| Search and Rescue (SAR) missions are crucial for the recovery of humans after a disaster. Traditional SAR missions often deploy drones on pre-planned paths or rely on being manually controlled by humans. The primary focus of this research is on optimising the routing of drones and establishing intelligent self-coordination mechanisms to accomplish SAR successfully and accurately. The proposed approach utilises Swarm Intelligence (SI), a subset of Artificial Intelligence (AI) that draws inspiration from the decentralised and self-organising behaviour observed naturally in populations of social animals, for the coordination of the swarm of drones. Drones are equipped with object detection models like YOLO to detect the probability of humans, which is fed to the proposed routing algorithm, HyDRA-PSOGWO, to coordinate and self-organise positions in the swarm. The PSO offers fast exploitation, while GWO, on the other hand, maintains exploration, avoiding premature stagnation. Using Gazebo modelling and ArduPilot in ROS, this was simulated in a 3D environment. The proposed algorithm has achieved an average success rate of 80%. Experimental results demonstrate better performance over traditional metaheuristic algorithms (PSO, DE, GWO), achieving 100% target detection with small drone swarm deployment. |
*** Title, author list and abstract as submitted during Camera-Ready version delivery. Small changes that may have occurred during processing by Springer may not appear in this window.