Performance of BCJR-DFE detectors over recording channels using pattern-dependent noise prediction

A noniterative receiver, namely Bahl-Cocke-Jelinek-Raviv (BCJR)-decision feedback equalizer (DFE), for achieving near capacity performance for intersymbol interference channels was proposed earlier. In this paper, we investigate the performance of BCJR-DFE receivers for recording channels with pattern-dependent media noise. To account for the media noise, we integrated pattern-dependent noise prediction (PDNP) into the BCJR-DFE receiver and evaluated its performance by comparing it to conventional iterative (turbo)-coded recording channels. Numerical simulations suggest that PDNP BCJR-DFE receivers provide superior or comparable performance as a turbo receiver, albeit with lower computation complexity and longer processing delays.