Underwater Acoustic Communication Receiver Using Deep Belief Network

Underwater environments create a challenging channel for communications. In this paper, we design novel receiver system by exploring the machine learning technique–Deep Belief Network (DBN) – to combat signal distortion caused Doppler effect and multi-path propagation. We evaluate performance of proposed in both simulation experiments sea trials. Our comprises DBN based de-noising classification received signal. First, is segmented into frames before each these individually pre-processed using pixelization algorithm. Then, algorithm, features are extracted from used reconstruct Finally, reconstructed occurs. does show better channels influenced propagation with improvement 13.2dB at 10 ?3 Bit Error Rate (BER).