The paper “Detection and Ranging of Outdoor Fires for Autonomous Firefighting Drones” has been accepted for presentation at the 2025 IEEE International Symposium on Safety, Security and Rescue Robotics (SSRR), to be held in Galway, Ireland.
The work is the result of the Master’s thesis of Mohamed Khaled Othman, conducted at the Smart Mechatronics and Robotics (SMART) Research Group at Saxion University of Applied Sciences, under the supervision of ir. Benjamin van Manen and Prof. Matko Orsag from the University of Zagreb – Faculty of Electrical Engineering and Computing (FER), in collaboration with Impulse Fire Fighting Solutions (IFFS Europe), a company developing handheld and aerial impulse-based fire suppression systems.
Developed within the framework of the Erasmus Mundus Master in Intelligent Field Robotic Systems (IFRoS), the research introduces a Fire Detection and Ranging (FDAR) pipeline that enables drones to autonomously detect, localize, and assess outdoor fires while ensuring human safety during suppression.
At its core, the proposed pipeline combines state-of-the-art deep learning fire detection, RGB–thermal (RGB-T) image fusion, and monocular depth estimation to identify both fire and human presence in real time. The work benchmarks several YOLO and transformer-based architectures, showing that the latest YOLOv12 model offers the best trade-off between accuracy and speed for onboard drone operation. Moreover, Mohamed’s novel RGT fusion approach — where thermal data replaces the blue channel in RGB imagery — demonstrated the most robust results, reducing both false positives from heat sources and missed detections in smoky or low-visibility conditions.
Beyond detection, the system incorporates a safety-aware targeting logic that prevents the drone from activating its water cannon when people are within the line of fire. The integrated visualization interface further allows human operators to monitor detections and depth estimations in real time.
This research marks a major step toward autonomous aerial firefighting, building the perception backbone for future fire response drones. The next stages aim to deploy the pipeline on embedded hardware and integrate it with the drone’s control and suppression systems for full autonomy.
The project was developed in collaboration with Impulse Fire Fighting Solutions (IFFS Europe) (https://iffs-europe.com), whose drone prototypes can already perform multiple impulse water shots mid-flight — with Mohamed’s perception pipeline paving the way for their next step toward autonomy.
Mohamed completed his studies in the IFRoS Master’s program and is currently a doctoral researcher within the MSCA Doctoral Network(AERIALIST) — a European research initiative on assistive health technology in unsupervised and home settings, focusing on safe human–robot interaction, sensorimotor AI, and embodied intelligence. He is based at the Institute of Mechatronic Systems (IMES) at Leibniz University Hannover (LUH), where he works on biomimetic, model-based reinforcement learning of motor primitives.
His journey reflects the IFRoS program’s commitment to fostering innovation-driven researchers who bridge advanced robotics research and real-world impact through international collaboration.