Formation stabilization of nonholonomic robots using nonlinear model predictive control

Abstract

This paper compares two approaches of multi-robots' formation stabilization using nonlinear model predictive control (NMPC), namely, centralized and distributed predictive controls. Centralized NMPC has been highlighted in the literature to lead superior performance; however, it has been remarked with high computational power requirements which limit its application to practical formation stabilization problems. Nonetheless, in this paper, the use of a recently developed toolkit implementing fast NMPC algorithms rendered this problem tractable. The performance of the two control approaches are compared in a series of numerical simulations. The results demonstrated that the centralized controller has a better performance compared with its distributed counterpart. Furthermore, it showed real-time requirements satisfaction.