Adaptive tracking control for cooperative multiple robot systems

Abstract

This study presents a novel adaptive controller for cooperative multiple robot systems to obtain both zero motion and force tracking errors. First, by introducing the concepts of force decomposition and virtual control input, we reformulate the control problem of cooperative manipulators as a problem of controlling a single constrained manipulator. Next, a stable subspace is defined by both motion and internal force errors. This subspace includes both zero motion and force errors as its largest invariant set. The controller is then designed to drive the error dynamics into the stable subspace. A non‐adaptive and an adaptive approach both systematically show the asymptotic motion/internal force tracking. Note here that the zero force tracking error is achieved without needing the persistent excitation (PE) condition in the adaptive control scheme. Furthermore, based on a decentralized method, the proposed controllers are implemented in multiprocessor systems to balance the computation load. Finally, two cooperative planar robots are used to demonstrate the developed methodology where the simulation results illustrate the expected performance.