Communications of the ACM
Appreciating a Flower's Texture, Color, Shape Leads to Better Drone Landings
A team of researchers at TU Delft and the Westphalian University of Applied Sciences have developed an optical flow-based learning process that allows robots to estimate distances through the visual appearance of objects in view.
Credit: Delft University of Technology
An optical flow-based learning process developed by researchers at the Netherlands' Delft University of Technology (TU Delft) and Germany's Westphalian University of Applied Sciences lets robots calculate distances through the visual appearance of objects in view.
This approach boosts the navigation skills of small aerial drones by mimicking flying insects' intelligence.
TU Delft's Guido de Croon said enabling robots to interpret optical flow, as well as the visual appearance of objects in their environment, would eliminate certain limitations to optical flow.
The solution involved the drones inducing optical flow oscillations to perceive distances to objects in the scene via visual cues.
Westphalian's Tobias Seidl said, "The presented learning process forms a novel hypothesis on how flying insects improve their navigational skills, such as landing, over their lifetime."
From TU Delft (Netherlands)
View Full Article - Link to Publication Homepage
Abstracts Copyright © 2021 SmithBucklin, Washington, DC, USA
No entries found