Research on Underwater Acoustic Channel Denoising Algorithm based on Auto-Encoder

Aiming at the complexity of underwater communication environment, a denoising algorithm based on autoencoder (AE-DNA) is proposed. In this algorithm, the autoencoder is used as the decoder, and the network model is established by learning the training data. At the same time, in order to improve the processing ability of the network, the idea of multiple inputs and multiple outputs is introduced to convert the input dataes and the output dataes. In addition, in order to improve the generalization ability of the network, when transmitting high-order signals, signal transformation is used to reduce the complexity of the signal. The simulation results of underwater acoustic channel show that compared with the denoising algorithm based on RBF neural network (RBF-DNA), the proposed algorithm has faster convergence speed, smaller mean square error and better decoding ability.