Robo-drones learn to land by going bug-eyed Landing a flying object can't be that hard: even bees can manage it. That's why researchers from Sweden, Germany and Australia have looked at what bees could teach about landing strategies for unmanned aircraft. Of course, if your UAV is a quad-copter with a human operator or good automation, landing isn't so hard: return it …
I'd go with a light field camera. Minimal focusing necessary.
Altitude clues would be provided by a wide lens, where the surrounding surface expands rapidly at lower altitudes.
Master how bees land on a level surface, the next hurdle would be how they land on a moving flower.
Master that and have 100% foolproof carrier landings too.
me three,
i want that job, does this mean a set of increasingly difficult set of landing conditions leading to awesome top gun bees ?
maybe some obstacle courses ....
and by extension robust drones, but I prefer to concentrate on the bees.
They have compound eyes. When the target fills up a certain percentage of the bee's eyes then they have arrived.
If a thing is below a certain level of cells then you are going to fly above it.
If there is a bright pretty blue light . . . oh so pretty . . . (DON'T FLY INTO THE LIGHT!) . . . ZAP! Aggghhh . . .
Bees crash land all the time, even without a spiral pattern. High speed films of bees flying around and landing show they are horrible flyers, and Mother Nature's solution was to make them rugged as hell, and make a lot of them per colony. If one breaks, another can take its place. Not sure if designers are going to take the cue.
You're destroying my idea of strapping bees to the head of UAVs and let them do their beeish things to control the aircraft.
But for the described model to work it doesn't matter whether bees are good flyers or landers. All that matters is that it's a simple model that could easily be implemented and is independent of other high-tech stuff such as GPS.
"You're destroying my idea of strapping bees to the head of UAVs and let them do their beeish things to control the aircraft."
Well there was a project in the 50's for homing pigeons to used for ballistic missile guidance (back when inertial measurement packages weighed several hundred pounds.
It does sound like the sort of thing you could do with a couple of cameras and cheap lenses.
Note that word universal as well.
This could be the core for a whole new generation of UAV control systems.
Thumbs up for this.
Actually it was a glide bomb guidance idea in WW2. Project Pigeon:
http://en.wikipedia.org/wiki/Project_Pigeon
I saw this in a documentary a few years ago and apparently it was working pretty well.
It was revived as Project Orcon in '48, but that was cancelled in '53 because electronic guidance systems were starting to work reasonably well, and were more practical than using "wetware".
I agree that this approach will handle the final approach quite nicely since its pretty close to what is taught and I do in a glider, though I wouldn't dream of using it for speed control (the ASI does a better job). However it breaks down in the final stages of landing because it offers no help judging when to flair or for flying a fully held-off landing: that is best done by monitoring landing area perspective changes after flairing.
IOW, any algorithm based solely on monitoring the expansion of a view of the landing spot is unlikely to be successful for landing a fixed-wing aircraft because it implies that the aircraft speed will become zero at or immediately before touch-down. However, it may work for aircraft with hovering ability, e.g helicopters and quadra- or hexacopter UAVs.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
