Reduction of PAPR by Convolutional Neural Network with Soft Feed-back in an Underwater Acoustic OFDM Communication

The intelligent neural networks have proved to be the most efficient and less complex for orthogonal frequency divisional multiplexing (OFDM) communication in the last decade. Therefore, in this paper, a convolutional neural network method with soft feed-back is employed for the reduction of peak to average power ratio (PAPR) in an underwater acoustic OFDM communication system. The PAPR performance shows a great enhancement by using a convolutional neural network. The simulation results prove the excellency of the overall system, as it is the simplest, intelligent, and reliable for mitigation of PAPR, considering an equitable performance of bit error rate (BER).