CHANNEL STATE INFORMATION ACQUISITION FOR ADAPTIVE UNDERWATER ACOUSTIC NETWORKING

Abstract

This dissertation investigates the design aspects of an adaptive cross-layer architecture to optimize the energy efficiency, the spectral efficiency, and the reliability of underwater acoustic multihop relaying networks by utilizing channel state information (CSI). Specifically, an energy efficient channel-aware routing protocol for reliably relaying data packets, as well as a media access control to maximize the network throughput and maintain connectivity are described. These tasks are approached by predicting CSI using a novel data-driven probabilistic model. As the main contribution, a CSI acquisition approach based on a Markov chain process is proposed that exploits information from the physical environmental conditions, including the tide phase and flow, to improve the accuracy of channel characteristics predictions. Specifically, the method is intended to obtain the channel characteristics, including the gain, delay, Doppler spread, as well as the standard deviation of intrapaths delays in time varying conditions.