Tip velocity tracking control for elastic manipulators

grantor: University of TorontoA unique approach to tip velocity control of an elastic robotic manipulator is presented. This method has potential application in teleoperation control and in applications where the trajectory is generated in real-time. Control is effected by reducing the tip velocity tracking error between the desired tip velocity and the measured tip velocity. Thus, in teleoperation, the concept of dead reckoning is used, so that while the manipulator's desired tip velocity is specified, the goal is for the tip to follow a path, leading toward a desired terminal position. This is done by concentrating the controller's effort on the manipulator tip while allowing the manipulator's links to deform., The controller utilizes a gain scheduling scheme to arrive at an appropriate feedback law. The control algorithm is implemented using a parallel-processing scheme on a multiprocessor system which consists of INMOS Transputers® . Experimental results are obtained using Radius, the Space Robotics Laboratory Facility at the University of Toronto Institute for Aerospace Studies. Radius was designed to serve specifically as a test-bed for the development and evaluation of control methods for elastic manipulators. For the present work Radius is configured as a two-link manipulator with both links structurally flexible. One of the trajectories used for the simulations and experiment is a square, which is quite difficult for an elastic manipulator to execute. This box trajectory is more demanding than trajectories typically executed by Canadarm and so link elasticity becomes an important consideration. The effect of elasticity is demonstrated using an independent joint PD controller. In simulation, excellent tip velocity tracking was achieved using the proposed controller. In the experiment, reasonable accuracy in following the desired tip path was attained, however, disturbances, mainly a result of unmodeled joint dynamics, caused degraded performance. The controller may be improved by making it more robust to disturbances and by adding control terms to assist in minimizing the tip position tracking error. Notable contributions include the design, parallelization, and parallel implementation of the tip velocity tracking controller. In addition, experimental results were obtained and a preliminary stability investigation of the controller was done.Ph.D