Virtual Decomposition Based Adaptive Control for Robot Manipulators: Theory and Experiments

Abstract

Virtual decomposition is a modular approach in which only the dynamics of the subsystems instead of the dynamics of the complete system are used in control design. The modular structure allows a variety of control objectives (position control, force control, constraint control, joint flexibility, and optimization) for a variety of robotic systems (single and multiple robot arms, space robots, walking robots, and human robots) without limitation on robot type. In this paper, the proposed approach is applied to adaptive control of an industrial robot manipulator. Both theoretical and experimental results are presented. In the theoretical aspect, modular control design together with modular stability analysis is presented. Lyapunov asymptotic stability is ensured. In the experimental aspect, preliminary results achieved on the industrial robot KUKA 361 demonstrate the advantage of the proposed approach compared to the original KUKA controller, particularly when the robot moves at a high speed.