Performance of a soft-output adaptive equalizer combined with soft-decision metrics generator based on differential detection

This paper proposes a soft-output adaptive equalizer combined with a novel soft-decision metrics generator based on differential detection, and assesses its performance by computer simulation and hardware experiments. The proposed soft-decision metrics generator performs differential detection by using the equalizer's soft-decision output and weights its results with the short-term average of squared errors in the equalizer. We confirm that the proposed scheme achieves good performance with lower complexity in generating the soft-decision metrics for Viterbi (1979) decoding compared with the conventional scheme.