Robot Force/Position Tracking on a Surface of Unknown Orientation

The problem of robot force and position trajectory tracking is revisited in the case of an uncertain mapping of a surface into the robot space; then, although it is possible to define the desired trajectories with respect to the constraint surface, the lack of knowledge of the constraint direction in the robot space, means that the position and force control subspaces are uncertain. Such a case arises when for example the surface is misplaced. A novel adaptive controller is proposed using estimates of the constraint surface normal direction that converge to the actual value; the controller drives the actual force and position errors to zero given a persistently excited desired velocity on the surface. The performance of the proposed controller is demonstrated by a simulation example.