Swarm Rescue Challenge Final 2025–2026: save lives by controlling a swarm of drones

The challenge’s initial scenario

This challenge consisted of deploying a swarm of drones to explore an unknown area, difficult to access and potentially dangerous, with the mission of locating and guiding victims—who may be immobile and on the ground—to a safe zone. 

Through a simulation written in the Python programming language, each team was provided with a fleet of 10 drones, all equipped with:

• A communication system with limited range, which may be subject to jamming
• A laser rangefinder (LiDAR) enabling distance measurement from obstacles
• A semantic sensor capable of identifying the nature of an object without data processing
• A GPS that may be unavailable in certain areas

The jury and evaluation 

The jury, chaired by David Filliat (AMIAD), consisted of Emmanuel Battesti (ENSTA), Philippe Guermeur (DGA), and Éloïse Berthier (ENSTA, CIEDS).

The team rankings were based on the average of two scores:

• A technical score, assessing the performance of solutions in environments unknown to participants, used to establish a shortlist of the ten finalist teams
• A presentation score, delivered in person for the selected teams

Scores were awarded based on the performance of the solutions across three unknown maps, as well as the quality of the final presentation.

Final results 

David Filliat, Research Director at AMIAD, congratulates the thirty-seven teams that took part in this challenge, and especially the prize-winning teams.