Precise control of underwater vehicle manipulator systems using iterative learning control

This paper presents a feedforward compensation approach for underwater vehicle manipulator systems (UVMSs) in order to realize precise and fast motion control. Iterative learning control (ILC) is used for controlling an underwater manipulator in this paper. Thruster forces calculated by using static Moment-Voltage (Control input[V]) characteristics of thrusters eliminate reaction moments acting on a vehicle during manipulator motion. Experiments were conducted using an underwater vehicle with a 1-degree of freedom (D.O.F) manipulator developed at Tokai University, Japan. In case of non-compensation, the vehicle attitude error was up to 10[deg] after a control period. On the other hand, the feedforward compensation approach proposed here could reduce the error down to plusmn 2.0[deg].