A market-based approach to tightly coordinating multirobot teams
Pacific Forum - 3:00 PM
A snapshot from our simulator that highlights three rovers exploring an environment while maintaining line of sight communication contact with each other. Here, the right-most robot will have to closely coordinate with its teammates to visit its next point of interest without violating the team's LOS constraints.
Nevertheless, these conditions often require that robots closely coordinate to ensure mission success. Consider a team of rovers exploring a hazardous environment. For safety and monitoring, we constrain teammates to maintain an ad-hoc network to complete their mission.
Two of our team of three e-Gator robotic platforms
In this talk, I explore the coordination challenges posed by these domains that require solving a distributed, tightly-coupled multirobot planning problem. Uncertainty in the environment and the team also necessitates persistent tight coordination between teammates throughout execution. Additionally, I present our framework, Hoplites, that borrows ideas from economics to achieve coordination. In Hoplites, the team simulates a market-economy in which robots negotiate tightly-coupled plans that efficiently meet both the mission constraints and objectives. I also highlight simulation and field experiments involving two different problem domains.