Detection and mapping of a toxic cloud using UAVs and emergent techniques

Unmanned aerial vehicles have gained a lot of interest in recent times, due to their potential use in several civil applications. This paper focuses on the use of an autonomous swarm of drones to detect and map a toxic cloud. A possible real-world scenario is the accidental release of hazardous gases into the air, resulting from fire or an explosion at an industrial site. The proposed method is based on the concept of swarm intelligence: each drone (agent) performs basic interactions with the environment and with other drones, without need for a centralized coordination technique. More precisely, the method combines collision avoidance, flocking, stigmergy-based communication, and a cloud exploration behavior called inside-outside. For the experiments we developed a simulator using the NetLogo environment, and tested different combinations of these emergent behaviors on two scenarios. Parameters were tuned using differential evolution and separate scenarios. Results show that the combined use of different emergent techniques is beneficial, as the proposed method outperformed random flight as well as an exhaustive search throughout the explored area. In addition, results show little variance considering two different cloud shapes.