Preprocessing for Iterative ML-Detection in Frequency-Selective MIMO Channels

We investigate the use of the minimum mean-squared- error decision-feedback-equalizer (MMSE-DFE) as a preprocessor to reduce the computational complexity of symbol detection in frequency-selective multiple- input multiple-output (MIMO) channels. The preprocessor delays and focuses the received signal energy which reduces the effects of inter-symbol- interference and multiple-access-interference. This property of the MMSE-DFE preprocessor is exploited to order and to reduce the number of symbol candidates considered by an iterative maximum- likelihood (ML) detector. Simulations show that the proposed detection scheme can achieve lower bit- error-rates than existing reduced-complexity algorithms in overdetermined and underdetermined receive scenarios. Near-optimum performance is achieved when the number of symbol candidates is increased.