Linear Equalizers for Turbo Equalization A New Optimization Criterion for Determining the Equalizer Taps

This paper investigates the subject of turbo equalizations in which a receiver combines the equalization and decoding process in an iterative fashion. In order to save in complexity, a linear equalizer with two inputs, one for the received signal and one for the extrinsic information, replaces the aposteriori probability (APP) processor used for the equalization. The equalizer tap gains are determined as a mean square approximation to the APP. We show that such optimization criterion for determining the equalizer tap gains is better than minimum mean square error (MMSE) optimization of such taps. Simulation results are presented for parallel-concatenated turbo code with BPSK modulation over channels that introduce severe amplitude distortion.