Abstract

This thesis is concerned with the problem of dynamic control of kinematically redundant manipulators. A robot manipulator is said to be kinematically redundant when it has more degrees of freedom than are necessary to accomplish a particular task. Kinematic redundancy in a robot manipulator is a desirable characteristic since such manipulators have increased dexterity and versatility due to their self-motion. However, kinematic redundancy in a manipulator's structure presents a challenging problem since the richness in the choice of joint motions for the same end-effector trajectory complicates the manipulator kinematics and control problems considerably. The main objective of the work presented in this thesis is to design useful control strategies for kinematically redundant manipulators in order to enhance their performance to achieve the main task as well as the secondary tasks. In particular, in this thesis the following three problems are considered: First, following the impedance control approach, the problem of minimizing redundant manipulator collision impacts is addressed. The configuration control approach is used to reduce impulsive forces, while a simplified impedance control scheme is formulated to minimize rebound effects. Second, a new Cartesian control strategy for redundant flexible-joint manipulators, namely the hybrid Cartesian-joint control scheme, is proposed. The main idea in this scheme is to control not only the manipulator's end-effector, but also its links so as to achieve, specified positions and velocities for the end-effector and the links. Finally, a new application of kinematically redundant manipulators is proposed, namely, that of using redundancy resolution to compensate for joint flexibility. This redundancy resolution scheme is incorporated in a control strategy for redundant flexible-joint manipulators. The problem of possible algorithmic singularities is addressed, and a scheme is proposed which makes the controller robust with respect to such singularities.