Researchers create "perching" robotic aircraft

Sure, your average helicopter can land in a whole mess of places, but what if you really need to squeeze that sucker onto a steep incline? Normally, you'd be out of luck if the slant were anything more than a measly 20-degrees, but some Georgia Institute of Technology researchers have managed something a bit more extreme. Eric Feron, an aerospace engineer, and colleague Selcuk Bayraktar have created a new breed of robot helicopter which can land, or "perch," at inclines up to 60-degrees. On a similar tip, MIT researcher Jonathan How has fashioned another unique unmanned aircraft which can take off and land from a "prop hang" position, wherein the plane stands on it's tail in midair. The hope is that these new technologies will help get aircraft into -- and back out of -- complicated spots with a new level of maneuverability. Also, the engineers probably just love blowing people's minds. Check the videos after the break to experience the gravity-defying moves yourself.

[Greg @ Nov 23rd 2007 11:45AM]

I want one

[Sean @ Nov 23rd 2007 11:52AM]

If only this were a Black Friday giveaway, you just might get one!

[Carbonize @ Nov 23rd 2007 11:53AM]

So long as you lived inside the US otherwise forget ever winning anything on Engadget.

[tallen @ Nov 23rd 2007 11:48AM]

This really isnt that ground breaking.. I made a model airplane that can do that too... all you need to do is overpower it, reduce weight everywhere you can, and make the flaps larger then usual.

[superfresh @ Nov 23rd 2007 11:54AM]

and your video is...where?

[Herman Manfred @ Nov 23rd 2007 11:56AM]

Where is the video?

Here:

http://www.youtube.com/results?search_query=rc+airplane+hover&search=Search

All the vidoes you could want.

[and that was a 5 second single search result - concentrating a little harder will get you even more, undoubtedly]

[EdZ @ Nov 23rd 2007 12:19PM]

Well done, you can fly a model plane. Now create a computer to do it autonomously.

[octoberasian @ Nov 23rd 2007 12:26PM]

If RC airplane enthusiasts can do this, then that tells us one thing:
Scientists have really lacked any sense of new innovation or discovery since the days of the transistor or light bulb being introduced, or the first telephone.

We're just rehashing current technology and make them new again. If not new, then smaller; or faster; or more energy efficient.

I think the one thing this technology above by Georgia Institute of Technology has proved is this: a robot is doing this, and not a human.

The videos of RC airplane enthusiasts doing it are still radio controlled by humans, this is done by software and other electronic parts-- automatically and pre-programmed. I think that's the only benefit here is reducing the human interaction involved with it. Give them a set coordinate and destination and let them off. Hence, why this could come in handy for UAVs for military reconnaissance. They could possibly have this perched on an incline discretely as it spies on someone or something else; or have it hover vertically in a small gap between two buildings. Of course, they would have to figure out how to make that electric motor a lot more quieter if they're intending this for reconnaissance and such.

[David Clark @ Nov 23rd 2007 12:42PM]

I have to agree.. this is very old.. very very old.

[firebat45 @ Nov 23rd 2007 2:21PM]

Actually, most of the RC pilots that hover the planes like this are using gyros to do a lot of the work for them. I'm pretty sure some of the gyros will even hold it in a steady hover for you if you let go of the sticks.

[Russ @ Nov 23rd 2007 4:12PM]

"Why the fax machine is nothing but a waffle iron with a phone attached." - Grandpa Simpson

[roger_huston @ Nov 23rd 2007 11:56AM]

Dude, this is not new. Radio Stunt planes have been doing this for years. So now someone programmed it, big deal.

-Roger

[Warren @ Nov 23rd 2007 11:59AM]

It's not Georgia Tech University. It's the Georgia Institute of Technology. Unless, that is, you are going to start calling MIT "Massachusetts Tech University."

[Chris @ Nov 23rd 2007 12:04PM]

Now do it in a 30 knot cross wind with flurries...yeah, that's what I thought.

[Josh @ Nov 27th 2007 4:45PM]

30 knot cross-wind with flurries... indoors, huh? maybe the A/C was on :)

[Mark Richardson @ Nov 23rd 2007 12:31PM]

*yawn*

[Herman Manfred @ Nov 23rd 2007 12:42PM]

Though I can believe the ENTIRE hovering RC plane video is "autonomous", the video and wording of the source article suggest the autonomous part is the launch only.

