NASA’s AI-Piloted Drones Race against Pro Human Drone Pilot

Researchers at NASA’s Jet Propulsion Laboratory in Pasadena, California built within the laboratory’s warehouse a twisting obstacle course for NASA’s AI-piloted drones and professional human drone pilot to race against each other.

The race, held in October of this year, culminated two years of NASA’s research into drone autonomy. The race fitted world-renowned drone pilot Ken Loo against three AI-piloted custom drones (dubbed Batman, Joker and Nightwing) all built by NASA researchers.

According to NASA, the three AI-piloted drones were built with complex algorithms for the drones to fly at high speeds while avoiding obstacles. These complex algorithms were integrated with the Tango technology of Google – the funder of NASA’s drone autonomy research.

The drones can fly as fast as 80 mph (129 kph) in a straight line. For the obstacle course built inside the NASA’s Jet Propulsion Laboratory in Pasadena, the drones could only fly at 30 or 40 mph (48 to 64 kph).

In terms of the official laps, Loo scored better than the AI-piloted drones. Loo averaged 11.1 seconds, compared to the autonomous drones, which averaged 13.9 seconds.

The results, however, showed that autonomous drones flew smoothly around the obstacle course, compared to Loo’s flight path which was jerkier as he tended to accelerate aggressively. Compared to Loo’s driving, the autonomous drones also flew more cautiously but consistently. Loo was also limited by exhaustion, something that the autonomous drones have no problems with.

“This is definitely the densest track I’ve ever flown,” Loo said in a statement. “One of my faults as a pilot is I get tired easily. When I get mentally fatigued, I start to get lost, even if I’ve flown the course 10 times.”

Racing inside a warehouse also presented a new dynamics for the autonomous drones which normally fly in open spaces and rely on GPS to find their way around. Inside a warehouse, however, the GPS will not work. The autonomous drones instead used camera-based localization and mapping technologies.

“These technologies might allow drones to check on inventory in warehouses or assist search and rescue operations at disaster sites,” NASA said. “They might even be used eventually to help future robots navigate the corridors of a space station.”