Experimental valuation of decentralized controllers for industrial robot manipulators

We present an experimental comparison among decentralized controllers for industrial manipulators. Three control laws (PID, Tarokh's nonlinear three term control, sliding mode) have been selected from those presented in the literature, which are suitable for the dynamic control of industrial manipulators since they do not need a system model and they require simple control architectures. Practical experimentations have been performed using a simple PC-based controller, and a 2-DOF system with a dynamic behaviour similar to the one of an industrial robot. The paper presents the problems which have encountered and the performances achieved in terms of tracking errors.