Search methods for an autonomous underwater vehicle using scalar measurements

Abstract

The continuing development of the autonomous underwater vehicle as an oceanographic research tool has opened up the realm of scientific possibility in the field of deep ocean research. The ability of a vehicle to travel to the ocean floor untethered, collect data for an extended period of time and return to the surface for recovery can make precise oceanographic surveying more economically practical and more efficient. This thesis investigates several scalar parameter searching techniques which have their basis in mathematical optimization algorithms and their applicability for use specifically within the context of autonomous underwater vehicle dynamics. In particular, a modified version of the circular gradient evaluation in the simulated environment of a hydrothermal plume is examined as a test case. Using a priori knowledge of the expected structure of the scalar parameter contour is shown to be advantageous in optimizing the search.