Scalable Path and Time Coordination for Robot Formation

In this paper, we consider several CPS challenges (e.g., responsiveness, scalability, adaptability) in multi-robot formation. In general, the response time of multi-robot formation task involves two parts: the computation time for path and time coordination to avoid any collision among robots and the actuation time for the control of the robots to actually move to their destinations. In terms of responsiveness, a shorter response time provides a higher quality of responsiveness. However, it is complicated to reduce the response time since reducing computation time and reducing robot actuation time are conflicting objectives, and such a trade-off varies over environment. We present a scalable optimization framework that explores such a trade-off dynamically and exploits it in a feedback manner to find efficient trajectory schedules. Our simulation results show that our framework successfully finds a shorter response time by adapting to various environments compared to a commercial optimization tool, and it is scalable for a large number of robots.