Basic Movement Behavior Research Of Swarm Robots

As swarm robots can take more complex task than single robot and multi-robots with swarm intelligence coming from its scality, behavior research about swarm robots becomes more and more popular. In this paper, how to aggregate swarm robots and realize theirs formation and walking is researched.Firstly, the coordination mechanism of swarm robots is researched. As a distributing system, robots inside swarm robots can not communicate directly one another to realize coordination. So an indirect communication method based on environment has been used in this paper. Then a existing coordination mechanism of swarm robots based on this indirect communication method has been analyzed. Through analyzing individual robot's environment model and relative movement model used in this coordination mechanism, two problems such as unreasonable global hypothesis and being lack of thinkness of maximum movement ability of individual robot have been found. To resolve these two problems, novel individual robot's general environment model and relative movement model are proposed.Secondly, how to auto-aggregate swarm robots has been researched. Through analyzing existing algorithm to auto-aggregate global swarm robots, it is re-proved that global swarm robots have strong auto-aggregate ability and perfect auto-aggregate performance when the individual robot's attraction/repulsion environment model and relative movement model are properly set. But for local swarm robots, it is proved that auto-aggregation can not be got easily by only utilizing auto-aggregation algorithm of global swarm robots. Through analyzing the reasons, a novel principal to design auto-aggregation algorithm of local swarm robots has been proposed. Then two algorithms based on this principal to realize auto-aggregation of local swarm robots have been designed. Under theoretical analyse and experiment simulation, the validity of these algorithms is proved. Comparatively, the second algorithm has better performance. Thirdly, how to passively aggregate swarm robots has been researched. While ugly environment makes auto-aggregation of swarm robots impossible, passive aggregation method has to be used. At First, new individual robot's attraction/repulsion environment model and relative movement model have been designed. Then, two methods such as leading method and aggregation-area-limitation method have been proposed. Under theoretical analyse and experiment simulation, the validity of these methods are proved. Comparatively, the second method has better performance.Fourthly, the formation of swarm robots is researched. Formation embodies the strong adaptive ability of swarm robots. Reference to formation algorithms used in multi-robots and based on particular coordination mechanism of swarm robots, some methods to realize formation of swarm robots have been proposed. For global swarm robots, center-reference method and constraint box method have been proposed. Under theoretical analyse and experiment simulation, the validity of these methods are proved. Comparatively, the second method has better performance. For local swarm robots, a normal passive formation method based on its environment has been proposed and it is proved that this method can randomly change swarm robots' form based on its environment.Finally, the walking of swarm robots is researched. Walking is a necessary ability of swarm robots. For global swarm robots, the walking mechanism has been researched and virtual introducing method has been proposed. Under theoretical analyse and experiment simulation, the validity of this method is proved. For local swarm robots, the neighbor-following method and real introducing method have been proposed. Theoretical analyse and simulation result prove the validity of these methods. At last, the pheromone has been proposed. With character of being easy to be used and perfect performance, this method can be utilized widely in walking of global and local swarm robots.