Sparse MAP equalizers for turbo equalizations

The turbo equalization scheme provides nearly optimal performance, but when it is employed in sparse multipath channel environments, the equalizer complexity can prohibit its use. In this paper, we develop a parallel trellis n-tap soft-input soft-output maximum a posteriori (MAP) equalizer which can consider n dominant taps of sparse channels with reasonable complexity. It is shown that turbo equalization employing the n-tap MAP equalizer can nearly achieve the asymptotic performance of the optimal maximum likelihood (ML) detector for the combined channel encoder and intersymbol interference (ISI) channel assuming perfect ISI cancellation. It is also shown that increasing n has favorable effects on the convergence of turbo equalizations employing n-tap MAP equalizer by EXIT chart analysis.