A networked formation control for groups of mobile robots using mixed integer programming

Abstract

In this paper we demonstrate an effective way of description and control for network-controlled formations of mobile robots using minimized inter-robot communication. Therefore we use a global centralized planning algorithm and solve the optimal trajectory problem of the formation of mobile robots by mixed integer quadratic programming (MIQP). This description of formation can be used for non-static formations as well as for formation switching between different kinds of formations where inter-robot collisions will be avoided and the formations are organized as unit-center-referenced formations. The environment itself can be formulated by convex polygons that are described as hybrid systems. The moving along the optimum trajectories is controlled by a controller structure in three levels, whereby the highest level is set on a master robot and controls the general formation using Bluetooth communication between the robots. The control of the position of each robot is realized by an individual controller on every robot and a motor controller for every wheel of each robot. The effectiveness of our formation control structure and the algorithm for the planning of the trajectory is demonstrated in simulations and experiments which also verify the dynamic models of the robots