An improved MMSE turbo equalization algorithm in frequency domain

Turbo equalization is an iterative technique for the equalization and decoding of data that has been transmitted over the inter-symbol interference (ISI) channels. A number of suboptimal and low-complexity equalization algorithms have been developed to replace the optimal maximum a probability (MAP) equalizer. In this paper, we present an improved turbo minimum mean-squared error (MMSE) equalization algorithm in frequency domain. It is motivated by the existing turbo MMSE frequency domain equalization algorithm and further improved to perform well even in the severe frequency-selective channels while keeping a comparable complexity. Simulation results show that the new algorithm achieves a performance gain of 1.6 dB in Proakis B channel and 2.4 dB in Proakis C channel in comparison with the existing algorithm.