Optimum space-time processors with dispersive interference-unified analysis and required filter span

We consider optimum space-time equalizers with unknown dispersive interference, consisting of a linear equalizer that both spatially and temporally whitens the interference and noise, followed by a decision-feedback equalizer (DFE) or maximum-likelihood sequence estimator (MLSE). We first present a unified analysis of the optimum space-time equalizer, and then show that, for typical fading channels with a given signal-to-noise ratio (SNR), near-optimum performance can be achieved with a finite-length equalizer. Expressions are given for the required filter span as a function of the dispersion length, number of cochannel interferers, number of antennas, and SNR, which are useful in the design of practical, near-optimum space-time equalizers