- Feature Articles
- CodeSOD
- Error'd
- Forums
-
Other Articles
- Random Article
- Other Series
- Alex's Soapbox
- Announcements
- Best of…
- Best of Email
- Best of the Sidebar
- Bring Your Own Code
- Coded Smorgasbord
- Mandatory Fun Day
- Off Topic
- Representative Line
- News Roundup
- Editor's Soapbox
- Software on the Rocks
- Souvenir Potpourri
- Sponsor Post
- Tales from the Interview
- The Daily WTF: Live
- Virtudyne
Admin
As a robotic technician, and industrial electrician, I have no doubt in my mind that this story is fictitious. There is really just no possible way this could ever happen, for many of the reasons already stated in the comments, as well as the description of how the robots acted in the story. It just doesn't work that way.
Admin
Admin
Admin
I think that this is the video that might illustrate the process: http://www.youtube.com/watch?v=e-r7z6JFwbk
Admin
Had a similar experience some years ago but way more lines and mostly C. It also appeared to work bug-free.
Admin
You're welcome.
Admin
It's not about size, it's about design. Industrial robots are generally designed to lift large weights, but not to move them especially fast (ie, 50+ MPH). Even unloaded, the robot probably couldn't move its arms that fast.
Admin
Admin
Ok, google for "dlr justin catches ball" you will see a human-sized robot built by German aerospace center catching balls thrown at him.
Now balls are round, and you only have to account for linear motion, add angular motion to it and problem gets harder. In case of an engine block, you would have to account for angular motion.
The story is bunk.
Admin
"as I'm sure you'll come to understand, the submitter wanted both them and their company to remain anonymous."
Keeping the company anonymous shouldn't be a problem since it is probably as fictitious as the story.
Admin
Standard conversion when picking 3rd party library:
Manager: So did you finish researching products that will generate PDFs for our reporting component? Developer: Yup, I found one from a legitimate company with rock solid reviews which constantly releases updates for their products. A server licence is $400. Manager: Oh, that's not too bad, anything else? Developer: Yeah, I found one created by some guy called Ken_1984 that was last updated in 2004 that is bundled with malware and written in VBS, but it's free... Manager: FREE YOU SAY?!?!?!
Then after the highly paid developer spends six weeks fighting with the POS API, they abandon the free version and end up buying the $400 licenced version.
Admin
I wish you'd been responsible for the PHP-spaghetti I inherited a few weeks ago, instead of the person who wrote it. I think even "Hello World" would have stretched their abilities...
Admin
Admin
I'll bet it had static classes in it!!! :P
Admin
This doesn't count all the other apps I've ever written that are stashed in source control somewhere and are no longer on my HDD.
But I agree with you partially: I am 10x cooler than you. But if you want way cooler stories than I have, read the sidebar. The regulars there always have something awesome to share.
Admin
Admin
Admin
Crazy Eddie - doing all the wrong things for all the right reasons!!!
But +1 to Amalie for talent-spotting...
Admin
Admin
To all the people who can't understand how the second robot can catch a heavy object that the first robot throws:
We assume that the second robot is at least as strong as the first robot, and that it is at least as high up as the first robot. It is therefore possible for the second robot to throw the object to the first object. By reversing time, you can then show that the second robot can catch the object thrown by the first robot, assuming that the first robot throws it along the trajectory that the second robot is expecting.
Admin
Admin
Horsefeathers.
The story has been "anonymized" because it is BS, not to protect someone's job.
Admin
I assume that what /actually/ happened was more likely a robot handing off the engine casing to another robot at high speed over a human passageway, or possibly robot A tossing the part towards robot B where it would land on the floor within arm's reach of robot B.
Both of those would account for clangs, danger and damage.
Admin
1- They may not be the engine blocks themselves, but they are still huge big heavy chunks of metal. 2- Owing to having the same shape as an engine, they are equally irregular objects that will tumble unpredictably when thrown through the air. 3- A big heavy piece of aluminium being hurled through the air would get seriously damaged when it lands, aluminium being such a soft metal, even if "caught". Odds of the engine block subsequently fitting inside it, and there not being any distortion causing leaks around all the seals? Tiny.
