End-to-End UWA OFDM Communication System Based on LSTM-DRSN

Orthogonal Frequency Division Multiplexing (OFDM) is a multi-carrier transmission technique characterized by high spectral efficiency and strong resistance to multipath interference. It has been widely applied in underwater acoustic (UWA) communication. With its powerful feature extraction and fitting capabilities, deep learning offers new opportunities for UWA OFDM communication. In this paper, we propose an OFDM autoencoder based on a Deep Residual Shrinkage Network (DRSN) and Long Short-Term Memory (LSTM) network to address the challenges posed by complex and dynamic UWA communication environments. The residual shrinkage module employs a soft thresholding function as a nonlinear transformation layer, which is embedded into the deep structure to effectively eliminate noise-related features and enhance the network’s ability to extract useful information. Simulation results demonstrate that the proposed model outperforms convolution-allayer-based OFDM communication systems regarding block error rate (BLER) performance in UWA channels.