Generalized decision feedback equalizer for approximate maximum likelihood bit detection of digital signals
Akira Sano, JUN-ICHI YAMAZAKI · International Journal of Electronics · 1982
A new type of non-linear equalizer which gives the maximum Likelihood (ML) bit estimates approximately is presented and its error rate performance is discussed. The structure of the non-linear equalizer is derived by introducing the idea of the decision feedback to the ML bit estimation. Since it is implemented by simple digital filters and decision slicers, we can greatly reduce this computation and storage required in the ML bit estimation and significantly improve the bit error probability as compared to the conventional decision feedback equalizer.