Mitigating the Impact of Ocean Ambient Noise in an Underwater Acoustic Communication System

Abstract

In wireless communication links, receivers are typically designed to mitigate the variability in the propagation channel in presence of arbitrary white Gaussian noise. However, the variations in the instantaneous channel properties introduce high statistical variability in the communication link reliability. In fact, the underwater acoustic channel in the ocean is particularly sensitive to the changes in environmental conditions. Extensive research has been done on the propagation but it is typically assumed that ocean noise compares with terrestrial noise in the design of underwater acoustic receivers. This assumption does not accurately represent the impact of ocean ambient noise on the performance of underwater acoustic communication systems, particularly when using receiver arrays.

This dissertation studies the unique properties of oceanic ambient noise, particularly the variability in its directional properties. The application developed in this work focuses on both anthropogenic noise due to vessels and naturally occurring ambient noise within the channel bandwidth. To this end, the characteristics of these noise sources and their impact on the underwater acoustic link are discussed in this dissertation.

Firstly, using a compact receiver array, an acoustic source tracking procedure is designed to characterize the directional properties of vessel noise. This is achieved using a maximum-likelihood beamformer to estimate the bearing and a coherence-based matched-field processor to estimate the range of a vessel over its travel duration. Although the performance of most methodologies developed for characterizing vessel noise in literature are evaluated using computer simulations, the algorithm applied in this dissertation are tested with actual measurements of vessel noise from ocean experiments. It is observed that the noise directionality can be estimated accurately using a compact array but relies on the geometry of the array.

Secondly, noise models are developed to characterize the unique properties of naturally occurring ambient noise at a compact array of acoustic receivers. Synthetic ambient noise is generated with defined properties and validated against measured ambient noise.

Thirdly, the performance of a space-time receiver for signals processed in measured ambient noise is validated against signals processed in synthetic noise processes. It is observed that the variations in the space-time properties of ambient noise do not compare with the usual uncorrelated noise assumption in the design of an underwater acoustic receiver. Also, the bit-error rate of the space-time filter depends on optimizing the training and payload duration in the received signal to adapt to the time-varying property of ocean ambient noise.