion, and these are applied to subsequent optimal minimum mean square error channel estimation. Performance is evaluated by simulation under the assumption of perfect timing, and presented against the known results of optimal coherent detection. The error performance is shown to be better than the Kalman filtering approach. The impact of finite cross correlations between the receive antennas is also investigated by simulation.