Design and Test of an Autonomous Drone SWARM for Chemical Agent Detection

Abstract

This technical report presents the details behind designing and testing a swarm of Unmanned Aerial Systems (UAS) that are capable of detecting and mapping a chemical plume. The sensor used onboard the UAS is a government-owned detector, although the technical approach used in this work is generalizable to other detectors of contaminants or chemical agents. The swarm of drones operates autonomously, with minimal user input but with operator oversight. Autonomous behaviors were built to search for and map a chemical plume over a given area. For UAS in search mode, a Lissajous pattern was chosen as the primary motion trajectory, owing to its surveying efficiency of minimizing time for distance traveled. If any of the agents within the swarm interact with a chemical plume (physical or simulated), the autonomy will cause the swarm behavior to adapt from the searching to mapping. In the mapping phase, each drone in the swarm performs circles around the chemical plume to map the perimeter of the plume. These circles are an important feature of the behavior because they allow the detectors to de-saturate from the chemicals. Finally, the team also developed and incorporated a visualization tool for real-time monitoring of chemical presence, to provide a Common Operating Picture for the user. Utilizing simulations, emulators, and real hardware, the team completed test flights of the swarm behavior to prove its functionality.