Sparse Directional Estimation of the Time-Varying Underwater Acoustic Channel

Abstract

This thesis focuses on the directional estimate of the underwater acoustic channel using modal beamforming with an orthogonal frequency division multiplexed (OFDM) communication system. An underwater channel model and simulator is first presented that is capable of implementing the time-scaling associated with the wideband time-varying acoustic channel. Modal beamforming is then introduced as a unique wideband beamforming method that makes use of circular arrays. The last Section explores the effect of a time-varying acoustic channel on an OFDM signal and produces a frequency domain channel that models the distortions that are introduced. This frequency domain model is then used to create a sparse approximation dictionary that, when coupled with the modal beamformer and a sparse approximation technique, is capable of producing accurate sparse directional channel estimates. These estimates include the complex amplitude, delay, time-scaling factor, and direction of arrival for each cluster arrival in the channel. This directional channel estimate is shown to greatly improve the performance of an OFDM receiver in a time-varying underwater acoustic channel when compared to classic estimation techniques.