Multi-agent Formation Control based on Distributed Estimation with Prescribed Performance

We consider the distributed simultaneous estimation and formation control problem for swarms of identical mobile agents with limited communication, sensing and computation capabilities. In particular, we develop a novel scalable algorithm that encodes the formation specifications of the swarm via geometric moment statistics, which are estimated by a distributed scheme with prescribed performance guarantees. Based on the locally available information, each agent calculates an estimate of the global formation statistics, which is then employed by its local motion controller, thus creating a feedback interconnection between the estimator and the controller. The proposed scheme guarantees convergence of the global formation statistics to the desired values, while decoupling the estimation performance from the control performance. Moreover, a minimum allowable inter-agent distance can be predetermined so that inter-agent collision avoidance is achieved. Finally, simulation paradigms are provided to validate the approach.