Projects
2024: Testing of efficient AI models for cost reduction of drone navigation modules
Project duration
08.10.24 - 30.04.25
Project manager
Martin Rebane
There is a practical need for navigation without GPS both in military zones, for example for the delivery of medical or humanitarian aid, and increasingly in civilian use, for example considering the Kaliningrad GPS jammers that recently disrupted air traffic over the Baltic Sea. An alternative to GPS-based navigation is camera-based navigation for drones. Conventionally, the camera is used for relative positioning with respect to the original location. It is accurate within 100-200 meters, but deviates from the actual location for a 10-kilometer distance. A less explored alternative is absolute positioning, where a camera image is matched to an aerial or satellite photo. For this, however, the drone would have to fly significantly higher. The project is looking for a cost effective way to use AI to combine high-frequency and low-altitude relative positioning and low-frequency but high-altitude absolute positioning for drone mapping module manufacturing.
Share this entry
Related projects
