Robot Arm Gives Kids The Roller Coaster Ride Of Their Lives

Unfortunately, [Dave Niewinski]’s kids are still too little to go on a real roller coaster. But they’re certainly big enough to be tossed around by this giant robot arm roller coaster simulator.

As to the question of why [Dave] has a Kuka KR 150 robot in his house, we prefer to leave that unasked and move forward. And apparently, this isn’t his first attempt at using the industrial robot as a motion simulator. That attempt revealed a few structural problems with the attachment between the rider’s chair and the robot’s wrist. After redesigning the frame with stouter metal and adding a small form-factor gaming PC and a curved monitor in front of the seat, [Dave] was ready to figure out how to make the arm simulate the motions of a roller coaster.

Now, if you ever thought the world would be a better place if only we had a roller coaster database complete with 4k 60 fps video captured from real coasters, you’re in luck. CoasterStats not only exists, but it also includes six-axis accelerometer data from real rides of coasters across Europe. That gave [Dave] the raw data he needed, but getting it translated into robot motions that simulate the feeling of the ride was a bit tricky. [Dave] goes into the physics of it all in the video below, but suffice it to say that the result is pretty cool.

More after the break.

Before anyone gets the urge to call Family Services and report [Dave], know that he seems to have taken great care not to build something that’ll turn the kids into jelly. He describes the safety systems in an earlier video, but the basics are laser light curtains to keep the arm within a small safe window, an e-stop switch, and limiting the acceleration to 1 g even when the real coaster would be giving its riders a good beating. That’s probably less than something like this real backyard coaster generates.

60 thoughts on “Robot Arm Gives Kids The Roller Coaster Ride Of Their Lives

    1. Even though I know these things are insanely reliable, with very mature safety technology and all the rest I still worry that he’s strapping a child into a machine capable of pulverising them in a split second.

      1. I’m not worried about the robot arm, I’m worried about the programming and the mechanical design of the harness.

        I don’t think I’d hesitate to choose this over many other things people seem to think are acceptably safe, but it’s still scary.

    1. They are quite affordable if you are looking for a used one. I got a Kuka KR16 with A KRC2 Controller from 2004 with 179h on it from a University nearby for 3k€. I only lack the time to implement all the interesting things like MIG 3D Printer, 6 Axis Plasmacutter or Robotic Sheet metal bending…

    1. Indeed, my first thought was, I’d like an adult sized version of that. I can just imagine that sort of thing in arcades of the future. Certainly cheaper and smaller than the military training simulators. I wonder if it could be modified to simulate F1 cars or fighter jets too.

    2. There is no 1g limit. He’s using gravity as the only acceleration force in the system and the arm just modulates it’s relative direction. That’s why it’s scaled to constant 1.

      1. While the predominant force is gravity, the arm is still moving and therefore still exerting its own accelerations on the rider. All of those are explicitly limited as well.

  1. … and when I was 10 years old, my dad bought a “Mighty Mouse” portable carnival roller coaster which had been sitting rusting in farmer’s field. Came apart and fit on 5 flat bed trucks. We spent a year rebuilding it, replacing rusted out parts, fabricating new parts, fixing broken welds. Dad taught me a lot of maker skills on that coaster. Eventually got it running and I was the most popular kid in the neighborhood! My mother however was NOT happy!

    1. There are active and passive safety systems. Most of these are active. If sensors are working correctly then the safety system will work correctly.

      When they fail; they can kill people in less than a heartbeat.

      If you’ve ever seen a video where a robotic arm failed you might understand how completely insane this is.

        1. But I would never trust myself to make a flawless enough implementation that I’d dare risk the life of others.

          You do understand that professional engineers design stuff that could potentially kill people literally all the time? That it’s pretty much their job description?

          Planes, bridges, industrial machinery… the list goes on.

          1. And yet the industrial robots are kept in wire cages, with complex safety systems and lockouts for good reasons. Certainly not because the professional engineers designing production lines don’t know what they are doing.

            This is insane, esp. in that small space – one programming error or failure – and his kid gets smashed against a wall like a bug.

        2. He and his wife can make another one, just like it.

          If I made a child catapult like this, I’d install physical end stops on every axis and a kid safety cage.

          That or a solenoid belt release, for maximum throw range.
          Of course I’d give the kid a parachute, I’m not insane.
          You could throw the kids to school, every morning.

      1. You do understand that that mechanisms are used in aircraft, cars, surgery, and any number of other life-critical functions?

        Right?

        And that PEs are the ones who are authorized to design and stamp the plans for such life-critical equipment?

        Right?

        1. You do understand that the critical failure of many robot arms is to go into a hard and repetitive jerking motion that doesnt stop until the power is disconnected and this guy is strapping his kids onto this thing. If a cars primary mode of failure was to ram into a wall over and over and over again you probably wouldnt fasten a car seat to the hood would you?

          1. “Hard repetitive jerking motion?” Why not wild spinning motion? Why not sudden and immediate return to origin (0,0,0)? Please. The guy knows more about it than you do.

    2. HaD article: something done with professional skills which is safe but looks scarier than it is

      HaD commenter: How dare this person even get out of bed in the morning, let alone do a thing? I’m incompetent and I have no capacity to assess this project but I can scream “what about the children” with the best of them!

        1. Boeing used to be full of competent engineers led by management that moved up through the ranks of engineers. Now, the MBA leaders have made their lives so difficult I think most of the competent engineers have moved out or taken early retirement. Thank you, McDonnell Douglas!

