back to article Family of technician slain by factory robot sues everyone involved

The family of a repair technician killed in an auto parts factory accident is suing five robotics companies they say are responsible. In a suit [PDF] filed to the Western Michigan US District Court this week, the family of Wanda Holbrook claims that the companies that built, installed, and maintained the robotics at a trailer …

Page:

  1. Pen-y-gors

    Trying to see both sides here...

    Sounds like the rules are all a bit vague and the legislation is a bit dodgy, but...

    1) Depending on the type of maintenance being done, could it be necessary to have some power to the unit? This is complex electronics, not a Ford Fiesta that needs an oil change.

    2) The deceased was a repair technician, so presumably fully trained and aware of how the robot works. Were all approved procedures followed?

    1. Stuart Castle Silver badge

      Re: Trying to see both sides here...

      "1) Depending on the type of maintenance being done, could it be necessary to have some power to the unit? This is complex electronics, not a Ford Fiesta that needs an oil change."

      I suspect it could. However, I suspect that with most robots, it's possible to test the electronics with the mechanical parts disabled. You can test various circuits for expected outputs with a multimeter for example. I would also suspect it's possible to control the mechanical parts directly.

      That said, I am not a Robot engineer. The only "experience" I have with Robots is seeing them on TV, reading about them (in publications like this) and controlling a few motors with a Raspberry Pi, so I could very well be wrong.

      "2) The deceased was a repair technician, so presumably fully trained and aware of how the robot works. Were all approved procedures followed?"

      You would assume that anyone who even gets near to doing something as potentially dangerous as maintaining a robot would be fully trained, and have access to any protective gear and equipment needed. Sadly, a lot of companies don't go beyond what is legally required, and some don't even go that far.

  2. Anonymous Coward
    Anonymous Coward

    Hydraulics

    If the power supply of a unit is disconnected, are there still hydraulic pistons under pressure? Can hydraulic pistons in such a unit extend/contract, even if no power supply is connected, in case a valve fails?

    1. Anonymous Coward
      Anonymous Coward

      Re: Hydraulics

      "Can hydraulic pistons in such a unit extend/contract, even if no power supply is connected, in case a valve fails?"

      In general, yes, though specifics here are unclear. In general on hydraulic lifting equipment etc there is often (should always be?) a warning along the lines of "do not work under unsupported platform", with the same basic meaning as "use axle stands as well as a hydraulic jack".

      Is there an equivalent of Health and Safety Executive in the USA? It would appear not :(

      Condolences etc to the victim's family and friends.

  3. Mage Silver badge

    Sensor?

    A few dollars per moving part fits sensors (even if just load sense on actuators) to "see" is something going to be in the way.

    Seems like saving a few hundred dollars, in once off cost, is more important than operators or maintenance workers lives.

    ANY automated system should stop if something that shouldn't be there is there. You don't even need software, though that makes it easier and cheaper.

    I made a prototype machine with a rotating drum with an aperture. The prototype had a "knife" sharp edge and wouldn't even nip a finger tip, it would reverse as soon as the travel was restricted. It could cut pencils in half without the sensor.

    So yes, sue everyone, this is a completely avoidable death, even if was a late 1930s automated machine.

    1. Anonymous Coward
      Anonymous Coward

      Re: Sensor?

      You greatly overestimate the ability of sensors.

      A reflective surface can throw off most optical sensors for example...

      Think about the Tesla incident where their autopilot sensor-system overlooked a large truck in front of the car and steered the Tesla right through it.

      To outfit every single robotic arm with a comprehensive list of sensors (to avoid such scenarios)

      would not only be prohibitively expensive, it would also dramatically decrease the productivity of that robotic arm, as all of those sensors have to be checked by a processor/program and need to be able to completely interrupt operation (so basically one major check and many minor checks per update cycle)

      AC cos I'm talking out of my ass.

      1. John Brown (no body) Silver badge

        Re: Sensor?

        "dramatically decrease the productivity of that robotic arm, as all of those sensors have to be checked by a processor/program"

        CPU power is cheap these days. So are most sensors. These sort of accidents have been preventable almost since the inception of automated production lines and the means of doing so is cheaper now than it's ever been. Any skimping on safety is almost always done be bean counters at the time of purchase/installation and who likely won't be there when the actual costs come home to roost.

      2. Mage Silver badge

        Re: Sensor?

        I'd not rely on an optical sensor.

        1. TRT Silver badge

          Re: Sensor?

