Underwater Acoustic Orthogonal Frequency-Division Multiplexing Communication Using Deep Neural Network-Based...

Area of Science:

• Electrical Engineering
• Computer Science
• Signal Processing

Background:

• Underwater acoustic (UA) communication faces challenges due to multipath propagation causing significant delay spread.
• Conventional orthogonal frequency-division multiplexing (OFDM) receivers struggle with channel estimation due to non-linear frequency responses between pilot subcarriers in UA channels.
• The unknown nature of underwater channel delay profiles hinders precise modeling of this non-linearity.

Purpose of the Study:

• To propose and evaluate a deep neural network (DNN)-based receiver for underwater acoustic (UA) communication.
• To address the challenge of non-linear channel responses in UA communication by leveraging neural network capabilities.
• To demonstrate the performance improvement of a DNN-based receiver over conventional methods in real-world UA environments.

Main Methods:

• Development of a deep neural network (DNN) architecture for underwater acoustic (UA) communication receiver.
• Training the neural network (NN) to learn and compensate for channel non-linearity inherent in UA communication.
• Comparative performance analysis against conventional least-squares (LS) estimator-based receivers using data from river trials.

Main Results:

• The DNN-based UA communication receiver demonstrated superior performance compared to the conventional least-squares (LS) estimator.
• River trials in Western Australia validated the effectiveness of the proposed DNN-based receiver.
• The results indicate successful learning and compensation of non-linear channel effects by the neural network.

Conclusions:

• Deep neural network (DNN) receivers offer a promising solution for revolutionizing underwater acoustic (UA) communication.
• The proposed DNN-based approach enables higher data rates, improved reliability, and better adaptability to dynamic underwater conditions.
• Future UA communication systems can benefit significantly from the integration of DNN receiver technology.