GUIDANCE AND CONTROL OF MARINE CRAFT: AN ADAPTIVE DYNAMIC PROGRAMMING APPROACH By PATRICK S. WALTERS A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY

of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy GUIDANCE AND CONTROL OF MARINE CRAFT: AN ADAPTIVE DYNAMIC PROGRAMMING APPROACH By Patrick S. Walters May 2015 Chair: Warren E. Dixon Major: Mechanical Engineering Advances in sensing and computational capabilities have enabled autonomous vehicles to become vital assets across multiple disciplines. These improved capabilities have led to increased interest in autonomous marine craft. As the technologies of these vehicles mature, there is a desire to improve the performance of their motion control systems so that these vehicles can better achieve their mission objectives. Path planning, station keeping, and path following are critical control objectives of marine craft that enable autonomous docking, surveying, etc. Improving the performance of these control objectives for marine craft directly correspond to improved range, endurance, accuracy, and robustness in different environmental conditions. Model-based adaptive dynamic programming has also seen considerable attention in the last decade, as a method of generating approximate optimal policies for classes of general uncertain nonlinear systems. Recent advances have made model-based adaptive dynamic programming a viable option for the control of uncertain complex systems such as marine craft. These recent advances motivate the exploration of model-based adaptive dynamic programming, as a means of improvement for motion control systems of autonomous marine craft. This dissertation focuses on the application of adaptive dynamic programming to the motion control of marine craft. Specifically, Chapter 1 provides motivation for the application of model-based adaptive dynamic programming to control objectives