BTW, if NOT then MIT has developed some dynamite instantaneous feedback and "video parsing" software that beats out even the DARPA cross-country challenge!

[Josh @ Nov 27th 2007 4:45PM]

FYI: launch from perch, transition from hover to level flight, transition from level flight to hover and perch were all done autonomously... there is no human in the loop. That is, the computer flying the airplane was programmed to tell the thing to launch, fly around a bit, return to hover, then return to its perch again.

[Herman Manfred @ Nov 27th 2007 6:48PM]

I read the first page of the original pub (AIAA site?) and it suggested (by omission!) that a human WAS in the loop for all those parts of flight IN-BETWEEN the stages you note.

that is - computer launch, human move to good location, computer hover, computer transition hover to flight, unknown fly around, computer transition to hover, human move to good location for perch, computer perch.

But, anyhow, this article has now moved way off front page so, by definition, of no further interest!

Lol!

[Herman Manfred @ Nov 27th 2007 6:49PM]

Oh, i see I made about the same comment down below a few posts a few days ago...

[Dan @ Nov 23rd 2007 1:02PM]

I agree that is would have been nice if it were made more clear exactly what we're seeing; how much of it is by live human direction, how much AI, etc. Along with that should be footage showing us the box(es) responsible for the real time flight. Do the boxes know the exact dimensions of the room? Is real time space measurement going on during flight? I'd call this footage a tease more than a test.

[Herman Manfred @ Nov 23rd 2007 1:17PM]

On re-reading of everything I could find, including tracking down the link to the actual conference proceeding (that Engadget got wrong in the article) it appears the plane takes off, hovers, docks, transitions to level flight, transitions to hover. Maybe ti also flies circles.

All "autonomously".

Note that transitioning from hover to landing mode is NOT covered by the above, etc etc etc.

[scolen2 @ Nov 23rd 2007 2:39PM]

OMG, seriously?

ITS CALLED 3D FLYING, LOOK IT UP!

Everything the plan or heli is doing can be down by anyone, the only thing going on here is that some "software" is making it do it. I can set up my 3D plan with 3 gyros to keep it locked into a position, i dont' see why this is special.

Helicopter: http://www.youtube.com/watch?v=C2m451SbvhE

Airplane: http://www.youtube.com/watch?v=Zc2zBEc6VU4

Want more?

Chip Hyde: http://www.youtube.com/watch?v=ISzt8t77qEY&feature=related

[David Fine @ Nov 23rd 2007 3:41PM]

weird

[Zeus the God @ Nov 23rd 2007 3:53PM]

I fly RC planes like that all the time. Whats the big deal other than the fact that its a robot?

[bob @ Nov 23rd 2007 4:53PM]

Any hobby grade r/c heli can land at any angle. Even up side down. Just reverse the collective pitch as soon as you hit the landing pad. Only problem is you can't power down.

[skhawaja @ Nov 24th 2007 4:29AM]

you guys don't read that much do you?

right now they are using external input - the helicopter move may not be that amazing when you look at it - and then the description about using velcro ads etc raises eyebrows - I would imagine throwing up a velcro roof top cover and having the drone target that with it's autonomous systems for landing while I'm driving along - mobile recon anyone?

as for the prop hanging plane - that's pre-programmed flight - and I sure as hell can bet that a computer controlling a RC plane is faster than my twitching fingers today - and by grabbing the info from the onboard sensors you can have a very stable and very agile plane that can blend with it's surroundings - tree top solar array for a recharging perch - go on patrol at a certain time - return to perch and upload data or just do a live feed - very nice implementation of control - and because it is in the air you don't have the extremely limiting environments that the DARPA challenge has to traverse :]

as for the environmental factors such as cross winds and flurries - come on - something that has a preplanned flight with internal systems telling it that it is going off course is not a huge stretch - lets look at my coordinates in relation to what route I am supposed to be at - am I getting thrown off course? let me check that atmospheric data - what flight deck should I be at now? you can extrapolate from there.

Can I get a patent on the above process?

[Brandon @ Nov 28th 2007 1:44PM]

There are a bunch of other videos of the MIT vehicle flying autonomously at http://aerobatics.mit.edu/videos.html.