CONSTRAINED MOTION TASK CONTROL OF ROBOTIC MANIPULATORS

Control of robotic manipulator during constrained motion task executio n is the subject of this brief paper. Our previous work in this area a ddressed control of manipulators during constrained motion subject onl y to kinematic position constraints. This paper addresses the problem of manipulator constrained motion control for the case of arbitrary ki nematic constraints acting on the system as well as dynamic parameter uncertainty. Hence, this brief paper represents a generalization of ou r previous results. A method is presented which permits the asymptotic regulation of both generalized forces and position of the manipulator . Regulation of these outputs is achieved in the presence of constant unmodelled disturbances which may act on the system. The control syste m synthesis is based on a general theory associated with the control o f descriptor variable systems. The linearized manipulator dynamics is decomposed into a ''slow'' and ''fast'' subsystem. The slow subsystem, corresponding to the manipulator states that lie within the subspace of constrained states, is stabilized to yield an asymptotically stable system. The dynamics of the fast subsystem may be ignored, as shown i n the paper. Synthesized from a linearized model of the manipulator dy namics, the method is valid only in a neighborhood about the point of linearization. It is assumed in this paper that the kinematics of the manipulator and the contact environment are known. A numerical example serves to illustrate the method presented here.