Who We Are

The Project

Argus Panoptes is a multi-drone autonomous search and rescue research platform. The system is designed to deploy a coordinated swarm of UAVs into unknown, GPS-denied environments and locate survivors using onboard computer vision, real-time mapping, and decentralised task planning.

The platform is built on ROS 2 and simulated in Gazebo before real-world hardware trials. Every design decision prioritises reliability under degraded conditions over performance in ideal ones.

Approach

Rather than adapting commercial drone software to SAR tasks, Argus Panoptes is built from first principles. The perception stack uses RGB cameras and trained neural networks to detect human presence from the top. Future improvements focus on thermal imagery as well.

Swarm coordination will be handled by using ROS2's mesh network capabilites, such a Zenoh. This open-source communication protocol and set of libraries is emerging as a leading middleware for ROS2 in mesh network scenarios, particularly for decentralized multi-robot systems (MRS) in challenging environments like planetary exploration.

"Current SAR drone systems are either human-piloted and resource-intensive, or autonomous only in controlled airspace. The gap between the lab demo and the mountain at 2am is enormous. That's the gap we're closing."

The Team

Core research team

Inigo Chandia
Project Lead / Robotics / Perception
System architecture, ROS 2 integration, simulation environment, and swarm coordination, computer vision pipeline, depth estimation, and target classification.
Spencer Sedano
Hardware / Robotics / Flight Systems
PX4 configuration, controller tuning, trajectory planning, and hardware-in-the-loop testing, path planning algorithms, map-building, GPS-denied localisation, and mission logic.

CORE STACK

ROS 2 Humble Gazebo PX4 Python C++ Computer Vision Docker