ABSTRACT
Robot manipulators are basically multi-degree-of-freedom positioning devices. The robot, as the “plant to be controlled,” is a multi-input/multi-output, highly coupled, nonlinear mechatronic system. The key in robot control system design is to model the most dominant dynamic effects for the particular manipulator under consideration and to design the controller so that it is insensitive or robust to the neglected dynamics. The notion of feedback linearization of nonlinear systems is a relatively recent idea in control theory, whose practical realization has been made possible by the rapid development of microprocessor technology. Feedback linearization is a useful paradigm because it allows the extensive body of knowledge from linear systems to be brought to bear to design controllers for the nonlinear systems. Passive force feedback requires neither a force sensor nor a modified programming and control system and is therefore simple and cheap.
