9701850 Lee Our primary research objective is to investigate and develop intelligent control strategies for underwater robotic vehicles (URVs) with manipulator workpackages. The motion of the manipulator, which is attached to the vehicle's main body, affects the motion of the vehicle, whose dynamics are also subject to parameter uncertainties, changes in payload and environment, and nonlinear behavior. It is necessary to develop an intelligent control system for such vehicles to provide: automatic compensation for the errors of the vehicle motion, due to manipulator motion and underwater currents. coordinated control of both vehicle and manipulator using deliberate vehicle motion; such motion will help task performance and add the degrees of freedom of the vehicle to those of the manipulator workpackage; learning and adaptation capabilities to parameter uncertainties and changes in the environment; and close coupling of control knowledge transfer between the high-level and low-level subsystems of the overall URV control system. The study: extends a theoretical modeling of the underwater robotic vehicle and manipulator, including dynamic interaction between links and vehicle main body, kinematic redundancy, contact stability of the system in contact with the environment, and fixtureless manipulation; develops a theoretical framework for intelligent control strategies based on fuzzy neural network (FNN) control using fuzzy clustering techniques and on-line reinforcement learning technique; and includes experimental demonstration of the proposed approach on the University of Hawaii's underwater robotic vehicle, ODIN (Omni-Directional Intelligent Navigator). ***
