Adaptive global tracking for robots with unknown link and actuator dynamics

We present a continuous adaptive controller for the robot adaptive tracking problem with unknown robot parameters, unknown actuator parameters, and unavailable joint acceleration measurement. We prove that the proposed controller has global convergence to zero tracking error for any non-zero gains. This is a distinction compared to existing approaches, so far as applied to robots, such as the adaptive controller of Bridges et al. (Conference on Decision and Control, 1993; 159–165) whose global-convergence proof depended upon the controller gains being larger than certain bounding functions containing bounds on the system parameter estimates. The present work solves the same problem in a different way, where the main difference is that the global convergence does not impose a requirement of a gain size proportionate to the parameters' uncertainty sizes.