COOPERATIVE ACCELERATION OF TASK PERFORMANCE - FORAGING BEHAVIOR OF INTERACTING MULTI-ROBOTS SYSTEM

摘要

We study the effectiveness of cooperative behavior in a society of interacting agents. After reviewing the problem and defining the concept of swarm intelligence, we examine collective behavior of many-body active clusters through a task to gather pucks in the field. In this study, we used a robot with a simple structure which has a driving system and the simplest interacting means; a light and some sensors. The effectiveness of group behavior was studied under various (homogeneous, localized) puck distributions with real experiment, simulation, and analysis. To evaluate the efficiency of group behavior, we examined the scaling relation between the task completion time and the number of robots, and the relation between the interaction period and the efficiency of group. We found that a cooperation between agents by a simple interaction is very efficient in enhancing the performance of the group compared with independent individuals. [References: 17]