Vision-based Distributed Coordination and Flocking of Multi-agent Systems

We propose a biologically inspired, distributed coordination scheme based on nearest-neighbor interactions for a set of mobile kinematic agents equipped with vision sensors. It is assumed that each agent is only capable of measuring the following three quantities relative to each of its nearest neighbors (as defined by a proximity graph): time-to-collision, a single optical flow vector and relative bearing. We prove that the proposed distributed control law results in alignment of headings and flocking, even when the topology of the proximity graph representing the interconnection changes with time. It is shown that when the proximity graph is ”jointly connected” over time, flocking and velocity alignment will occur. Furthermore, the distributed control law can be extended to the case where the agents follow a leader. Under similar connectivity assumptions, we prove that the headings converge to that of the leader. Simulations are presented to demonstrate the effectiveness of this approach.