An Experimental Demonstration of Flocking with Multiple Autonomous Air Vehicles

Autonomous air vehicles have many exciting applications such as point-to-point passenger transport, distributed airborne sensing, and point-to-point mail delivery. All of these applications require safe and efficient use of the national airspace, which necessitates advanced coordination schemes for large groups of autonomous air vehicles. For example, it is neither desirable nor efficient for many small air vehicles to travel on independent trajectories because it is impractical to route and track a large number of vehicles safely. Instead, vehicles could travel together in a large group formation called a flock. Flocking behavior requires decentralized methods for coordinating and controlling large groups of autonomous vehicles. Safe and effective flocking methods will help to unlock many applications for multi-agent autonomous air vehicle systems. In this project, we aim to demonstrate autonomous flocking with a group of 3 autonomous rotorcraft. The Principal Investigator's research group recently developed a new multi-agent control method for coordinated flocking (including flock formation and flock breakup). This method requires real-time measurement of the air vehicles' relative positions and velocities. The Co-Principal Investigator's research group has developed vision-sensing methods, which are capable of providing relative position and velocity measurements in real time. These vision-sensing measurements will be used (in combination with GPS) to provide each air vehicle's guidance, navigation, and control system with the measurements needed to perform coordinated flocking maneuvers.