Digital twin technology for wildfire monitoring using UAV swarms

Abstract

Deploying aerial swarm robotic systems in real-life scenarios can be challenging. Using them to monitor wildfires requires the system to be easily used by a swarm operator. To achieve this with the minimum associated costs, advanced frameworks must be developed to optimise, monitor, and control the swarm in real time. One approach to achieve this is the creation of a digital twin where physical counterparts can be mirrored in a virtual world. Our aim was to create a digital twin to support and accelerate the design, testing and deployment of aerial swarms. Our framework is composed of the following main parts: a digital twin system for development and optimisation of swarm algorithms as well as real-time monitoring and control of swarm deployments; a cloud infrastructure to allow data passing between our system components; and a swarm of uncrewed aerial vehicles (UAVs). We developed a user interface that allows a swarm operator to define missions for a swarm to monitor areas in search for a digital wildfire. We tested our system in field trials using a mix of three physical and three digital aircraft. During the trials an operator was able to task the UAVs to perform autonomous search amongst three different search strategies, to stack at specific locations, and finally to land.