抄録
In this paper, we discuss to control of the number of clusters formed by distributed robots only with local information. This work was motivated by Swiss Robots, which collect scattered obstacles into some clusters without any global information nor intelligent concentrated controller. In this paper, we define fundamental event rules in this cellular world, and introduce two types of local rules for robot action: one is the Push & Turn rule, which can collect obstacles, the other is Pull & Turn rule, which can scatter obstacles. First, we show that the number of clusters can be estimated based on intrinsic memory of individual agents. Then, we propose a “mode switching strategy” that a robot is autonomously switching 4 modes: core-creating/growing/estimating, and fracturing mode. Repeated trials indicate that we can statistically control of the number of clusters by adjusting the switching threshold. And, we also show that timely switching between two rules in core-creating mode achieves to form a single cluster. At last, we describe that introducing anisotropic robots will lead to shift a clustering position.
