Dynamic control of multi-robot formation

The issue of multi-robot formation is address in this paper. Usually robots can only handle local information described by an adjacency graph. We define structure error of formation based on the graph theory and propose a local control strategy for individual robot based on NN control with robust term which can ensure robots converge to a unique formation. Furthermore we discuss the performance on two aspects: the first one is dynamic characteristics during formation process; the second one is obstacle avoidance during formation.