External Interference on Autonomous Robotic Swarm

Daniel E. Villarreal*, Di’Quan M. Ishmon*, Trevor R. Smith*, R. Michael Butts*, Xi Yu, and Yu Gu
Department of Mechanical & Aerospace Engineering, West Virginia University, Morgantown, WV 26506-6045

Presentation No.: 69

Assigned Category (Presentation Format): Engineering (Poster Presentations)

Student’s Major: Aerospace Engineering

This study investigated a situation where a semi-autonomous adversary robot directly interacted with an autonomous robotic swarm to infiltrate a defensive waypoint. This was intended to mimic a real-life situation where robotic swarms are used in foraging tasks and security is a concern. In the simulation we created, a state machine was implemented to enable the swarm to complete foraging tasks assigned by the operator, and take up defensive formations towards the invader and/or the defensive waypoint. Once the adversary was detected by an agent, it utilized fluid flow kinematics and set a sink flow to the defensive waypoint. This allowed individual agents to decide whether to protect the defensive waypoint (sink) or pursue the invader (doublet). The Doublet formation enabled the swarm to funnel the invader towards one direction and continue wherever the invader struck. We simulated this for a visual representation of the state machine, the operator’s way of assigning tasks, and to observe the emergent behavior that resulted during defensive formations that kept the invader from reaching the defensive waypoint.

Funding: NSF, Award # 1851815

Program/mechanism supporting research/creative efforts: the WVU Robotics REU (Yu Gu & Jason Gross)