It's just not a believable WTF. It is a great story, but it's not a believable WTF.
Admin
First I think the throw was more like 10 or 15 feet. Also I think the item being thrown was part of the anonymizing. The part that best fits the description of the toss is a Valve Cover. They weight 8 to 10 pounds. They are not a precision part. They would reverberate like a bell when caught by a sudden force. There is plenty of flat area to grab. The shape symmetrical.
Admin
My most impressive was a lot shorter than 1200 lines, but it was Z80 machine-language (in hexadecimal) inserted in an MT-Pascal "inline" statement, pulling the opcodes out of memory with no reference manual handy after not having touched a Z80 for several years (because the compiler's inline-assembly statement would only let me use Intel mnemonics and I only knew the Zilog mnemonics, and I was using some Z80-only opcodes that weren't on an 8080 anyhow). So yeah, first-try bug-free happens.
Only problem was that it was at a client site, not at my own office, and there was nobody in the room to high-five.
(I stuck the Zilog assembly-language mnemonics in comments next to the hex machine code, in case anybody else had to maintain the program later. To do otherwise would have been irresponsible.)
Admin
Call me sexist, but I smelt a rat when the story started with a female doing preventative maintenance in a factory. Also, "the conveyors were buckled and the motors were exposed"? A conveyor motor should always be exposed. "Greybeard" changing robot code? I struggle to convince anyone to have a go at changing robot code even after plenty of training. Network+Robot+Vision System will never produce anything in under a month. I've enjoyed a few dubious stories on this site, but this one made me cringe. Thats all before you go into laws of physics and feasibility of throwing/catching (changing tool load data mid motion? does any brand of robot do that?). And yes, programming robots IS my day job.
Admin
Network+Robot+Vision System will never produce anything in under a month.
I dont agree to that, and i used to write code for different industrial robot systems, some with vision systems ( the first ABB Flexpicker, i think its branded different now , Adept SCARA arms, a bunch of smaller japanese brand cartesian robots etc. )
There were decent enough development tools to set up an industrial robotic system to do certain tasks plenty fast, at least from certain vendors. And i got out of that line of work years ago, i assume the tools have only improved.
There are other reasons why the story as described is bunk, as i posted above. Getting a big enough arm to toss an object is pretty easy, catching is the hard part. I have tossed a few grippers off the arms accidentally, thanking plexiglass safety cages.
Catching WITH AN ARM though while doing vision tracking, is very much at the bleeding edge of tech ( see DLR Justin, above ) and there is no way any standard industrial system would be able to perform that stunt with anything more than a soft ball and even then you would have to have a ton of bandwidth in the integrated system between vision tracking and servos. But like i said, tracking objects angular pose in in air accurately enough for the gripper to catch it in 3D is quite far in science-fiction-land.
What may have been the case indeed if the arm would toss the objects to a floor in a gripping area of another robot to pick up. But presumably, that would damage the objects.
Hence, i call bunk.
Admin
Most likely the "catching" robot didn't pick it out of the air, but caught it on some large flat surface (like a pallet). This would explain the loud noise when the casing was "caught."
Seems plausible to me. Although the distance may have been exaggerated.
Admin
ffs, the arms are perfectly capable of good throws, thats not the issue ( see http://www.youtube.com/watch?v=LAtdsDTt__s or look up KUKA Titan with max payload of a ton or so )
Catching is the issue.
Admin
Engine block implies cylinders are included... I also think watercooled.
Engine case implies cylinders are separate... I think small air-cooled
what could have been meant is a engine block CASTING (ie no finishing had been done on it, it was only just pored)
This and the fact that it was going to be sent back to the "Blast furnace" implies iron engine blocks of indeterminate size.
none of this matters because, as many have pointed out, its not easy to catch something when your a robot. Id like to point out that its not JUST the figuring out where it will be and when its also figuring out how to grab it.
