Guidance and Navigation Controls for Drone Swarm Applications

This paper describes the research utilizing Quanser equipment to create GNC systems to maneuver a swarm comprised of QDrones, QBots, and QCars in different formations and specified movements [1]. Drones have become commonplace in many industries, taking place of human agents doing tedious or dangerous tasks such as infrastructure inspections, deliveries, or search and rescue missions. While Drones are utilized in many industries, research and innovative projects have been focused on swarm applications. Swarm applications refer to the use of a group or ‘swarm’ of agents which can be drones, ground vehicles, or other semi-autonomous equipment which can send and receive data or commands. This research has been applied to various issues across many fields, but mainly in industries which deal with data collection and exchange, where the swarm’s efficiency can be most readily observed. There are many challenges involved with swarm applications as controlling many agents can lead to issues with safety, legislation, and computational limitations specifically as it pertain to drone agents. In this study, a method was developed to control multiple drones and unmanned ground vehicles synchronously. The primary challenge to overcome was the creation of a control system that can command each agent individually, this will require an individual controller alongside a central server which collects data and calculates the desired maneuvers for each agent in real-time. The project yielded a variety of successful maneuvers that could be applied to indoor environments and eventually to outdoor platforms. Each maneuver had a different control system that took advantage of many different control types such as state flows, handheld controls, and many more. Future research will include developing other large swarm maneuvers that incorporate more agents, eventually allowing the swarm to gather data to train Graphing Neural Networks (GNNs) to the point where GNNs will plan and control swarm agents with minimal human input.