Abstract
The model-based adaptive control of robot manipulators has been studied actively in recent years (e. g., Slotine and Li (1987), Sadegh and Horowitz (1987), Koditschek (1987), and Berghuies et al. (1992)). In these adaptive controllers, global asymptotic stability is guaranteed. But, it is difficult to apply them to manipulators with more than four degrees of freedom because of its computational complexity.
This paper presents an efficient computational algorithm of model-based adaptive control for manipulators. It adopts the basic method of Slotine and Li (1987) and uses the minimum parameter set of manipulator dynamic model, which was derived by Kawasaki and Kanzaki (1991). The number of operations in the algorithm is about one half of that presented by Niemeyer and Slotine (1988). This algorithm can be easily applied to the other model-based adaptive controllers with few modifications. We also present experimental results on the adaptive control of a robot manipulator with six degrees of freedom. Its control system has a 32 bit DSP for the computation of the adaptive control. The sampling time is 1.65ms.
