Hacker News new | past | comments | ask | show | jobs | submit login
PhD Student gets quadrotor helicopter moving aggressively (geek.com)
281 points by ukdm on May 28, 2010 | hide | past | favorite | 100 comments



Serious question, is there anything driving mobile robotics research besides the defense dept?

I spent my grad school time in a robotics lab and it always bummed me out that anytime you see something amazing like this it's invariable funded by a military research agency which can't wait to strap a gun to the damned thing.

Same thing for computer vision, except it was all about surveillance.


this doesn't answer your question at all, but it's just food for thought: imagine we traveled back in time to the 1960's ... i suppose that people would ask "Serious question, is there anything driving reliable robust nationwide packet-switched data network research besides the defense dept? it always bummed me out that anytime you see something amazing it's inevitably funded by military which can't wait to use it to plan and coordinate military strikes against the soviets and to plan for some ridiculous doomsday nuclear holocaust that might never come" ... without huge amounts of DARPA funding throughout the 60's and 70's for the Internet, we'd all still be pen pals sending snail mail to one another ;)


Fair enough, but there's nothing overtly weaponizable about communication technology.

Don't get me wrong, I'm thrilled the money is there to advance the field, I just wish it wasn't coming from a bunch of dickheads trying to strap guns to it.

I guess since the conventional wisdoms says pornography is the biggest catalyst for new media adoption, it shouldn't be a surprise that killing people drives the other engineering fields forward.


Fair enough, but there's nothing overtly weaponizable about communication technology.

Reliable and secure communication technology is what enables the military to function. Without it, all the weapons technology in the world wont do you any good. So on a moral level, if you are against war you should actually find working as a researcher on communication technology more reprehensible than working on weapons. On an intuitive level I would guess that the breakdown or compromise of military communication has resulted in a greater loss of life than advances in robotics.

The reason why I'm pointing this out is that the real world isn't as simple as "a bunch of dickheads trying to strap guns to [robots]". You're making a simplification because it's convenient, but that simplification is morally inconsistent with approving of DARPA's work on the Internet. The fact that technology has a dual use is something we all have to deal with as researchers. But I don't think it helps matters by treating the issue as a binary.


"On an intuitive level I would guess that the breakdown or compromise of military communication has resulted in a greater loss of life than advances in robotics."

Up to this point, yes. Not many robots. I suspect the next 300 years will change that ratio hugely. Networks won't kill people...bots with guns will kill people.

It's part of the researcher mindset to rationalise their (enjoyable, paid, 'important') work, otherwise it wouldn't get done. What is the likely result over the long term? Probably not something most researchers are too worried about.


As the great Tom Lehrer put it:

"Once the rockets are up, who cares where they come down? That’s not my department," says Werner von Braun.


"Fair enough, but there's nothing overtly weaponizable about communication technology."

Um, what? DARPA was funding it for a reason. How do you kill someone with a network packet? Quite simply. The packet contains "Hey, sarge, go here and kill those guys." Without modern communication you don't have a modern military.


You can also look at it from a different perspective:

The market is a much harsher master when it comes to fiscal accountability. Stuff doesn't get funded unless it's going to make money. Military spending is a magical well of infinite money with little accountability. You can think up any project you want and get money to do it so long as you say you're going to strap a gun on your robot, or that the decentralized communication network is really to help survive a nuke attack.

I used to work in bioinformatics and money from the typical funding sources was getting tight in the late 90s, early 00s. After 9/11, my department just recast every project as "terrorist bio-defense" and was financially refueled with defense money.


This is supposed to be a positive aspect of defense R&D spending?


> I just wish it wasn't coming from a bunch of dickheads trying to strap guns to it.

Dickheads?


Everything good starts out for war, entertainment or porn. Later that technology benefits regular uses which become the mainstream.

Throughout history all great minds and technological advances were close to defense development. This here internet is an example, GPS another. Good things come from war competition thats the rub.


The Wright Brothers and Ford weren't doing it for any of those, and neither were the people who invented chocolate or GM rice. People researching the solar system aren't working for the defense industry. Many, many things originate outside of war, entertainment or porn.


Using the Wright Brothers as an example might not be the best. They had a lot of dealing with US Army.



