Low-computation dynamic generation of distributed connectivity control for varying sized multi-agent systems

摘要

Dynamic communication topologies are an important yet challenging topic for autonomous communication. This paper proposes the use of a low-computational algorithm to generate and maintain connectivity with multi-agent systems wherein the number of agents can change over time. This methodology utilizes communication links described by a price criteria that is decided by what sensors are available to each agent to calculate the communication quality between each potential link. By using this pricing methodology the communication graph is generated in a distributed manner that allows for agents to leave and enter the communication flock. This method does not require complex calculation and can be used by each individual agent to create the communication graph with the fastest communication speed given the price of each communication link. This type of methodology is very promising for many applications where multi-agent systems need a communication protocol in a distributed way. Such applications can be seen in cooperative-UAS, ground based robotics, multi-agent tracking, and heterogeneous systems.