A nonlinear robust controller for robot manipulators

An improved accuracy variable structure robust controller for robot manipulators is discussed. In comparison with the well-known robust controllers utilizing saturation functions with a linear zone in a close vicinity to the desired trajectory, the proposed controller exploits properly selected nonlinear functions instead of linear functions. Thus, a smooth control law and lower steady-state tracking error are achieved. Global stability of the proposed controller is proved for a class of curves including third-order splines, Bernstein-Bezier curves, etc. The controller was applied to a second-order single-input-single-output (SISO) dynamic system, as well as to a two-joint SCARA-type robot. >