It is forever intertwined. Ford was against financing war (but also supported Hitler), but the roads and interstates that made the car viable in the US was military driven by Eisenhower's interstate system (and the railroad monopolies).

I agree things are made/innovated on outside of war, but somewhere along the line most technology is defense driven or funded.

What was the Wright brothers invention used for first? Later came commercial flight.

The solar system research is driven by science but the science programs are largely defense supported and dual purpose. For instance, many of the shuttle trips have a defense purpose or classified payloads. Even the trip to the Moon was for defense leverage on Earth.

From Archimedes to Tesla to science and technology innovation today, most technological advances are closely linked to defense funding or war technology.


I thought about this for a while a couple of months ago and wrote down my thoughts on war, robots & culture here:

http://linusthiel.posterous.com/about-war


In the USA, no, not really.


So where is this different? And how? Who is funding this outside of the US and for what?


Apparently a lot of robotics work in Japan is aimed at assisting the elderly and making up for labor shortages.


Common topics here include assisting the elderly, industrial automation, disaster response, (minorly) entertainment, (minorly) consumer goods, and one other application area I studiously ignored because smart foreigners have no opinions about Article 9.

There were a couple of interesting discussions at the prefectural technology incubator about possible applications for research which was in the news at the time. It was a (very early prototype of a) man-portable cloaking device. Clearly, we decided, it was destined for the Tokyo fashion scene.


The japanese do a lot of robot research and AFAIK most of it is not military.

I'm not sure if the non-military clause in their constitution means they can't do military research.


The Japanese have a fairly large and technically advanced military, they just call it a "self-defense force". They can and do carry out military research.


Robotics research in Japan is primarily aimed at defending against giant space monster invasion. Oh wait, I guess that falls under defense.


you'd be all paranoid too if godzilla kept invading.


Ironically, it's the civilian sector that's limiting this research to the DoD. The FAA has very strict regulations on flying autonomous drones outdoors (basically, you're not allowed to do it), so this has seriously hobbled a lot of research into UAVs, etc. Israel, Canada, Western Europe and other places are far ahead now in terms of commercial autonomous drones.


In Japan, they'd like to use robots for earthquake rescue.



The "only funded by the military" gripe only applies to the US, and is a peculiarity of American culture. In other areas robotics research is funded by a lot of other organisations besides military ones.


A lot of cognitive science (especially AI related) and cog psychology is DoD related as well. For FY2010 60% of the US governments total R&D spending is military 40% non-military.


Reference please?


http://budgetinsight.wordpress.com/2009/11/03/research-and-d...

>Historically, between 50 and 60 percent of the federal R&D investment is made in defense.

>[mentions the American Reinvestment and Recover Act]However, despite this shift in R&D investment priorities, defense R&D spending continues to outpace non-defense R&D spending by $20 billion.

>For psychology claim, my own experience

sry takes a while to respond with /noprocast on


Holy. Crap.

Imagine what could become of this research in 5-10 years of sustained development?

God help the poor sucker that needs to deal with one of those things when coupled with less than benevolent intentions. Imagine five of these, armed and working together?

Good bye urban combat, hello robotic assassins.


Yeah, my thoughts exactly. When I saw it, I instantly thought of Man Hacks from Half Life 2 - flying drones with buzz-saws attached. They were far less maneuverable than these quadcopters. Buzz-saws are for chumps, though - I'm sure these things could carry more than sufficient C4 to ensure that any human-sized target was eliminated completely in the sort of way that their relatives hope their dental records are up to date.

Terrifying.


With a Gyrojet style weapon, it would be easy to fit 4 or 5 rounds of ammo with the stopping power of a 45 caliber on such a platform. Taser rounds could also be quite doable, and even adapted to the same launcher. This would be complimentary, as Gyrojets have a minimum range of 15' or so, which is a practical max range for a Taser.


If the quadrotor helicopter gives you the willies, get a whiff of BIGDOG and his cousin LittleDog, which can accomodate rugged or unstable terrain:

http://www.youtube.com/watch?v=nUQsRPJ1dYw


And what is on BigDog's back? A pack of six of these buggers and support infrastructure for them (larger battery, for instance, or a generator). Localized air surveillance and potentially semiautonomous to boot.

