AlphaPilot Challenge: Winning Team’s AI is Slower than DRL Pilot… For Now

Last year Lockheed Martin and The Drone Racing League (DRL) announced the AlphaPilot Challenge, a contest that would explore the boundaries of robotics and AI by inviting teams from around the world to develop software capable of piloting a racing drone.

On Friday at an old newspaper printing plant in Austin, the fourth and final race took place on DRL’s Artificial Intelligence Robotic Racing (AIRR) Circuit. The winning team, MAVLab, was made up of academics from the drone research lab of the Delft University of Technology in The Netherlands.

They walked away with a $1 million cash prize, sponsored by Lockheed Martin, after their drone autonomously flew through the track in the quickest finishing-time – 25% faster than the drone in 2nd place.

The wider aim of the four-event series was to advance the development and testing of fully autonomous drone technologies with real-world applications in mind, including disaster relief, search and rescue and space exploration.

In total, nine teams from around the world took part. Each as tasked with developing AI capable of piloting DRL’s RacerAI drone without any GPS, data relay or human intervention. The code was applied alongside AI-at-the-edge compute platform from NVIDIA, the Jetson AGX Xavier.

“The AlphaPilot open innovation challenge is about going fast, taking risks and pushing the boundaries of AI and autonomous flight,” said Lockheed Martin Chief Technology Officer Keoki Jackson.

“We are excited to recognize Team MAVLab and award them with the $1 million prize, but the most energizing part of this competition is seeing how Lockheed Martin’s partnership with DRL inspired great emerging global AI talent to help our world leverage AI and autonomous technologies.”

“Robotic sports, like drone racing, push the limits of speed and performance while creating thrilling opportunities to test and refine tech innovation for the real-world — as seen through DRL’s groundbreaking partnership with Lockheed Martin,” said DRL CEO and Founder, Nicholas Horbaczewski.

“We are incredibly excited for Team MAVLab, who designed the winning AI for high-speed racing drones in AIRR, contributing to the future of autonomous flight, which will revolutionize emergency response, aerial surveying and urban package delivery.”

Read more: DRL’s RacerAI Drone is Here to Push Boundaries and Take on Human Pilots

Read more: Ethics in the Drone Industry & AI’s Slippery Slope

Team MAVLab takes on DRL Pilot

After beating their fellow AI competition, team MAVLab participated in an AI vs. human-piloted drone race against DRL Pilot Gabriel “Gab707” Kocher – widely regarded as one of the best drone pilots in the business.

Gab707 completed the circuit five seconds faster than Team MAVLab’s AI, so human pilots have the upper hand for now. The gap between the two will no doubt narrow in the years to come.

“I can’t wait to see how autonomous drone technology innovation will continue to evolve over the next few years,” said Christophe De Wagter, Team MAVLab Lead.