NASA drone race pits human pro vs AI
The JPL engineers had recently finished developing the artificial intelligence for three custom drones designed to autonomously navigate an obstacle course at high speed. The three drones – dubbed “Batman,” “Joker,” and “Nightwing” – were built to racing specifications and could fly up to 80 mph in a straight line, but only at 30 or 40 mph on the obstacle course before needing to brake.
To test their progress, JPL set up a timed trial between their AI and world-class drone pilot Ken Loo. According to NASA, the race capped two years of research into autonomous drones.
“We pitted our algorithms against a human, who flies a lot more by feel,” says Rob Reid of JPL, the project’s task manager. “You can actually see that the AI flies the drone smoothly around the course, whereas human pilots tend to accelerate aggressively, so their path is jerkier.”
Initially the AI and human pilot started out with similar lap times, but after dozens of laps Loo learned the course and became more creative and nimble. While the drones flew more cautiously and consistently, say the researchers, Loo attained higher speeds and was able to perform impressive aerial maneuvers.
For the official laps, Loo – despite being limited by exhaustion – averaged 11.1 seconds, winning against the autonomous drones, which averaged 13.9 seconds.
“This is definitely the densest track I’ve ever flown,” Loo says. “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.”
The AI drones did sometimes move so fast that motion blur caused them to lose track of their surroundings, so more work is needed, the researchers say. Still, it may only be a matter of time before they are flying professionally, says Reid.
The drones used camera-based localization and mapping technologies to navigate the indoor race course. These technologies, say the researchers, might allow drones to check on inventory in warehouses or assist search and rescue operations at disaster sites, or even be used to help future robots navigate the interior of a space station.
The drone algorithms were integrated with Google’s Tango augmented reality computing technology, which JPL also worked on. The race capped two years of research into drone autonomy, which was funded by Google.
Related articles:
“Drone Days” at research campus to showcase aerial robotics tech
‘Slaughterbots’ video depicts autonomous killer drones
Amazon drone tower described in patent application
