| Autonomous Underwater Vehicles (AUVs) are extensively being used by the scientific, oil and gas, and military communities. Many of the missions require the vehicle to function in complex, cluttered environments; to react to changing environmental parameters; and, to find a collision-free path through a workspace containing a significant number of obstacles. Many of the AUV missions currently involve high risk for human lives and excessive costs. State-of-art vehicles are not maneuverable enough to successfully accomplish most of the desired tasks. Desirable vehicle control capabilities include the ability to drive at very low, controllable speeds, the ability to maintain a set distance and attitude (pitch and roll) relative to some surface for optimal sensor (both sonar and video) effectiveness, and the ability for the operator to intervene to change the mission activities. Moreover, a vehicle capable of rotating in place or having a fraction of a meter turning radius is needed to conduct desired missions. Novel controllers to implement these specific behaviors are expected to be nonlinear due, for example, to the fact that the vehicle is maneuvering at nonzero attitude while translating parallel to the surface. A specific mission that this research addresses is ship hull inspection. This dissertation works through the details of a method to control the vehicle's attitude and translation relative to a surface. The surface of interest for example being a ship hull.;This dissertation describes the derivation, design, simulation, and implementation of a Behavior Based control system. Each behavior is designed using a command filtered backstepping (CFBS) approach. Each behavior and the switching among behaviors is provably stable in the sense of Lyapunov. We use the results from Hybrid System Control in order to prove stability during behavior switching, and thus the overall control system stability. This dissertation presents the simulation and in-water testing results of our control design applied to an AUV. |
