Active Modeling and Control for Manipulators

In this paper, an active model-based control scheme is proposed for the manipulators with parameter uncertainties and disturbances. Firstly the Euler-Lagrange dynamic model of n-link manipulator is linearized by Lie derivation. Then Kalman Filter is utilized to estimate the model error in real-time. Moreover, an active model enhanced state feedback controller is proposed based on the estimated model error. Finally, several simulations are conducted on a manipulator to validate the feasibility of this method. With the proposed control scheme, the tracking errors of all joints are reduced by more than 47% comparing with the conventional state feedback controller.