Blind maximum-likelihood sequence estimation with decision feedback

This paper presents a new reduced-state, decision feedback Viterbi equalizer (RSSDF-PSP) which works blind. In the RSSDF-PSP, the whole length of the channel impulse response (CIR) is divided into two parts. The first part is handled by the MLSE and the second part is handled by a soft decision feedback filter. The first part channel estimate was obtained by the blind per-surviving processing (PSP) algorithm, the second part channel estimate was obtained by the decision feedback filter. The complexity of the MLSE is significantly reduced by the second part CIR truncation. Simulation results show the RSSDF-PSP has nearly the same performance as the MLSE. It provides the best tradeoff between complexity and performance of MLSE.