Learning control for robot tasks under geometric constraints

A learning control scheme for robot manipulators whose endpoint is moving under geometrical constraints on a surface is proposed. In this scheme, the input torque command is composed of two signals updated separately at every trial by different laws. One is updated by the angular velocity error vector which is projected to the tangent plane of the constraint surface in joint space. The other is updated by the magnitude of contact force error at the endpoint. A theoretical proof of the uniform boundedness of velocity and position trajectories and the convergence of these to their desired ones is given, together with the convergence of force errors. Computer simulation remits by using a 3-DOF manipulator are presented to demonstrate the effectiveness of the proposed method