          1. I guess the only way to make Boeing healthy is to fine the company for such an enormous amount of money for the 400 or so passengers they killed that the whole company goes bankrupt. Then the curators can oust most of the management and use that money to restart the company. But it still would not be fair without some jail time for gross negligence for a lot of those people. But in the land of limitless greed that is all unlikely to happen.

    1. Those are COMPLETELY different machines!
      They have nothing in common with ‘normal’ industrial robots. Industrial machines like this are made to be cheap, and this is a cheap industrial arm and controller, who lacks any kind of safety needed for this kind of entertainment.
      And yes, i worked for 23 years with Robots (developing Software and some Mechanics) of many kinds. So, i feel qualified for an opinion.
      Still, not my kid, so meh.

      1. They’re both from Kuka, and to me they pretty much look the same, except that those in the amusement are two-seaters, and maybe one or two sizes bigger.

        It would be interesting to know what the difference of safety features would be between these robot arms from the same company.

        I know there are “safety PLC’s” which have forced relay contacts, and I think they have two (maybe more) cpu’s inside checking each other. Is there an equivalent for this for robot arms?

    2. Harry Potter and the Forbidden Journey in the US.

      Robotron on the MSC Seascape

      Epcot “Sum of all Thrills (Presented by Raytheon)” (2009-2016). Guests could “build” a virtual roller coaster, and then get strapped to the end of a Kuka arm to “ride” it.

    1. And did you notice how much free space is around that robot there?

      And how little space he has in his garage? If anything goes wrong and the robot arm stretches out for whatever reason (software bugs and hardware failures are a thing – this is not a triple redundant aerospace-rated system but an industrial factory robot), whoever is riding it will get smashed against the wall or ceiling like a bug before anyone could react.

      Still sounds like a good idea to you?

      1. You can be walking in front of your parked car and have it suddenly decide to roll forward and crush you. Or a plane could suddenly fall out of the sky directly on your head. Or your dog could decide to chew a battery and set your bed on fire.

        NOTHING is 100% safe, but things can be acceptably unlikely, and you’re being completely disingenuous if you’re claiming you know for certain this is riskier than being in a car going at highway speeds.

        1. If you worked in a factory with this robot arm and disabled the safety cage door sensor switch, they’d fire you as soon as they saw the video.

          If you were a PE and did that, they’d also report you to the engineering board.
          Even though they’d be less liable…’Missing wiener and fleshlight’…He’s the expert, wrote program, confused twist axes…Nobody to blame but himself.

          But no child shortage.
          I say he should proceed.

          There are robots legal to have around people uncaged.
          Basically they have to be stop-able and pushback-able by a normal size person.

          He want views too much.
          Will regret this.

        1. It would not be too difficult to install a completely separate system onto the robot arm itself that triggers an emergency shutdown if certain parameters would exceed some kind of safety limit. And that would not be visible in video’s like this.

  2. Interesting idea and fun to see it executed. The kids look like they’re having a blast on it.

    One thing I would suggest is using Quest VR headset with ‘Epic Roller Coasters’ It’s made by B4T games, and they have a complete telemetry setup for motion simulators where you can mimic the rollercoaster antics as you’re riding in the game (VR). yawvr, simtools, simracingstudio are all supported.

  3. This is a crazy, bordeline insane, thing to do in such a small space. It is rather shocking to me that the dad who built this is a professional engineer programming these things for a living – he really should know better.

    Such motion simulators based on robot arms are in use in practice, both for research (e.g. at Max Planck Institute in Tubingen, Germany: https://youtu.be/oMLarxR-q08) or in amusement parks – but you always have a ton of free space around the swinging arm so that it can’t smash into anything. And you need a ton of safety waivers for operating something like that.

    If the robot ever “goes nuts” (be it because of a programming error or some failure), the wide open space limits the number of things which it could crash into before the operator could intervene and hit the big red E-Stop button.

    That’s not a theoretical possibility of some “concerned carebear” wanting to spoil someone’s fun – there has been an incident like that in Tuebingen, despite the system being operated in a large hangar and with strict safety protocols in place. I believe something has failed on the robot and the person riding it got hurt. They had to rebuild the simulator because of that (or to even decommission it, don’t remember anymore – it has been years ago).

    Having the robot this close to the wall, where it is clear the arm can reach it and crash into it, with the only safety being some software motion limits preventing some joints from moving, would make me super uncomfortable – esp. when a child is being strapped to the wrist of the robot. If anything fails here, the kid could be dead before the dad even manages to realize something has gone wrong.

    There are some good reasons why these robots are normally only ever operated in wire cages, with interlocks and complex safety systems in factories, despite professional engineers designing and building those systems. Squishy humans and robot arms capable of lifting hundreds of kilos of weight don’t mix – and when they do, the results tend to be a literal bloody mess.

    Don’t do stupid.

    1. Did you watch his safety video? He has multiple light screens wired to the estop circuit (which is dual channel) to stop it in the event it get too close to the walls/floor/ceiling. He mentions he accounted for distance in his placement.

      I’m trying to find an article on your mentioning an incident in Tuebingen. I can’t find anything. When was this? Do you have a link?

    1. Hey, if it works, he’ll get to pass on his “good engineer” genes. If it turns his kid to paste on the garage floor, then we won’t have to worry about him passing on his “overconfident engineer” genes. Win/win. Let Darwin do his thing.

    1. Speaking as someone who operates a similar arm in a research setting. This would never pass a safety review. It relies on software to avoid smashing the child into a wall at several metres per second. Yes, I’ve watched the safety measures video.

Leave a Reply

Your email address will not be published. Required fields are marked *

Please be kind and respectful to help make the comments section excellent. (Comment Policy)

This site uses Akismet to reduce spam. Learn how your comment data is processed.