Robust Nonlinear Predictive Control Based on State Estimation for Robot Manipulator

In this paper a robust nonlinear model predictive controller based on state estimation for rigid n-link robot is designed. The control law is based on prediction model, which is carried out via Taylor series expansion. The optimal control is computed directly from the minimization of receding horizon cost function. There is no need to an online optimization. A disturbance observer is designed to deal with system uncertainties such as unknown external disturbances, unmodeled quantities and parametric uncertainties. Two methods are used to treat uncertainties estimation. The most known one is designed on basis of the theory of guaranteed stability of uncertain systems. Then, an estimation based on model predictive control law is carried out. Compared to the first method, this observer can eliminate the steady errors and enhances the robustness of the control scheme. This is due to the integral action occurred in the observer structure. The control law is carried out with state estimation through a nonlinear state observer. The global stability closed loop system (robot + controller + state observer) is proved analytically via Lyapunov stability theory. The performance of the control scheme is tested for the two link rigid robot.