And I mention this because that's just the direct application of these two devices with no (significant) new research necessary.

While the news is flooding with doom and gloom (and with some justification), times are getting interesting real quicklike where few people are looking.


It reminds me heavily of some of the bots in The Terminator movies.


Or the manhacks in Half Life 2.


That's exactly what I thought. As a clunky non-flying biped, I for one welcome our new quadrotor overlords.


Imagine ten thousand of them swarming in over the horizon, all sharing data and building a super accurate 3d map of everywhere they go. In 5-10 years time the cost of something like this would have to be less than $100. I guess power supply/range would be the main problem.


Operate them in swarms around base vehicles, where they can dock with a battery-exchanger. Each unit's endurance is 14 minutes, but the whole swarm is only limited by the fuel available to the base vehicle's generator.


Based upon my observations there is no such thing as "sustained development" in academic robotics research. Otherwise I agree that larger versions of this with a weapon could make war crimes or acts of terrorism easy to commit.


Or save some civilian lives in a war zone. The robot drone soldier might actually be able to take the time to assess the target instead of split second shooting. If a normal soldier hesitates, the the soldier could die. If a person commanding a remote drone hesitates, well then maybe you have to order a new drone or some spare parts.


Obligatory Half-Life 2 reference: http://www.youtube.com/watch?v=6PEPt2OCYSk#t=2m00s


Nice! This is more of what I'd like to see on HN.


Might be asking too much, but I'd also like to see a thread discussing the possible parts, schematics, and source code I could use to hack one of these together myself :) To start: what kind of rotors are those? Are they expensive?


It sounds like the control system is the expensive part. It uses a Vicon motion capture system to track the quadrotor using 20 cameras arranged around the room, plus custom control software running on a laptop, and an XBee wireless link from the laptop to the quadrotor to send the control signals.

The interesting part to me is the overshoot you see when the quadrotor shoots through the window, then pulls back nearer into a resting position. This implies that the control software can work out a dynamic trajectory (one where the quadrotor wouldn't be able to hover at every point along the trajectory), then reconcile it with a desired end point that won't fit onto the trajectory.

That sounds simple, but many walking robots still can't do that -- their gaits are designed so they could freeze in place at any time and still not fall over.


$100 in parts, according to this guy: http://vicacopter.com

I tried (unsuccessfully) to make his design a year ago and it was more like $300 though, and that assuming you have the RF equipment.

Guy works at NASA I think. The claim that his is the only fully open source autopilot is to my knowledge true. No docs or schematics though, but these could be adapted from other projects, http://mikrokopter.de http://diydrones.com


What about the Rotomotion guys? I thought their stuff was also released as OSS.


None of the hardware construction looks all that expensive. (I've watched the video a few times, tried to catch a few details.) They're pretty high-speed motors, obviously, and otherwise lightweight construction.

The real trick is how they're getting them to calculate vectors. They're certainly not just remote controlled -- not unless they had an awful lot of takes of that 3" clearance shot -- so that raises the question, are they precalculating the vectors, or are they using some really advanced method to calculate them on-the-fly (so to speak)?

If they're precalculating, then the impressive thing here is the method they've developed to get these things to do tricks with rotors. If they're doing on-the-fly calculation, then I really want to know what kind of sensor they're using. I don't know of anything available on the hobbyist market with that kind of resolution and speed.

EDIT: Well, this is embarrassing. It spells out the answers to my questions right in the linked article. :-( Hooray for not paying attention.


I work with the exact same setup. You're looking at 100-250k for the "Vicon" (motion capture) setup and 4k+ for the quad rotor (an Ascending Technologies Hummingbird).

That sounds like a lot, but the motion capture system can accurately track things with sub-millimeter precision at, at least, 100fps. The Hummingbird is very sophisticated, with a 3-axis accelerometer, 3-axis magnometer, 3-axis gyro and high performance brushless motors, and of course, the R&D cost for developing all of it.

DIYDrones has a lot of good info on hacking these together, but you won't get the performance you're seeing in that movie.

