AI-Powered Drone Beats Former World Champions Without Satellite Navigation

2 Min Read

Researchers at the Delft University of Technology in the Netherlands have built an autonomous racing drone that uses just a single camera and a motion sensor to fly complex courses at high speed. In other words, the system does not rely on satellite navigation, which is usually required for such flights.

During tests, the drone reached speeds of up to 100 kilometers per hour and beat three former world champion pilots in a race.

According to the researchers, the system could help create faster and more energy-efficient autonomous robots.

The result is interesting not only from a racing perspective. The drone navigated complex courses without satellite navigation, meaning it made all flight decisions on its own based on data from the camera and motion sensor. This directly affects areas where satellite signals are unreliable: indoor spaces, dense structures, and possibly challenging environments as well.

The victory over three former champions shows that a system with limited sensors can compete with human expertise. However, racing is a controlled environment, and it remains unclear how well this approach would work in real, unpredictable settings.

The researchers point to the prospect of fast and energy-efficient autonomous robots. Where exactly the technology will be applied first is not yet known.

Share This Article