Hybrid learning control for constrained manipulators

Abstract

The problem of controlling a manipulator constrained to move on a rigid and frictionless surface is dealt with in this paper. A desired end-effector trajectory, lying on the surface, is given, together with a reference evolution for the contact force acting between the terminal device and the surface itself. The dynamic mathematical model of the robot is instead assumed to be unknown, together with the geometric characteristics of the surface surrounding the desired trajectory. The control signal allowing the execution of the task (i.e. the tracking of the desired trajectory with the specified contact force profile) is found by means of a trial-and-error procedure. The effectiveness of the learning method is shown by means of a fully non-linear analysis. Finally, some simulation examples are reported.