Study Of Pattern Formation By Swarm Robots 2 Coordination

Pattern formation, as the classic problem to be solved in the research of swarm robots, could supply a good scene to swarm robots coordination, communication and learning. The study of swarm robots coordination is an important research area in the research of swarm robots, which is the base of swarm robots completing a task and decides the result of the task completion. Through quantities of materials filing and investigation, we can come to that at present the popular coordination algorithms make swarm robots completing tasks well. But they may still not come to the optimal way of swarm robots.In this research, there are swarm robots of high homogeneous and self-organized, with no leader. It applied a market-based Distributed Bidirectional Auction algorithm to swarm robots to achieve pattern formation. This innovated coordination algorithm bases on classic auction algorithm, and sets a novel way of making dual roles of auctioneer and bidder to a single robot. We propose a new auction mechanism of normal auction, reverse auction and iterative auctions composing the whole auction, which makes swarm robots system have the advantages of completely distributed algorithm, and the optimal result for robots is worked out. For homogeneous and simple characteristics of swarm robots, in this paper we simulated robots functions such as simple move, sensing, and communication, managed to design the system simulator to swarm robots, and carried out plenty of experiments of pattern formation on the simulator by Bidirectional Distributed Auction algorithm and made comparisons with other algorithms.Through smoothly analyzing and evaluating the Distributed Bidirectional Auction algorithm, it can come to these conclusions: 1) Distributed Bidirectional Auction algorithm on swarm robots is a novel and successful market-based swarm robots coordination algorithm. For its characteristics of completely distributed, it has the advantages of good scalability, robustness, high self-organization, and parallel running. 2) To complete the same scale of pattern formation, with the number of robots growing, Distributed Bidirectional Auction algorithm made swarm robots consume the total energy decreased obviously, which is better than other algorithms of same type.3) To complete the different scales of pattern formation, with the subtasks number growing, the robots number growing, the trend of total energy consumption of swarm robots by Distributed Bidirectional Auction algorithm increasing gets lower, and the trend of getting lower is greater; meanwhile, the average energy consumption decreased by steady trend. The result is better than other ones by other algorithms of same type, and with robots number growing, this advantage is more obvious.