Behavior sequence analysis of incrementally evolving robotic swarms in a foraging task

Swarm robotic systems typically comprise many homogeneous robots that operate autonomously without a global controller. Swarm robotics aims to establish desired collective behaviors through multiple interactions with other robots or between robots and their environment. Robot learning or artificial evolution leads to self-organization, which underlies the emergent behavior that ultimately governs the robotic swarms. However, we are unaware of any method that accurately mimics true macroscopic collective behavior. In this study, we propose a novel method for analyzing the collective behavior of robotic swarms. In particular, we adopt the behavioral sequence concept, which stems from ethology. Analysis of behavioral sequence reveals that robotic actions evolve into specialized roles, providing insight into the role of subgroups in the robotic swarm. Applying this method, we observe collective behavior in a foraging task of autonomous mobile robots that evolve in an incremental manner.