A drone piloted by NASA’s artificial intelligence has taken on a DRL FPV pilot, with interesting results.
If futurists are to be believed, most of our jobs are under threat from the rise of robotics, artificial intelligence and automation. The ironic thing is that some of the first to go could be drone pilots; a professional that has only existed for a handful of years. Under that umbrella is the even more recent vocation of the professional drone racing pilot. Surely these guys and girls won’t be replaced by ‘bots before their dreams have even come to fruition?
That was the scale of the challenge faced by DRL pilot Ken Loo, also known as FlyingBear, when he took on a drone packed with AI technology developed in NASA’s Jet Propulsion Laboratory. Sure, to him it was just a race like any other. But really he was representing mankind against machine.
So What Happens When AI Takes on a Professional Pilot?…
Take a look at how the race unfolded:
To be fair, the AI drone was pretty smooth and did remarkably well on the tiny and very compact track. But what it had in smoothness it lacked in racing aggression. It had an average lap time of 13.9 seconds, compared to Loo’s 11.1 seconds per lap.
However, the NASA team behind the AI drone were keen to stress that their quad had completed the course with more accuracy.
So how does it all work? The race was the culmination of a two-year Google-funded project into autonomous aerial systems and computer vision. The drone uses cameras to track its position and matches those images with a map that’s pre-loaded into its memory.
“We pitted our algorithms against a human, who flies a lot more by feel,” Rob Reid, of the Jet Propulsion Lab, said in a statement. “You can actually see that the A.I. flies the drone smoothly around the course, whereas human pilots tend to accelerate aggressively, so their path is jerkier.”
This Is Just The Beginning
Thanks to Ken Loo, it looks like humanity is in a strong position for the time being. But NASA’s automated drone did perform impressively. In fact, NASA’s Reid ominously insists that “our autonomous drones can fly much faster. One day you might see them racing professionally…”
Before that day comes, NASA and Google are hoping to refine the technology to the point where it can be used to operate drones in tight spaces, such as warehouses.