Adaptive Control of Robot Manipulators: Design and Robustness

In this paper we present a globally convergent adaptive control algorithm for n-link rigid robots. The algorithm can most easily be described as a combination of the passivity-based approach[14] and the guaranteed stability approach[4]. The algorithm relies on the passivity and the linearity-in-the-parameters properties of rigid robots, as in passivity-based parameter adaptive algorithms; however, the number of parameter update laws required in the adaptation is independent of the number of links of the robot. In the simplest case, presented here, two update laws are required. This scheme thus has certain advantages both with respect to design and with respect to robustness. With respect to design, the difficulty of tuning numerous adaptation gains is avoided. With respect to robustness, since the joint tracking error vector always has codimension two, the transient response is easier to constrain and problems of parameter drift are more easily handled.