Neural Network Equalizer and Decoder for LinearBlock Codes in ISI Channels

This paper delves into the design and evaluation ofneural decoders, efficient neural network-based algorithms fordecoding linear block codes like the Hamming code. The studyexplores the influence of neural network architectureparameters, including the number of layers and neurons, ondecoding performance in AWGN, Rayleigh fading, and ISIchannels. Specifically, the investigation focuses on (۷,۴) and (۸,۴)Hamming codes. To counteract the impact of the ISI channel, aneural equalizer is employed. Comparative analyses with MLdecoding and ZF equalizers are conducted to assess theeffectiveness of the proposed approach.