On dynamic distributed control and its application to multi-robot systems

Abstract

This paper contributes to addressing the rendezvous problem of multi-robot systems. A dynamic distributed controller that adopts the multi-sensor information fusion is designed for the position control of robots. It is shown that if the communication digraph has a spanning tree, then the positions of all robots reach rendezvous. Moreover, two algorithms are proposed by applying the dynamic distributed controller although the robots are subject to nonholonomic model dynamics. The first one adopts a certain transformation to convert the kinematic model of the mobile robot to an integrator model, based on which the dynamic distributed controller can be directly applied. The other one develops a two-layer control strategy. The dynamic distributed controller is applied in the position control layer, and a desired angle of the direction control can be obtained simultaneously. A tracking algorithm is designed in the direction control layer for each single robot, which is shown with the ability to track the desired angle. Simulation tests are conducted to validate the effectiveness of the proposed algorithms.