Stabilizing NMPC of wheeled mobile robots using open-source real-time software

Abstract

In this paper, a recently developed open-source toolkit implementing fast nonlinear model predictive control (NMPC) routines has been utilized to achieve the two main control objectives of nonholonomic mobile robots, namely, point stabilization and trajectory tracking. The stability of the controller has been guaranteed by adding final state equality constraint, which generally requires, for feasibility, long optimization horizon and hence is computationally demanding. In order to use the toolkit for real experiments, a C++ code, which couples the toolkit and a wheeled mobile robot research platform's software, has been developed. Full scale experiments have been conducted showing the applicability of the stabilizing terminal equality constraint NMPC, to wheeled mobile robots' control problems, in a real-time framework.