End-to-End Underwater Acoustic Communication Based on Autoencoder with Dense Convolution

• DOI: 10.3390/electronics12020253
• Publication Date: 2023-01-04T07:15:42Z
• Authors: Fangtong Xie; Yunan Zhu; Biao Wang; Wu Wang; Pian Jin
• Keywords: Autoencoder; Underwater acoustic communication; Convolution (computer science); Feature (linguistics); Convolutional code

To address the problems of high complexity and poor bit error rate (BER) performance traditional communication systems in underwater acoustic environments, this paper incorporates theory deep learning into a conventional system proposes data-driven filter bank multicarrier (FBMC) communications based on convolutional autoencoder networks. The proposed is globally optimized by two one-dimensional (Conv1D) modules at transmitter receiver, it realizes signal reconstruction through end-to-end training, effectively avoids inherent imaginary interference system, improves reliability system. Furthermore, dense-block are constructed between Conv1D layers connected across to achieve feature reuse network. Simulation results show that BER method outperforms FBMC with channel equalization algorithms such as least squares (LS) estimation virtual time reversal mirrors (VTRM) under measured conditions certain moment Qingjiang River.