Adaptive Robotic Control in the Presence of Uncertainties

This paper presents new design procedures for adaptive robotic control with uncertain plant models. We consider cases where the uncertain terms of the dynamic model are completely not available. The uncertain plant terms are not assumed to be "slow varying" or bounded by a constant bound. Instead, they are bounded by a first order polynomial of the system states. This assumption allows the proposed adaptive controller to be applied to fast robot motions and varying external loads. A stability analysis proves that developed adaptive laws provide bounded tracking errors, and the error bounds can be made arbitrarily small by a proper choice of the controller parameters. The paper also includes an analytical study on the effects of adaptation gains and measurement noise. A series of simulation results is included to demonstrate the performance of the new adaptive controllers.