Drones are not entirely under the control of humans, make way for the artificial intelligence. NASA Jet Propulsion Laboratory (JPL) has performed an experiment in which AI-controlled drones were pitted against human-controlled drones, and the AI has performed beyond expectations.
It was interesting to see that a robot-controlled drone was able to perform so well and pave way for a new opportunity for AI in the field of drone missions.
AI's Performance In Drone Flying
The race took place on Oct. 12 and is funded by Google. The funding is part of drone autonomy research.
"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," says Rob Reid, projects task manager.
JPL researchers have been able to develop a vision-based algorithm for navigation of the spacecraft. Now, researchers are using the same algorithm for drones. To demonstrate as to how the algorithm works, the researchers built an obstacle course for the drone to clear.
After the AI drone did its job, the researchers called in Ken Loo, a professional drone pilot to fly the same course with his drone piloting skills.
The results of the experiment were published on NASA's website and the video of AI drone's performance was also uploaded to YouTube.
Results Of Human Pilot vs. AI Are Here
The results can surprise anyone. The AI-controlled drone completed one lap with an average time of 13.9 seconds and the human pilot Loo cleared with a flying average of 11.9 seconds.
With this result, it is clear that a human professional drone pilot can win against artificial intelligence. At Least for the obstacle course set up by the JPL researchers.
The real meat of this experiment lies in the AI's ability to fly the drone with more cautious and consistent performance. The human piloted drone flew with more speed and was jerky. Whereas the AI took its time which, is slightly more than the human-controlled drone, but its performance was far superior when it comes to accuracy.
Another striking thing that the researchers noted was that that the human drone pilot was prone to fatigue, whereas the AI does not constraints itself within such a parameter. It can chart its course and correct its moves many times over, without any signs of lag.
"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," says Loo.
It is interesting to see how far and how quick AI drones will start to fly and its implications are far-reaching.