Learning control of cooperating robot manipulators handling an unknown object

Abstract

A learning control scheme is proposed for learning of cooperating multiple-robot manipulators which handle a rigid object under the uncertain operating conditions such as parametric uncertainty, unknown external disturbance and the commonly held object being unknown. The learning controllers drive the multiple-robot manipulators to follow the desired Cartesian trajectory with the desired internal forces to the unknown object. It is also shown, after perfect learning, the learning controllers provide an approximate inverse dynamics solution along the desired object trajectory with the desired internal force.