Signal Demodulation without Channel Equalizer Using Machine Learning Techniques

With the unprecedented availability of computing and data resources, there has been a widespread trend in using machine learning techniques in the field of communication as well. In the existing communication system, signal demodulation in multipath channel works with the aid of equalizers to mitigate the effect of inter-code cross talk and inter-symbol interference. Owing to the ability other machine learning models to learn the complex structure in feature space, deep learning architectures such as Deep Belief Network(DBN) and Stacked Auto Encoders(SAE) are used for demodulation in this paper. The experiments were conducted in a multipath Rayleigh channel. Simulation results shows the feasibility and hegemony of the deep learning model over the conventional demodulation techniques using equalization.