MLP/BP-based Decision Feedback Equalizers with High Skew Tolerance in Band-Limited Channels

A multi-layered perceptron neural network with backpropagation algorithm (MLP/BP) is realized as a waveform equalizer for distorted nonreturn-to-zero (NRZ) data recovery in band-limited channels. Moreover, the proposed approach can tolerate sampling clock skew and channel response variance. According to simulation results, the proposed design can recover severe distorted NRZ data with better performance than LMS DFEs in the band-limited channel that the data rate is ten times as much as the channel bandwidth. Under the 20% channel response variance and the 30% sampling clock skew, the proposed approach can provide an acceptable performance. Key-Words: Clock Skew, Decision-Feedback Equalizer (DFE), Intersymbol Interference (ISI), Multi-Layered Perceptron Neural Network with Backpropagation Algorithm (MLP/BP), Minimum Mean Square Error (MMSE), and Nonreturn-to-Zero (NRZ).