The Generation of Bidding Rules for Auction-based Robot Coordination

Robotics researchers have used auction-based coordination systems for robot teams because of their robustness and efficiency. However, there is no research into systematic methods for deriving appropriate bidding rules for given team objectives. In this paper, we propose the first such method and demonstrate it by deriving bidding rules for three possible team objectives of a multi-robot exploration task. We demonstrate experimentally that the resulting bidding rules indeed exhibit good performance for their respective team objectives and compare favorably to the optimal performance. Our research thus allows the designers of auction-based coordination systems to focus on developing appropriate team objectives, for which good bidding rules can then be derived automatically.