Formation Control With Multiplex Information Networks

Abstract

Current distributed control methods have a lack of information exchange infrastructure to enable spatially evolving multiagent formations. Specifically, these methods are designed based on information exchange rules represented by a network having a single layer, where they lead to multiagent formations with fixed, nonevolving spatial properties. For situations where capable agents have to control the resulting formation through these methods, they can often do so if such agents have global information exchange ability. Yet, global information exchange is not practical for cases that have large numbers of agents and low-bandwidth peer-to-peer communications. Motivated from this standpoint, the contribution of this paper is to show how information exchange rules, which are represented by a network having multiple layers (multiplex information networks), can be designed for enabling spatially evolving multiagent formations. In particular, we first consider the formation assignment problem and then the formation tracking problem and introduce new distributed control architectures that allow capable agents to spatially alter the size and the orientation of the resulting formation without requiring global information exchange ability. In addition, tools and methods from differential potential fields are further utilized in order to generalize the proposed distribute control architecture for the formation tracking problem to allow for connectivity maintenance and collision avoidance needed in real-world applications. Stability of the proposed architectures is theoretically analyzed and their efficacy is illustrated on numerical examples and on multiagent formation experiments.