I've talked to people who are trying to build a poor man's version of these and they're aiming for 600-800, but it's not going to be as high performance.


Am I the only one who thinks a future filled with robots like this is scary and pretty undesirable?


Speak for yourself, I look forward to joining the resistance and shooting at them all day long.


Who are you kidding, the resistance is going to have open source versions of all these things shooting back.


Or the resistance will "jailbreak" some of these and turn them against their evil masters.


A crowbar would likely be a more effective method of dispatching them, from prior experience.


No, but what're you gonna do? Ban technology?


Limiting unscrupulous funding seems like the better option. There's no reason for DARPA and NIST to be what funds all the coolest tech, it could be some other govt sector.


Burn the witch!


I wonder what the control system for that looks like?

I had a professor who engineered control systems and I remember him showing me this giant PDE-filled matrix that described the control system for the helicopter.

One of his grad students even had a computer simulation where you could try to fly a helicopter with no control system on it. It was pretty much impossible to control. Whenever you tried to go forward, you were likely to flip over.


This is interesting. It reminds me of manhacks from half life 2 though which scares me somewhat but thats probably intensional since its so aggressive.


I'm sure the DOD would probably want a word with that guy


Who do you think probably funded this? His advisor is funded by (at least) ONR, ARL, and DARPA.


You don't think the DOD has tons of this stuff under its belt already?


If it doesn't already, it will do soon. 3 ... 2 ... 1 ... oh look, the guy just got a DOD research grant.


This was my first thought. It's sad that whenever new technology is developed, governments think of ways to use the technology to kill people.


A simple look a history suggests that 99% of all human invention begins with looking for a better way of killing each other.

Don't blame governments, we're all like that.


That's kind of a weird thing to say, since the main reason so many inventions are related to weapons is directly because political entities war with each other so much. Serial killers, cuckolded husbands, and street gangs don't usually pay for new weapons inventions the way governments do.


Yet, this is really only because every other government is doing the exact same thing. An arms race naturally has a self-defense component.


Forget recon and bomb usage. I spy swarms of assassin drones. Would make fine horror story.


As the author suggested, Is there any indication that they aren't preprogrammed flight trajectories? And with so many cameras, the instrument setting is so "brute forced" that it seems hard to imagine a setting where this work could be applied. This http://heli.stanford.edu/ seems a lot more impressive to me


The maneuvers look like something out of The Empire Strikes Back.

If the maneuvering could be calculated locally in relation to sensed objects and other quad rotors, then robust swarming behavior combined with this degree of agility is easily achievable.

Small swarms of such quad rotors equipped with video cameras could autonomously patrol a certain volume around a base vehicle. Low res video feeds could be used for automatic heuristic recognition of bipedal targets.

Perfect a lightweight weapon like a Gyrojet, and this becomes a very effective area denial and recon system. Combine this with Taser projectiles and you could have a very effective crowd control or prison security system.

14 minute battery life doesn't seem like a lot, but with a swarm, individual units could be programmed to return and perch on a battery exchange dock. Such a system could be installed on a vehicle like a Humvee or an APC.


well...crap. This makes the AR.Drone ( http://ardrone.parrot.com ) I've secretly been craving for the past couple of months seem utterly boring.


This is pretty cool. However, as a technical note for others, if you decide to develop a state of the art fast-moving RC helicopter I suggest you record your experiments with a video camera that has a higher frame rate than, seemingly, 5fps.

Edit: my apologies, after re-watching the video it looks fine. It must have just been a problem on my end (although I've never seen anything like that from youtube before).


Looks pretty smooth to me... I'd give it >10fps for sure.


If you find quad-copters doing cool stuff check out also this video: http://www.engadget.com/2010/03/18/quad-copter-responds-to-y...


I love the sound it makes when the rotors speed up to catch the copter after a maneuver.


doesn't seem that crazy when you see someone doing this kind of stuff without computer assistance: http://www.youtube.com/watch?v=-zzhriAN9Bc


With respect to the skill displayed in that video, it's different from this research. The video you linked doesn't have the same precision displayed in the research; I don't think they'd be able to go from hovering, through a window, back to hovering as precisely as was shown in with the single copter craft.

