Feature relative navigation of an autonomous underwater vehicle

This thesis presents a technique for adaptively mapping features in the ocean using an autonomous underwater vehicle (AUV). An adaptive behavior was developed in response to the challenge of locating and mapping an unknown number of target fea-tures without the aid of a priori maps. The design is an extension of the concept of state configured layered control, a modified form of behavior-based intelligent control of AUVs. The new adaptive feature mapping behavior incorporates planning and mapping capabilities which allow the vehicle to alter its trajectory on-line in response to sensor data. New waypoints are selected by evaluating the expected utility of vis-iting a given location (is it close to previously mapped portions of a feature while not having been visited before?) balanced against the expected cost (weighted distance and heading change) of visiting a particular cell. The technique has been developed based on the assumptions of a point sensor attached to a non-holonomic, dynamically controlled survey-class AUV such as the Odyssey II. Testing has been conducted in a simulation of the Charles River basi