On the exact parameter estimation of robot manipulators with a predefined minimal amount of excitation

Abstract

Abstract Adaptive control of robot manipulators has been widely used in the last decades. The main control goal is usually trajectory tracking, but exact parameter estimation is always desirable. In the recent years, the so called dynamic regressor extension and mixing (DREM) procedure was developed to provide an alternative in the design of adaptive laws with more relaxed parameter convergence conditions than the Persistency of Excitation (PE). This article proposes an innovative DREM based adaptive scheme with the ability of increasing a minimal amount of excitation to the level necessary to obtain exact parameter estimation in finite time. Furthermore, a new robust control law is proposed to guarantee trajectory tracking even if exact parameter estimation does not take place. Simulation results are in good accordance with the developed theory.