U. Penn Super Quadcopter Learns New Tricks 124
Freddybear writes "University of Pennsylvania's GRASP lab posted new video of their scary fast maneuverable quad-rotor drone. It can now fly through openings (hoops) which are themselves moving." The entire list of GRASP projects is worth crawling through.
Killing me in my sleep? (Score:2, Informative)
I don't think so. That sucker is NOISY.
Cool aerobatics though.
Re:Killing me in my sleep? (Score:5, Insightful)
Re: (Score:2)
Can't wait for a version with two swiveling turbines at the tips of small wing, eh?
Re: (Score:3, Informative)
Can't wait for a version with two swiveling turbines at the tips of small wing, eh?
Something similar was : http://en.wikipedia.org/wiki/Hiller_Hornet [wikipedia.org]
Re: (Score:2)
Fairey Rotodyne was probably a more sucesfull example of those ideas. But...no, not really similar at all.
Similar would be Bell/Agusta BA609, V-22 Osprey, perhaps NASA Puffin. This one [wikipedia.org] pretty damn close, ze Germans had something even closer [wikipedia.org], and here is almost the real deal [wikipedia.org], just not with turbines or ducted fans ("cancelled", riiight ;) )
Pretty close to proper aerial HK [goingfaster.com]
Re: (Score:2)
"It does make a distinct sound and sounds like that in a war-zone can have a terrifying effect, psychologically on enemy troops."
Until they learn to range it by sound and take countermeasures...
"Frightfulness" gets the low-hanging fruit, but disciplined men have stood incredible barrages (WWI being the best example) and stood fast despite losing tens of thousands killed in a single day.
Re:Killing me in my sleep? (Score:5, Funny)
Re: (Score:2)
Re: (Score:2)
This should be modded insightful. The aircraft have no inertial references AFAIK. All they are is four remotely controlled motors. All of the real-time processing for the control is done elsewhere in the room. The aircraft have retro-reflective markers and external motion capture is used to get feedback for the control system. If you were to try and reproduce their results, the stumbling block would be the price of the Vicon motion capture system with 1000Hz sampling rate cameras. Those aren't cheap.
Re: (Score:2)
Re:Killing me in my sleep? (Score:5, Funny)
They could use it to track autonomous Audis, better than crashing helicopters.
So technically you're awake for a second (Score:3, Insightful)
Sure it's noisy, but did you see how fast that thing moved? You'd be awake for about a second, max, before it was on you.
Re: (Score:2)
Not faster than a Phalanx with Lockheed Martin's new laser system mounted on it.
http://en.wikipedia.org/wiki/Phalanx_CIWS [wikipedia.org]
So, until they're faster than light, I'm not concerned =)
Re: (Score:2)
Not faster than a Phalanx
So you're saying you have a Phalanx in your bedroom?
Wait, don't answer that - TMI.
Re: (Score:1)
"See-whiz"
Re: (Score:2)
Now if we made a cross between that, and this [youtube.com], we'd have the perfect defense :-)
Re: (Score:2)
Not faster than a Phalanx with Lockheed Martin's new laser system mounted on it.
Excellent! So we can build machines to destroy the machines we build to destroy machines!
I'm finally starting to understand the whole "logic" of this "war" stuff: instead of using this incredible technology to create a world of plenty with nothing much worth fighting over, we will use it to destroy things, thus sapping the economic productive capacity of the world, reducing our opportunities to trade, and cause our enemies to invest in more useless, deadweight loss electro-mechanical junk to stop our dead
Re:Killing me in my sleep? (Score:5, Informative)
It would need much more sophisticated navigation in order to operate outside its little custom-made "cage".
Re: (Score:2)
Re: (Score:3, Informative)
Actually, it does not need to be a killer. With a camera, gps, and a targeting laser, it can zoom in, acquire a target, zoom up out of range, and let a remote site fire a missile up to kill targeted item.
Scary.
Air-dropped drones with man-portable controls (Score:2)
I've always wondered if air dropped drones with man-portable field controls would ever be practical.
I'd see them as disposable, capable of low-level surveillance, and armed with lighter armaments (7.62 machine gun, or RPGs, or maybe a pair of dumb rockets) and have all of them packed with high explosive so they could also be flown kamikaze when their primary armaments were used up or their useful life was completed and detonated in emeny areas as a bomb.
Control would be via field-portable briefcase type con
Re: (Score:1)
Re: (Score:2)
Re: (Score:3, Interesting)
I'm ok with using it to kill clowns. That seems like a public service.
Re: (Score:1)
Re: (Score:2)
Nope. The "bright red lights" are infrared light sources of the motion tracking system, mounted on the walls. Most likely a Vicon system with 1000Hz cameras. The IR illuminators show up on color CMOS-sensor cameras as red.
The hoops don't have to be colored at all -- they have retro-reflective markers for the motion capture system, IIRC.
There's nothing on the aircraft that would "see" anything, it's merely an airframe, battery, motor controller, and a radio receiver. All of the control is done elsewhere in t
Re: (Score:2)
I don't think so. That sucker is NOISY.
Cool aerobatics though.
First: make it work
Second: make it quiet.
Works for me. If these people don't have a military contract, the they're doing it wrong...or...they have the intestinal fortitude to tell them to FFFFFFFFUUUUUUUUUU!
Way cool (Score:4, Funny)
Re: (Score:2)
Re: (Score:2)
I would prefer lasers instead -- fired bullets generate quite a bit of recoil.
Yeah, it obviously has difficulty dealing with erratically changed flight patterns, so I'm sure the known force of the recoil would be a huge problem. I doubt this thing could carry much more than a laser pointer anyhow. Leave the freakin laser beams where they belong, on sharks heads.
Re: (Score:2, Funny)
still no match for airsoft i suspect:
http://www.youtube.com/watch?v=4HGmoxotBi8 [youtube.com]
Navigation (Score:5, Insightful)
Re: (Score:2, Informative)
Re: (Score:2)
I dunno what all the practical applications of this tech are but all I can say is, I want one.
If you can shrink the vehicle down to fly size and make it as quite as a fly with the addition of a microphone or a mic w/ a camera it certainly would give new meaning to the expression "fly on the wall". I think the flight path control and anticipating the location of a moving object real time is the impressive part more so than the vehicle.
Re: (Score:3)
http://www.diydrones.com/ [diydrones.com]
http://code.google.com/p/arducopter [google.com]
How about a Hexacopter! It's got 6 rotors (Score:1)
If you like this, then you should have a look at the hexacopter. It is available as a kit from a company in Germany. It has on-board GPS and gyros and is controlled with a basic RC controller. The cost is about $1000. Not cheap but not outrageous either. The big one can handle a payload of up to 1 Kg! Think about the uses!! Cameras, terrifying the neighbor's dog, pizza delivery in minutes no matter what the traffic conditions are . . .,
Link to the YouTube Video here:
http://www.youtube.com/watch
Manhack (Score:5, Interesting)
Reminds me of the Manhacks from Half Life.
http://half-life.wikia.com/wiki/Manhack [wikia.com]
Re: (Score:2, Interesting)
Weaponize! (Score:2)
So when can we mount some freaking laser on it?
Re: (Score:1, Funny)
Re: (Score:1)
Re:Nothing new to see here (Score:5, Insightful)
"We've already got quadricopters. And you can control them with an iPhone. And they could already do this sort of maneuvering."
(A) Yes we do. (B) So what? An iPhone is hardly an optimized flight controller. It's a toy. (C) No, they can't. If you can maneuver your quadricopter with your iPhone anywhere near the way this thing can move, I'll eat your shorts.
I am pretty sure that my digestion is safe for the time being.
Re: (Score:1, Insightful)
Uh, he is safe for the time being seeing as you need a $500k VICON system to pull off the sorts of maneuvers the quad rotors are pulling off in the video. You seriously think this is controlled solely by on board cameras or even one external camera? Why do you think in all these videos you see these glaring red lights? They are the locational cameras (GRASP Lab has 16 I think) identifying where in space the quad rotor is and it's orientation. There is a huge external computation component that you don't see
Re: (Score:3, Informative)
For those modding me down.
Eat your already-outdated quadcopter.
http://www.youtube.com/watch?v=gvH2f-AewX8&feature=related [youtube.com]
with built-in GPS location services, auto-stabilization, call-home function, can carry a camera, and can likely carry a micro-radar for autonomous navigation.
Keep on dreaming while I've got investments in this stuff.
Re: (Score:1, Informative)
no it isnt. the source code is proprietary. only the SDK is available and it is highly crippled. please show me how you can EASILY make this work with anything. idiot.
Re: (Score:3, Insightful)
Yes, because those things fly so blindingly fast, and autonomously. Right...
Re:Nothing new to see here (Score:4, Informative)
"no it isnt. the source code is proprietary. only the SDK is available and it is highly crippled. please show me how you can EASILY make this work with anything. idiot."
And so it is. The page you linked me to clearly states that it is an SDK, not source code.
Uh, he is safe for the time being seeing as you need a $500k VICON system to pull off the sorts of maneuvers the quad rotors are pulling off in the video. You seriously think this is controlled solely by on board cameras or even one external camera? Why do you think in all these videos you see these glaring red lights? They are the locational cameras (GRASP Lab has 16 I think) identifying where in space the quad rotor is and it's orientation. There is a huge external computation component that you don't see doing most of the heavy lifting required to control this quad rotor.
I'm fairly certain this can't be replicated by a quad rotor with merely an iPhone. That's kinda laughable.
I repeat: if you can maneuver a Parrot AR Drone with your iPhone anywhere near the way this thing can move, I'll eat your shorts. (I think it's obvious that would include flying through both stationary and moving hoops of comparable size.)
And I repeat: "I am pretty sure that my digestion is safe for the time being."
Don't misunderstand: I think the AR Drone is very cool in its own way. It's a great toy. And it might even be useful for a few semi-practical things. But comparable to the device in the video OP linked to? Not.
Re: (Score:1)
I repeat: if you can maneuver a Parrot AR Drone with your iPhone anywhere near the way this thing can move, I'll eat your shorts. (I think it's obvious that would include flying through both stationary and moving hoops of comparable size.)
It doesn't matter if dude can replicate these maneuvers with a Parrot -- after hours and hours of practice and maybe nail it once in twenty tries. The computers performing the Penn maneuvers can reliably repeat them over and over again, and could fly as many quadrotors as whoever's footing the bill cares to buy. That's the difference the "autonomous" part makes.
And if you think "Bah, that's impractical - you need a big computer somewhere off-board to do the calculations" then you need to reread Moores Law [wikipedia.org]
Re: (Score:2)
As for the impracticality I agree with you, except that it is in fact impractical today. Tomorrow, maybe not. But tomorrow isn't here yet.
Re: (Score:1)
And I repeat: "I am pretty sure that my digestion is safe for the time being."
It's your taste buds that I'm worried about.
Re: (Score:2)
No way your average quadcopter controlled by an iPhone is going to be as maneuverable as the one in the video. The ones controlled by an iPhone use stability control and the pilot is simple telling it to move in a certain direction. It's very different to what you can do once you learn to fly manually, or in this case, an advanced auto pilot system.
While impressive (Score:5, Informative)
Oblig Skynet (Score:3, Insightful)
Those little buggers are sure going to be hard to shoot down. Y'all better start skeet shooting right away if the human race is to have any chance!
Re: (Score:3, Insightful)
http://en.wikipedia.org/wiki/Phalanx_CIWS [wikipedia.org]
No, no they won't be hard to shoot. Fight science....with SCIENCE!
Re: (Score:2)
The autocannon! The Centurian, land-based Phalanx model that can fire at ground targets sounds pretty scary, especially if made mobile and well camouflaged.
Re: (Score:2)
I was able to see the Centurion in a live fire operation. It is a formidable opponent.
Not just aggressive, VERY Aggressive (Score:2)
Skynet's Very-Aggressive Quadrotor Parody [youtube.com]; the video that had to be done...
Re: (Score:2)
This should do it - optional dog for fetching (Score:1)
Remington 1100 Tactical [remington.com].
Best. Tie-In. Ever. (Score:2)
So, how long will it be before we can buy remote-controlled mini Starfuries?
Re: (Score:2)
Point-Picking systems are old-hat at this point, and walking between picked points is equally easy. I'd agree its a training-wheel approach that keeps the processing load much lower, but this is no great hurdle or limitation.
Not coincidentally, this is exactly how a human pilot flies, you are specifically taught to NOT trust your inner ear, but rather only instruments and what you see out the windshield (picked reference points). So what you're calling a weakness is strikingly close to regular old flight sc
Re: (Score:2)
Nope. The only sensors on the quad are motor speed sensors.
Good timing (Score:3, Insightful)
If I throw a hoop up at the right time, I can get a rubber-band to fly throught it.
Re: (Score:2)
If I throw a hoop up at the right time, I can get a rubber-band to fly throught it.
Cool, now show me how you can also get it to go through three hoops at different heights multiple times in a circular pattern.
Re: (Score:2)
That part WAS impressive, assuming it wasn't a preprogrammed pattern. I'm just not sure what I could assume about the thrown hoop demo that would make it interesting.
Re: (Score:3, Interesting)
I thought the recovery after being thrown was rather impressive as well.
I think if they have had several hoops swinging from strings at different speeds and heights as well as off set and it could somehow navigate that, I would be very impressed with is path negotiation abilities. As it is, I agree with you, it was had to tell exactly what they were demonstrating by throwing the hoop.
Re: (Score:3, Insightful)
Re: (Score:2)
The motion is pre-programmed in the sense that once the hoop is flying, it follows the ballistic trajectory. The control algorithm is set up to depend on that, and it choses an optimal trajectory for the quad to get through the hoop, knowing how the hoop moves in the immediate future. If you added a thruster on the hoop to make it non-ballistic, you could easily get the quad to crash. To make the quad become a universal hoop-jumper, whether the hoops move or not, the control system would need to learn how f
Re: (Score:1)
Re: (Score:2)
Air Force knocking (Score:2)
Re: (Score:1)
OMG ... reminds me of the V-1 buzz bombs the Germans dropped in London between 1942-45.
Something tells me that you were not actually in London at the time.
Re: (Score:1)
Jedi training (Score:4, Funny)
Looks like the lightsaber training ball used by Luke
They zippin' through your window (Score:4, Funny)
So y'all need to
Hide your kids, hide your wife
Hide your kids, hide your wife
And hide your husband
Cuz they killin' errbody out here
Re: (Score:1)
Hide your kids, hide your wife
And hide your husband
Who do you think you're singing to, bud, bi-sexual Mormons??
Twenty cameras. (Score:2)
It's very cool stuff and amazing work. Though, until I install twenty IR cameras around my room and send the information back to the controlling computer, I think I am safe for now.
Gizmodo. (Score:1, Flamebait)
It's trash, stop linking to it. They're just sensationalist writers and the site amounts to little more than the national enquirer of tech blogs. These morons need to be boycotted.
Forget the sharks (Score:1, Redundant)
I want lasers on these things!
Showed video to wife (Score:2)
Agreed.
I, FOR ONE, WELCOME OUR NEW QUADROTOR DRONES OVERLORDS
and btw, those rotors sound scary, sounds like a giant swarm of angry bees
No reason it can't be fully mobile (Score:5, Informative)
Very nice. For research purposes, they're using a cheap copter and expensive fixed motion tracking gear. That saves money during debugging crashes. It doesn't have to be that way. With a slightly bigger copter they could carry around 3 axes of fibre-optic gyro, good accelerometers, and a good dynamic GPS. Expect to see that soon, with DoD funding.
Robots are going to have faster reflexes than humans. Humans are stuck at 200ms or so, while computers get faster.
Re: (Score:1)
Re: (Score:1)
Video end card says "Ascending Technologies" which is here: http://www.asctec.de/ [asctec.de] (site is in German)
Re: (Score:2)
more specifically: http://www.x3d-shop.de/shop/catalog/details?sessid=sLj0f5goZLJPRcsCgtkhjXnEGJRrPR5tbCBSB2XfNGJzT0s1TOFT2Rss5cj1Svaq&shop_param=aid%3D17188182%26 [x3d-shop.de]
Not sure I'd want to pay 989 EUR for one though... ouch...
Also it does seem the team modded it to reduce those really long (compared to the rest of it) antennas....
Re: (Score:2)
Except that your 'slightly bigger copter' is 'not even fractionally as capable' - it's the motion capture gear (I.E. the environmental sensors) that make the quadrotor capable of performing the tricks it do
Re: (Score:2)
Except that your 'slightly bigger copter' is 'not even fractionally as capable'
It exists. [youtube.com] That's what an autonomous helicopter with onboard sensors can do as of a year or two ago.
Environment sensing is coming along. Check out Advanced Scientific Concepts' flash LIDAR. [advancedsc...ncepts.com] (Still too expensive, but it's a tenth the size of what it was five years ago. I saw the optical bench prototype in 2003, when it was the size of a desk.) Simultaneous Localization and Mapping finally works.
All that fixed motion trac
Re: (Score:2)
Well, either you didn't actually watch the quadrotor video or you didn't actually watch the autonomous helicopter video. When you do so, you'll note not only the vast differences in performance between the two - you'll also note that the autonomous helicopter did no obstacle avoidance, no coordinated maneuvers with other autonomou
Re: (Score:2)
"Humans are stuck at 200ms or so, while computers get faster."
While you are right, that 200ms is great for multipurpose multi-input multi capable Humans. While you can build specialized machines that can do one thing better than a human, machines aren't very good at multipurpose tasks.
Making a complicated multipurpose machine that can do more than one complex task at any given time, good luck with getting it to 200ms.
Re: (Score:1)
1:00 (Score:5, Interesting)
Did anyone else catch that at the 1 minute mark, the guy who just caught the hoop looks kinda frightened and drops the hoop, as if the Quadrotor wants to go through the hoop again?
Re:1:00 (Score:4, Funny)
He's probably taken a few hits while they were debugging. Ouch!
Onboard it (Score:1)
Not in MY house, you don't! (Score:1)