Slashdot Mirror
Volkswagen Factory Worker Killed By a Robot
m.alessandrini writes: A worker at a Volkswagen factory in Germany has died, after a robot grabbed him and crushed him against a metal plate. This is perhaps the first severe accident of this kind in a western factory, and is sparking debate about who is responsible for the accident, the man who was servicing the robot beyond its protection cage, or the robot's hardware/software developers who didn't put enough safety checks. Will this distinction be more and more important in the future, when robots will be more widespread?
Really Slashdot editors....
Try 1979 at a Ford plant in Detroit.
http://nl.newsbank.com/nl-sear...
Really it is a new low when the editors on slashdot can not be bothered to use Google
.
See my blog http://ilovecookes.blogspot.com/ for light hearted technical information.
Are you kidding me? No, it is most certainly NOT the first severe accident with industrial robots. Seriously, thousands and thousands of factories using them, why in the hell would anybody think for a second that accidents had never before happened??? I guess the submitter is so sheltered that he has no clue at all about what it is like to do physical labor in a place that makes actual things!
The regular safety measures weren't in place because they were installing the systems, so most likely they had people working on different things and someone started testing their piece without realizing it was already connected.
Yep. See second paragraph.
If the robot must be moving (typically, when you're teaching the robot the path it should follow), then every single person in the workcell must have an active deadman switch (anyone lets go, the robot emergency-stops). And you run the program at 10% speed so that you have time to trip the deadman or get out of the way. The workcell itself is fenced off, usually with either a tripwire or electric-eye switch that will e-stop the robot if triggered.
I used to work for a robot company, and we enforced these rules religiously. When I went to visit plants and work on the robots, they issued me my own padlock and tags for lockout/tagout. Someone had to have skipped some safety procedures in this case.
Indeed, in most places, a bug where the system crashes is the most severe possible bug. When dealing with robots, that's only the second most severe. The most severe were "unexpected motion" bugs, where the robot didn't follow the path in the correct way or otherwise didn't behave predictably. Those got everybody's attention.
PHEM - party like it's 1997-2003!
C'mon folks. This is basic stuff when working with any hazardous machinery. This is entirely a human error, and the 'robot' aspect of it is unimportant. The word 'machinery' would have been less provoking. It's about the same as saying "Factory worker dies after jumping into industrial tire shredder with insecure controller hardware". The controller has nothing to do with it.
Granted, Europe doesn't have the same OSHA requirements as the US, but still, it's pretty obvious.
If you're not familiar with this concept, here's a summary & scenario.
Summary: You use a device to physically stop the operation of the machine that requires a lock, and then you keep the key to that lock with you so only you can re-enable the machine.
Situation: You need to rewire half a building. You shut down the power and lock the panel so no one can turn it on. You start work and now your hands are full of wire. At the same time, a co-worker's air compressor loses power because it's plugged into that downed grid, he comes over and wants to turn it back on, but since you have the only key, he can't. As a result, you stay alive. Alternatively, you don't lock the panel, and your co-worker electrocutes you.
ReiserFS
I can't believe I heard about this story on the radio this morning, with the radio hosts likening it to the movie Terminator. I work in industrial automation and let me assure you that these industrial robots have absolutely nothing even remotely approaching "AI". An industrial robot is no more than a multi-axis motion control system with some fancy co-ordinate transformation math on top of it. The programs are as simple as "wait for this input, then move to this point, turn on this output, wait for this input", etc.
When we're starting up any industrial automation workcell (whether it as a robot or not), the cell design has to be certified (stamped by a professional engineer in our jurisdiction) that the safety system meets appropriate regulations and is built with certified components, all of which are specified to specific safety requirements based on hazard, etc.
The thing is, those regulations are there to protect factory workers and people interacting with the cell in normal operations. If you take any machine apart using a wrench, you're supposed to be properly trained in how to lock out all sources of energy in the machine. That said, when you're programming the cell, you're allowed to be inside the cell and power up the robot using a teach pendant with a special enabling switch you have to hold down. This requires you to put the robot in a special teach mode which also limits the robot speed to less than 250 mm/s. If the cell was built correctly, the interlock switches on the gates have to be wired into the gate inputs on the robot, and when you open the guarding, the robot can only be energized while in teach mode with the teach pendant enabled.
The system isn't fool proof. We all know impatient people. Maybe the person programming the robot didn't check that the gate switches were wired in properly, or maybe he asked his buddy to close the gate behind him and press the reset button because he wanted to see what was going on (something I've seen several people do, and have always chastised them for). Maybe the guarding wasn't completely installed yet. Maybe he mistakenly put it in "Teach 2" mode which allows full speed operation with the teach pendant enabled. This mode is generally illegal in the United States, but some jurisdictions do allow it as long as you take other safeguards, like striping out a dedicated area on the floor where the robot can't reach where you're allowed to stand.
That's why this is most certainly human error. The question is, who is liable? Did a manager pressure the guy to continue programming the robot even though proper safeguards weren't in place? Did he just get impatient and ignore his own safety training? I see lots of people do that, and I also see lots of people with missing fingers - go figure.
"I have never let my schooling interfere with my education." - Mark Twain
What we call industrial "robots" really are just fancy remote control/programmed toys. They got slightly more smarts than a woodchipper. They follow a programmed dance --rather stupidly. If something is between them and the next step they go THRU it with 500-1000lbs of force.
In this case, Western is referring to Western culture rooted in Greece & Rome, spread throughout Europe, and was brought to the Americas. If you think about early people coming out of Africa into the Middle East, some of them went west toward Europe and some went east toward Asia, hence the Eastern and Western cultural labels. Here is a Wikipedia article: https://en.wikipedia.org/wiki/....
Not only the title is wrong. Also no grabbing was involved.
The sentence in the summary (and article) that the robot "grabbed him" appears to me as a non-native speaker's translation of this newspaper artikel. It says "Der Mann sei von dem Roboter erfasst und gegen eine Metallplatte gedrückt worden."
Yes "erfassen" can mean "to grab" (although one would normally just use "fassen" for that) but in this context it means "to hit and push". You will find lots of sentences were people were "erfasst" by a car and I think we can all agree that cars usually do not grab people.
So instead of a malicious robot grabbing his tormentor and throwing him against a wall, the poor guy probably was just caught between one of the joints of the robot and a metal plate when the respective part of the robots arm moved towards that plate.