Abstract
The neural network equalizer (NNE) for the partial response maximum likelihood (PRML) system with a post-processor in perpendicular magnetic recording is studied. First, a design method of NNE is proposed for suppressing a high-level noise at the discrimination point. In our method, the NNE is designed by using a hybrid distributed genetic algorithm with a combination of a distributed genetic algorithm and a simulated annealing algorithm, and the suppression of a high-level noise and the simplification of NNE are attained. Then, the bit error rate performance of PR1ML system with a post-processor using the NNE is obtained and compared with those using a conventional NNE without a noise suppressing function and a conventional transversal filter as an equalizer. The results show that the performance of the proposed system is superior to those of conventional systems.