          I used to work a guillotine. Had a mesh of light beams above the work area, and you could only move the blade if you pressed two buttons under the table which were about 4 feet apart and had one foot on the clamp control. It was still possible to chop an appendage off if you had two people in the control area, so there was an exclusion zone painted on the floor where you were not supposed to operate the machine if there was anyone else stood in that painted bit. But there was no electrical or mechanical interlock on THAT part of the safe working method, and if the blade was NOT in the top position, the mechanical interlock wasn't engaged, so the cutting bar could still drop under its own weight, something that it was supposed to do when you had to change the blade. You really, REALLY, learned respect for the machine when you knew the forces involved in it. It could happily drive itself clean through a broom handle under its own weight if the blade was freshly ground, under power it would go through 6 reams of card stock in under half a second without even a change in pitch from the motor. And that was 1960s technology, let alone what modern machinery is like.

          A modern (1980s) addition to the unit meant you could programme a cutting sequence into it. The backstop would move to the next position automatically after a cut, and all the operator had to do was make sure the stock was squared up to the backstop. Hands on the buttons, foot on the clamp then BAM! another cut. A really practised operator working it was like a ballet. Could cut a whole stack down to the finished trim in under a minute easily.

    2. Anonymous Coward
      Anonymous Coward

      Re: Sensor?

      All components involved in safety related functions have to be safety rated. These are special components with special expensive price.

      Consider a motor spinning some blades chopping salad or whatever. The layman's approach is to enclose it in a cage, add a switch to a door. The switch is placed electrically so that it cuts power to the contactor that energizes the motor.

      Motor 300 bucks, contactor 50 bucks, switch 10 bucks. 360.

      However, since this is a dangerous machine, you need to use safety rated components. The switch Costs now 100 bucks. You need two contactos in series in case one of them fails. You need a 500 buck safety relay that evaluates that the switch is working properly, evaluates whether any of the contactor have failed, and also evaluates whether itself has failed. Failure should be safe and result in the safety relay denying the system from running.

      Our system cost just went from 360 bucks to 1000 bucks.

      For this reason it's common to enclose very large areas behind a single safety circuit. This results in larger areas than strictly necessary to lose power whenever an operator needs to access a small part. This unfortunately leads to crafty maintenance engineers devising ways to bypass safety systems and rely on just switching off the relevant motor from a control panel or PC or what have you..

      It gets even worse if you want some actual "smart" safety functions. A safety rated PLC needs to be able to survive abuse and still able to shut down itself if its input or output circuitry gets damaged by surges or overloads, or the operator forcefully inserting a knife through its circuitry (happened at my workplace when the system refused to run with a vague safety-related error message).

      It needs to have two redundant CPUs,from different manufacturers but running the same program.

      So no, it's not a few dollars. A smart safety system costs many times more than the robot itself...

  4. kwhitefoot

    Forty years ago I was a test technician at Emerson Electric in Swindon. When I discovered a fault in the UPS or variable speed drive that I was testing I immediately padlocked off the main supply breaker and called the foreman of the department that would fix the problem. He padlocked the breaker. When the wireman came to fix the wiring fault he padlocked the breaker.

    To power up the machine again all three of us had to agree that it was safe and unlock the breaker.

    Surely the same rules apply today?

    1. Sean o' bhaile na gleann

      I was thinking of something like that myself.

      Surely it must be possible to make a padlock that can itself be padlocked?

      Engineer 1 has to do some work in a hazardous area, & so 'safes' things by padlocking the power supply (or whatever the working energy source is).

      Engineer 2 needs to do some related work, & so padlocks the padlock that is already in place

      Engineer 3 padlocks the padlocked padlock

      etc etc

      Engineer 1 finishes work & needs to test the result... sorry guy, you'll have to sit and wait until 2 & 3 have finished, thus allowing you to unlock your own padlock.

      Or am I preaching to the choir, here? .

      1. Anonymous Coward
        Anonymous Coward

        "Surely it must be possible to make a padlock that can itself be padlocked?

        [...]

        Or am I preaching to the choir, here? ."

        Preaching to the choir, probably The issue in this picture isn't that the required safety technology doesn't exist - it's existed for decades, as have works procedures to use it safely and productively.

        It's that someone in this picture chose not to use the appropriate safety mechanism and process, in circumstances where it seems likely to have been applicable.

        US-centric related article:

        https://en.wikipedia.org/wiki/Lockout-tagout

        1. Yet Another Anonymous coward Silver badge

          Not always possible.

