One of the technical objectives of INGENIOUS is to develop an autonomous drone for unsupervised exploration and assessment of high-risk environments. This Multi-purpose Autonomous eXploring drone – MAX – navigates by itself in combined indoor and outdoor spaces and scouts the environment, acting as an “extra teammate” for first responders.

The concept

With its on-board suite of utility sensors, MAX provides visual and infrared images, 3D models, and environmental parameters such as temperature and indications of dangerous gases. This helps first responders assess the situation, inspect structural elements, locate victims, or understand how to approach a building with minimum risk.

To assist the first responders in the complex task of interpreting large amounts of sensor data, the MAX drone sends data to a Ground Control Station (developed by INGENIOUS partners ITC/Univ. of Twente) that runs machine learning-based algorithms for image enhancement, scene analysis and object recognition. The purpose is to raise alerts when MAX encounters elements that require human intervention, e.g. presence of victims, closed doors, important signs, damaged structures, etc.

MAX uses a dual-system SLAM approach (Simultaneous Localization and Mapping) to increase its robustness and ability to operate in complex environments; A short-range forward-looking active near-infrared stereo camera provides high-resolution depth images up to a few meters, while a laser scanner mounted on top gives 360 degrees coverage and detailed measurements at longer ranges.

Real-time 3D mapping and analysis functions identify how to move in the environment to explore it efficiently, directing MAX towards unexplored parts of the scene and assessing how the exploration progresses.

Example from a data collection in an indoor environment. Left: 3D point cloud acquired with MAX’s on-board laser scanner. The MAX icon and arrow show current position and orientation. Right: the view from a visual camera with the direction of movement shown by the red arrow.

Latest addition: Obstacle detection

The current development concerns obstacle avoidance functionality. At the lowest level, a set of ultra-sound sensors (sonars) act as an emergency brake that prevents MAX from colliding with objects not seen by the cameras and laser scanners and thus not taken into account in the path planning step. Among other things, this is an important safety measure for avoiding crashing into closed windows whose glass may sometimes seem transparent to optical sensors.

In addition to this, the laser scanner and the stereo camera detects small obstacles in the vicinity of the drone, and then raises an alert. The figure below shows two examples where such obstacles have been detected. As a future feature, a dynamic path-following technique will be implemented that tries to re-plan a path locally around such objects and find an alternative route to the next waypoint. This feature has been successfully tested on a ground vehicle and will be integrated onto MAX in the near future.

Alert signs mark the position of obstacles detected by MAX
MAX demonstration during the Scale Filed Test #5 (SST#5) in October 2021 at the Berrozi (Basque Country) facilities in Spain

The team behind MAX

The MAX drone is developed by FOI – the Swedish Defence Research Agency – a government authority under the Swedish Ministry of Defence and one of Europe’s leading research institutes in defence and security with around 1000 employees working in a broad field of applications. The team behind the MAX drone works at FOI’s Division of C4ISR in Linköping, Sweden, and combines expertise in unmanned aerial vehicles, electro-optical sensor technology and sensor informatics.

Share this post!