Adaptive formation control of dynamic nonholonomic systems using potential functions

In this paper an adaptive formation control for nonholonomic mobile robots with unknown dynamic parameters is proposed. The control scheme is based on a saturated artificial potential function which allows a decentralized formation control design including collision avoidance. For each agent, a cascaded control strategy combines an adaptive motion controller which takes into account its uncertain nonlinear dynamic, and a kinematic control, leading to an overall globally stable adaptive system. The proposed control strategy is a velocity free scheme, where the control law only depends on the neighbors position information.