An autonomous quadrotor beat two human pilots in a drone race


Unlike self-driving cars, autonomous drones can generally come from “A” to “B” for sure, but could they beat a human pilot in a drone race? So far the answer has been “no way”, but now, researchers from University of Zurich (UZH) created an algorithm that allowed an AI drone to beat two human pilots on an experimental racetrack. The work could lead to more efficient rescue drones, deliveries and other jobs.

In the past, researchers have built simplified models of quadrature systems or flight paths to calculate the optimal path. However, this time they fully took into account the limitations of the drone. “The key idea is, instead of assigning parts of the flight path to specific waypoints, that our algorithm only tells the drone to go through all the waypoints, but not how and when to do it,” said PhD student and author Philipp Foehn.

For AI versus the human race, the researchers let human pilots train on the track to make the comparison fair. They set up external cameras to send the exact position of the drone to the algorithm in real time. Once both humans and AI were trained, the algorithm beat humans in each round and had more consistent startup performance.

Research could lead to faster drones for real-world applications, even in complex multi-point environments. The next step is to make the system less computationally demanding and allow it to work with the ship, instead of with external cameras. “This algorithm can have major applications in drone package delivery, inspection, search and rescue, and more,” said Davide Scaramuzza, head of UZH’s robotics and perception group.

All products recommended by Engadget are selected by our editorial team, regardless of our parent company. Some of our stories include associated links. If you purchase something through one of these links, we may earn an associated commission.

Source link


Please enter your comment!
Please enter your name here