AI Drone from TU Delft outpaces human pilots at World Championship in Abu Dhabi

For the first time, a fully autonomous drone has defeated human competitors in an official drone race. A team of students and researchers from TU Delft claimed victory at the A2RL Drone Championship in Abu Dhabi last weekend, outpacing not only 13 other AI teams but also professional human pilots.

The Delft drone flew using artificial intelligence, relying on a single forward-facing camera as its only input. This setup is comparable to the perspective of human FPV (first-person view) pilots. Despite that limitation, the drone reached speeds of up to 95.8 km/h on a technically challenging track. The team first won the AI competition and then went on to triumph in the knockout tournament against human finalists.

The racing AI was developed by students and researchers from the MAVLab, part of the Faculty of Aerospace Engineering. “I always wondered when AI would be able to compete with human drone racing pilots in real competitions,” said team leader Christophe De Wagter in a press release. “I’m extremely proud that we were able to make it happen already this year.”

Springboard

The A2RL championship aimed to explore the limits of ‘physical AI’ by having drones perform under demanding conditions with minimal hardware. Whereas earlier breakthroughs in AI mostly took place in virtual environments such as chess games, this is an example of successful AI in the real world.

De Wagter hopes his team’s achievement will serve as a springboard for real-world robotic applications. “Flying drones faster will be important for many economic and societal applications, ranging from delivering blood samples and defibrillators in time to finding people in natural disaster scenarios.”