Semi-blind training-like estimation of spatio-temporal filter coefficients for finite alphabet signals

A semi-blind algorithm for equalization and jammer rejection for short-burst wireless communications is presented. It is based on the idea of using a few information symbols to virtually enlarge the length of the accessible training sequence. The proposed algorithm is a simple and useful solution in applications where the length of the available training sequence is shorter than the minimum required by existing training-based or semi-blind algorithms to reach a valid solution. Simulation results for spatio-temporal processing of 4-QPSK signals transmitted by simultaneous intra-cell users demonstrate the ability of the proposed algorithm