Novel physicomimetic bio-inspired algorithm for search and rescue applications

Decentralized algorithms are often used for multi-agent search in scenarios in which it is difficult or computationally intractable to use a centralized control strategy. “Bottom-up” multi-agent search techniques, or algorithms that are based on local interactions between agents and their environment, allow for highly scalable and fault tolerant systems. Moreover, they often demonstrate remarkable emergent properties, such as foraging behavior and emulated states of matter, which can be found in bio-inspired swarm algorithms and physics-inspired methods. However, it is often difficult to control these systems, especially when area coverage and target localization are the primary goals. In this paper we propose using stigmergic techniques combined with physicomimetic force laws to guide a multi-agent system toward efficiently exploring a pre-defined search area. We found that our method provides a reliable and effective method for decentralized area coverage by a multi-agent system. Furthermore, we implemented our approach using Robot Operating System (ROS) and the Gazebo simulation environment.