Model-Error-Observer-Based Control of Robotic Manipulator with Uncertain Dynamics

Abstract

In this paper, a nonlinear controller for robotic manipulator with unknown model parameters is proposed to reach high accurate trajectory tracking. The unknown parameters such as uncertain moment of inertia, uncertain geometry of manipulator, unknown friction torque, unknown gravitational torque and payload variation are addressed. Model-based control methods require accurate model parameters, while it is difficult to get these parameters. To solve this problem, a model-error observer is proposed, and it observes the parameter error effectively. In the proposed control law, the model-error observer is adopted to handle unknown model parameters, and this controller solves the problem of model-based control methods effectively. The robust performance of the control law is confirmed in simulations, and the results show accurate path tracking in spite of the existing friction and unknown model parameters.