Robust Nonlinear Model Predictive Control for Robot Manipulators with Disturbances

Abstract

T0641. The robust NMPC algorithm is developed based on tube methodology. First, by making use of the upper-bound of the disturbances, the tube cross-sections are calculated to contain all possible states of the actual systems, which are centred on a nominal trajectory optimized online. We then use the tube cross-sections to tighten the constraints on the nominal predictions and design a tailored terminal region and a corresponding terminal controller. By implementing the proposed algorithm, robust constraint satisfaction and recursive feasibility are ensured even in the presence of disturbances. In addition, the perturbed closed-loop system is input-to-state stable (ISS) and converges asymptotically to a neighborhood of the reference trajectory.