Comparison of two adaptive control methods for robotic applications

Abstract

In this paper two adaptive methods, the self-tuning regulator and the adaptive tracking method, are compared in performance when applied to a two-degree-of-freedom robotic manipulator handling a load of unknown mass. The comparisons are based on digital simulations of the robot with load, the control algorithms and the adaptation algorithms. Both methods yielded stable system performance; however, the adaptive tracking algorithm showed faster convergence in adjusting to a sudden change in load mass. Further, implementation of this method did not require measurement of joint angle accelerations whereas the self-tuning regulator method did. In summary, the adaptive tracking algorithm appears to be superior for applications of this type. >