A Qualitative Navigation System for Autonomous Underwater Vehicles

Navigation is a critical process for the successful operation of any autonomous underwater vehicle. A very small number of feasible solutions to autonomous navigation currently exist which are in use today. These solutions bring with them strict limitations on the duration of deployment, beyond which the vehicle will run the risk of becoming lost indefinitely or the mission disrupted. The source of these limitations comes from the inability of the vehicle to localize underwater in the absence of external localization aids. We present a novel approach to the navigation of an autonomous underwater vehicle through the use of side-scan sonar image registration. The Qualitative Navigation System introduced in this thesis implements an image-based route following strategy. Image registration provides the vehicle with a real-time estimate of its displacement and orientation relative to a previously trained route. It is the objective of our strategy to utilize the output of the image registration process to continuously guide the vehicle into correct alignment with the route. This simplified approach to navigation sidesteps the problems inherent with attempting to maintain a pose estimate within a global reference system. Simulation results are provided which validate the proof of concept for our approach.