          1, You are teaching the robot a new sequence of moves so the power needs to be on

          2, You are calibrating the touch points so you need to drive the robot to those points

          3, You are changing the working head and so need to be able to move the arm

          In this case it seems that it wasn't the robot they were working on but one in the next station that was triggered to move an assembly.

      2. LeahroyNake

        A T shaped metal bar that has to be pulled through a slot, several holes it for as many padlocks as you need. Simple design that could allow every worker to carry their own padlock and key.

    2. thames

      @kwhitefoot - "Surely the same rules apply today?"

      Having rules and following rules are not necessarily the same thing.

      I won't jump to any conclusions in this particular case, as I don't know what happened. I do know that the chances of getting a complete and balanced description of what happened out of one side's or the other's lawyer is pretty slim.

      Instead I'll describe a case that also in an automotive plant in Michigan that happened a number of years ago. A maintenance person was working in a large robot cell. When you are talking about making something the size of a car, you are talking about machines that you can walk around in. When he was finished, he closed the safety gate, re-enabled power, and restarted the production line. However, he later discovered that he had left one of his tools inside. Rather than admitting his mistake, he got a ladder and climbed over the guard and jumped down inside to retrieve the tool. One of the robots then moved in a manner he didn't expect and crushed him against part of the machine.

      In the case described in the story, I'll wait for the results of a safety investigation before I jump to any conclusions. Generally though, it is the responsibility of the person working in the machine to follow the appropriate procedures to render the machine safe before entering it. If the machine cannot be rendered safe to enter, it should not be entered. Companies such as Fanuc and Nachi have sold many thousands of robots for decades, this is not science fiction technology. Locking out robot 130 using the lockout on the front of the robot control cabinet would have rendered it safe. The real question is why this safety measure was not applied before the machine was entered.

      1. DropBear

        " The real question is why this safety measure was not applied before the machine was entered."

        I think you already answered that - it's always the same thing... peer pressure. In your case, the expectation of not stopping the restarted line again just to retrieve a tool. In the article's case, who knows... what I do know though is that proper procedures are never convenient to properly follow, and the expectation is always, always there for you to just get something done (that proper procedure might make impossible or much delayed) regardless. It's always very hard to point blank refuse such expectations based on nothing but safety concerns - even though one always should. It's not hard to imagine a young, inexperienced technician may have had a much harder time resisting that pressure.

  5. Anonymous Coward
    Anonymous Coward

    time for Asimov's Laws of Robotics

    to be made mandatory.

    The safety case here would never have been approved in many industries in this Country.

    Having been around control systems including high power Radars not locking the effing things down or not being able rto lock them down are huge great warning flags.

    No one repeat No one should be made to work inside these cages without safety interlocks.

    One Radar system I worked on had keys with huge red tags attached. The tags had to be visible on the operator's body ay all times. Failure to display the tags was cause to get fired on the spot.

    If the Radar powered while you were in the wrong place up you would get fried in less than a minute.

  6. Anonymous Coward
    Anonymous Coward

    Just

    Get Trump to shout at it

  7. Anonymous Coward
    Anonymous Coward

    Isolation 101

    Sounds like endemic failure on the company's part.

    1. Jason Bloomberg Silver badge
      Unhappy

      Re: Isolation 101

      Sounds like endemic failure on the company's part.

      Yup. And that is the one entity the victim's family are prevented from suing, the one who is exempted from blame by state law. Absolutely appalling; a travesty of justice.

      I sincerely hope we don't get that sort of crap once we have brexited. But I fear that we eventually will.

  8. jonnycando
    Mushroom

    Alternatively...

    Death to all robots....hire people to replace them.

  9. razorfishsl

    And this is why Asimov 's three laws of robotics are complete bollox.

    1. Dave 126 Silver badge

      For sure, today's CNC milling machines are ignorant of what constitutes a human. There are circular saws that retract their blade if they detect a level of capacitence - that being one way to distinguish between a human and a piece of timber.

  10. jrchips

    Many automated industrial systems have 'kill switches' (or e-stops) that allow a human being to freeze the operation of the equipment. Typically 'kill switches' are in readily accessible places and spaced in such a way that an operator can reach one quickly. Now the EU is considering making such 'kill switches' mandatory on robots.

    So does that mean that all fully autonomous vehicles should have 'kill switches' on them reachable by someone on the outside? One on each corner perhaps?

    1. Charles 9

      And then the complaints will start flying when those kill switches trigger spontaneously...

  11. The Real Tony Smith
    FAIL

    Elysium

    https://www.youtube.com/watch?v=qR5Ptfyi9tQ

Page:

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Other stories you might like