BOLTZMANN ENTROPY FOR UNDERSEA TERRAIN AIDED NAVIGATION

Abstract

The ability to navigate accurately without Global Positioning Systems (GPS) or beacon systems remains an elusive yet important capability for unmanned systems. Reliance on external beacon systems, like GPS and long baseline, create system vulnerabilities if those beacon systems are damaged or removed. This research explores the ability to improve Undersea Terrain Aided Navigation (UTAN) with respect to coverage path planning and creates a robust coverage path planning algorithm utilizing a new paradigm in configuration entropy. It highlights a classic challenge in the path planning problem between exploration and exploitation. Exploration ensures complete coverage of the region, while exploitation utilizes terrain features to reduce positional uncertainty of the Autonomous Underwater Vehicle (AUV). Using stochastic estimates of the vehicle location and map accuracy, it is possible to combine the variances of each together with exploration and exploitation policies to develop a coverage plan for UTAN. To do so requires development of a terrain entropy measure to influence the path planning algorithm. The thesis presents a new entropy measure that is a key architectural component for an information-theoretic framework to develop near-optimal paths for the AUV coverage problem.