Decentralized control for uncertain robots using only position feedback

A model following decentralized robust controller for uncertain robots is proposed. The robust controller is designed in two steps: firstly, an artificial nominal plant is introduced and based on a reference plant a nominal controller is designed, which enables the nominal closed-loop system to achieve desired tracking properties. Secondly, a robust compensator is added to restrain the influence of uncertainties and nonlinear couplings. By applying the controller with a sufficiently wide frequency bandwidth, local uniform ultimate boundedness is achieved. Only using local position feedbacks, the controller has the advantage of tuning on-line and easy realization. Analytical proofs and simulation results validate the effectiveness of the method.