Drones with Omni-Event Vision for Drone Neutralization – DEVIN

The project relies on a multidisciplinary approach structured around several complementary methodological axes. Initially, we aim to develop a 360° spherical sensor based on event-based cameras. This novel device will require a thorough calibration and alignment phase to ensure precise measurements in dynamic three-dimensional environments. Subsequently, a dedicated software suite will be developed to enable localization, tracking of moving targets, and inter-drone communication, the latter being based on the exploitation of luminous events captured by the sensor. In parallel, particular attention will be given to the control of the drone fleet, whose mission is to intercept and pursue a hostile drone. To this end, several control strategies will be investigated: model-based approaches, data-driven methods, and hybrid techniques combining both paradigms. Finally, all developments will be validated through rigorous experimental campaigns, carried out under predefined scenarios representative of real-world use cases.

At this stage of the project, several significant milestones have been reached. A first functional version of the 360° spherical sensor, based on event-based cameras, has been designed and integrated. In parallel, initial algorithms for object tracking and inter-drone communication—relying on the detection of luminous signals—have been developed and successfully tested under controlled experimental conditions. On the control side, a preliminary strategy based solely on physical modeling has been implemented and validated through practical trials, demonstrating both the feasibility and robustness of the approach in representative scenarios.

The project is progressing as planned, in line with the initial timeline. Ongoing developments will be continued in accordance with the objectives defined in the original project proposal, with a focus on algorithm refinement, system integration, and large-scale experimental validation.

[1] Mixed guidance law for capturing a reactive target by coordinated Multi-UAV, Felipe Kataoka Ishikawa, Sarah Aouiche, Bojan Mavkov, Guillaume Allibert, in International Conference on Control, Automation, Robotics and Vision, Dubai, December 2024 (https://hal.science/hal-04742967v1)

[2] A New Stereo Fisheye Event Camera for Fast Drone Detection and Tracking, Daniel Rodrigues Da Costa, Maxime Robic, Pascal Vasseur, Fabio Morbidi, in IEEE International Conference on Robotics and Automation, Atlanta (USA), May 2025

Over the past 15 years, the popularity and accessibility of unmanned aerial vehicles has seriously increased. The number of incidents involving drones flying over or landing on critical infrastructures around the world (the White House, the Japanese Prime Minister's office, the Golden Gate Bridge, nuclear sites in France, prison facilities, etc.) is increasing.
Although regulations are strict and require a number of rules depending on the class of the vehicle, these devices can be easily disconnected on commercial aircrafts and can be omitted on clandestinely manufactured aircrafts.
In the DEVIN project, we are interested in the detection, localization, tracking and encirclement of an enemy drone by a UAVs swarm.
In order to address the scientific challenges identified and stated above, we assume that the swarm evolves in an unknown and potentially dynamic environment. In addition to their Inertial Measurement Unit, they will be equipped with a spherical event camera. The latter will allow each drone in the swarm to detect and locate a potential enemy and to ensure its autonomous pursuit. It will also serve as a means of communication between the drones in the swarm using single-pattern light flashes. The control of the swarm will be ensured by decentralized controllers based on reinforcement learning easily transferable to the real world and having stability properties.
This project will allow to obtain an easily transportable solution by using as a means of detection, localization and tracking, a swarm of UAVs each embarked with mobile and small size sensors. DEVIN is positioned as an indispensable complementary approach in the short term to the solutions developed to date, whose main drawback is that they cannot ensure the security of mobile sites (demonstrations, processions, etc.).