Industrial robots dont have hands, they have end effectors and they are highly specialized.
Admin
I reread the article. it says "engine block casing" the first time the term pops up. It says "engine block" the second time.
Engine casings could be whole lot lighter and flyable than engine blocks themselves.
Admin
Quoted FTW Now we're having 'commented' comments as well as articles? Nice one :)
Admin
Blablabla impossible.
a) it's totally doable (pay me and I'll do it) b) production kit sucks compared to bleeding edge kit c) story overexagerated but quite fun
For all the people who think that's impossible ... seriously lol . you're prolly of the kind who think f-22 is high-tech for today or C&C-style ion cannons are impossible to make .
Tell you what, if Echelon was possible when it was made, the current technology is far beyond anything any of us here know about.
Admin
Some of the other threads are full of haters, thank you for your support.
Admin
Agreed on the +1 - however much the story is enhanced, it's still pretty clear the guy is overqualified for what he's doing.
Admin
after reding all coments, i reach concluson this is toy factory.
Admin
Pounds are weight, weight is a force, acceleration is also a force.
Accelerating a 200lb engine block from rest to a speed sufficient to fly 100' through atmo on a parabolic trajectory is probably pushing the limits or exceeding their load capacity, as well as their shock (rapid acceleration) limits.
Admin
Echelon wasn't one dude tinkering with a program involving math complex enough to warrant the creation of ENIAC. Not to mention robotics manipulation to exert those forces.
I remain highly skeptical.
Admin
Clearly someone is an MST3K fan...
Admin
I thought I was the only one who remembers the vindow viper...
Admin
So do I . the story is most likely bullshit/ massively overexagerated, but it's totally doable ;) Speaking of which, a bunch of people have been saying "blablabla craprobot catching ball" : well cool .. you know about ferromagnetism don'tcha ? ^^
Admin
Good one Made me chuckle
Admin
Admin
This is made of 100% pristine and infallible WIN!!!!&()^$#%.
Admin
Like a number of stories on dailywtf, I find this one extremely hard to believe. There are many problems with the story, including:
I'll only cover the first two; the others are similarly formidable problems.
Timing accuracy: If the robot doesn't close its grippers at the right moment, the block either falls out of open grippers or bounces off of closed grippers. Variation in network congestion is one of many variables that will significantly affect timing. Another two are the air pressure used to operate the gripper, and the temperature and viscosity of the oil that lubricates the pneumatic switch which actuates the grippers.
Also, robot computers are generally not running a hard real-time operating system. This would be a necessity for any operation requiring sub-millisecond timing.
Mechanical accuracy: Large, strong robots, such as those needed to pick up engine blocks, are considerably less precise than smaller robots. Consider the acceptable angular error for the throwing robot.
I'll assume that the catcher's grippers open one inch wider than the block (they almost certainly open less!), and the two robots are 100' (1200") apart. The acceptable angular range for the throwing robot can be computed as atan(1/1200), and is approximately 0.0477 degrees. This means that the robot's angular tolerance must be below +-0.0239 degrees or 66 parts per million. This is a very small tolerance, and if the block and arm weight aren't distributed perfectly symmetrically about the robot's axis of rotation, it will turn to the side a little as it lobs the block. Given this insane tolerance, it will also be necessary to take into account distortions in the robot, its base, and the factory floor, as the temperature rises and falls.
Admin
Admin
Now this sounds like a fun series of experiments: Programming robots to throw human bodies through the air!
We can reduce the requirements somewhat by using dwarfs. I understand this may soon be legal in Florida. http://www.globalpost.com/dispatches/globalpost-blogs/weird-wide-web/florida-legalize-dwarf-tossing-ban-ritch-workman-update
Admin
What regulations would OSHA impose? That man is the sole operator.
From the way that i see, the operator is automating a business process, keeping operating costs low and reducing risk by minimizing human error.
I think amaliee stole the operator's plans and got herself a nice promotion.
Admin
It was a cute story, but madmanguruman is right with all of the above.