The single copter craft does turns, flips and dances all around, but it's over a much larger area and nor is the human pilot trying to accomplish a particular task. I think he's just having fun, free-form style. The key difference, I think, is that a single copter craft can't change direction as quickly as the quad copter craft.


I do not dismiss this research: it is amazing. But, I also think you are underestimating the skill and sophistication of RC helicopter hobbyists. There are international competitions involving very complex, but well-defined and precise maneuvers.

For example, there is a maneuver named "chaos" in which the pilot continuously pirouettes while continuously flipping while continuously rotating the axis of the flip while holding the aircraft in a single position a few feet off the ground. It is tremendously difficult to even visualize this fully.

The research video also reminded me of this video, where an RC pilot lands on a vertical wall (without velcro): http://www.youtube.com/watch?v=lZ2HlePbD20#t=1m07s


That was an amazing stunt. Has something like that been used as a plot device in an action movie yet?


There was a stunt at the end of the Italian Job (the new one, of course) where a stunt pilot actually flew a full-size helicopter around inside a parking garage. Can't find the clip right now, probably DMCA'd :-/


A human has a really good vision system. We can easily differentiate the window (for example) from the rest of the environment without problem. Also, we have a pretty good 3D system, and with practice we can process the relative position of objects and use this processed information to control the helicopter/quadrotor.

In this project they are using Vicon to determine the position of the quadrotor and the window. Once that information is processed, the quadrotor is properly controlled and has good enough "feedback", the human control can no longer be a match for the precision of a computer.


Some useful research, in PhD field. Hope he keeps this one, up. Though the thought is scary if you have to confront one of these things armed with a tazer or some such nefarious immobilizing device.


Is... is it wrong to get turned on by this?

My uncertainty about what I'd actually do with one is equaled only by my irrational conviction that I nevertheless need one.


At this stage, I wonder if there is there a way to move these little beasts with cameras and weapons without making them too big for windows though...


This is cool.

But it looks like it has too many degrees of freedom to make such fast/complicated manoeuvres steered by human.


That is what they need us (ie, programmers) for. Layers of programming jobs for a militarized swarm of these things. From twitch decisions to overall unit formation and attach plans. These would also be useful as bomb delivery devices. Think about it flying straight down, either reversed propeller spin or (more likely) upside down flight to avoid AA fire then sneaking through a window with plastic/gel explosives to take out a general hiding in an otherwise public building. Could be used for micro biological/poison attacks, like anthrax to take out key targets, or as cheap satellites for the 3rd world to spy on their people.

Of course civilian uses are vast as well, but harder to find really high priced jobs in the private industry.


Terrifying, but so damn awesome!


Isn't that thinking about the control on a very low level? Why do you need to know about the degrees of freedom -- if you're able to say exactly where you want the device to be, do you need any more details about the path?

Sure - the fun factor of the complete control is gone, but the usability goes through the roof.


I just like to feel in control of machine, here it wouldn't be possible.

Besides - the more direct control is the more predictable behaviour of this vehicle can be.

I don't think every possible route between 2 points will be possible even using these vehicles. So steering this will probably look more like RTS - I pick control points, unit chooses the path.

This can cause some pain.


How many degrees of freedom do you have in the location of your hand with your elbow bent? Back, forwards, up, down, left, right, rotation in any direction around the wrist - it's pretty full.


Interesting as a media-grabbing stunt, but not terribly useful. There might be applications in reconnaissance for emergency services or police, but that market is not very significant. Really roboticists need to be focussing their efforts on solving problems which will allow the industry to grow and break out from its traditional niches.


Nice companions for that little dog-like robot.

The military applications are kind of creepy


But can they interface with Skynet?


I soon as I saw the video I recalled Terminator 3. http://www.youtube.com/watch?v=jpYJg5obUkg


awesome. the first out-of-control robot swarm is gonna be so freakin scary.


If you see me walking around outside with a fishing net draped over me, well, now you know why.


Watching that video gave me one of those moments when you think you are seeing the future. Something from science fiction suddenly brought into the real world. Amazing achievement from an engineering standpoint. Kudos to the guys who made it happen!


DOD comment: took the words right out of my fingertips...




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: