A space-time FIR equalizer training algorithm for MIMO-OFDM systems

In this paper, we propose several new direct training-based methods for FIR space-time equalizer design in multiple-input multiple-output orthogonal frequency division multiplexing (MIMO-OFDM) systems employing spatial multiplexing. We propose a two-stage receiver structure consisting of an initial spacetime equalization for suppressing the co-channel interference (CCI), shortening the channel, and mitigating asynchronicity followed by parallel equalization and detection in the frequency domain. The coefficients of the equalizers and the shortened post-equalization channel response are jointly derived from the training data of the desired users under various constraints. Simulations show that training based methods have good error rate performance and are robust to MIMO co-channel interference.