EQUALIZED MAXIMUM-LIKELIHOOD RECEIVER WITH A UNIT ENERGY CONSTRAINT

In order to reduce the complexity of implementing the Maximum Likeliho od Sequence Detector (MLSD), we propose the Equalized Maximum Likeliho od (EML) receiver which still approaches optimum performance, In this scheme, a feed-forward filter is used to shape the channel to the targ et channel response, to which the Viterbi detector is matched. A unit energy constraint is imposed on the target response in a magnetic reco rding channel as well as a DC free constraint. A comparison to Partial Response Maximum Likelihood (PRML) is made in a magnetic recording sy stem. Simulation results indicate that significant performance improve ment over PRML can be achieved especially at high recording densities. A DC free constraint on the target response is shown to yield a bette r performance comparing to solutions without the constraint. Implement ation issues such as the adaptive algorithm for finding the optimum se ttings recursively and the complexity of the Viterbi detector are also investigated, Multiplication operations in the Viterbi detector can b e avoided by employing a look-up table.