An improved fractionally spaced decision feedback equalizer structure

Equalizers are used to combat inter symbol interference (ISI). Adaptive equalizers are either symbol spaced or fractionally spaced. Fractionally spaced equalizers offer many advantages over Symbol spaced equalizers. The most well known and widely used algorithm to adapt the tap weights of the FSEs is the ubiquitous Least Mean Squares (LMS) algorithm. More recently other algorithms have been presented in the literature which use a more general error criteria to adapt the tap weights of the adaptive filters. The Least Mean Mixed Norm (LMMN) family of algorithms has been found to yield improved performance over the LMS algorithm in certain cases. In the first part of this work we consider the use of three different algorithms from the family of LMMN algorithms to adapt the fractionally spaced decision feedback equalizer (DFE). We have used white-gaussian, colored-gaussian and uniform noise environments to compare the performance of these algorithms. In the second part of this work an improved fractionally spaced decision feedback equalizer (FS-DFE) structure is presented and evaluated using simulations. Simulations have been carried out for static and time-varying channels to check the performance of the proposed structure in static and time-varying environments. It has been found that the proposed scheme is giving better results than the conventional FS-DFE structure without any increase in the complexity.