3D Cable Robot Uses The Building As Its Exoskeleton.

There’s not much information about this commercial product, but it looks so interesting, we just had to share it. It doesn’t seem there is anything too magical happening here: some motors (presumably some type of servo or stepper with positioning feedback) some cables and pulleys, and an end effector of your choosing. Oh, and just some clever math to solve the inverse kinematics – not that inverse kinematics is all that easy! You can see the robot at work in the video after the break.

Most likely you’ve already seen the end results of such a three-dimensional cable driven system on your TV. If you’re a fan of most field sports, the SkyCam system is what’s used to deliver the stunning aerial shots that really put you into the game. We’ve covered this sort of mechanism before, but only in two dimensions. Usually we see the concept used as a white-board plotter like this extremely methodical Polargraph or one built with K’NEX.

We can’t help but wonder how this might be adapted into other situations?  Perhaps, you could use small light-weight cables (fishing line) and pulleys to make a living-room beer delivery system or TV remote retrieval claw?  Or could it become the mechanics of a really large format 3D printer? If any of you do rig up some sort of house-hold beverage fetching robot, be sure to let us know via the tipline.

28 thoughts on “3D Cable Robot Uses The Building As Its Exoskeleton.

    1. Yes, but a crane that can be extended in size by more or less just adding some length of cable. Useful when working with large, light structures as you do when building aircraft. I can see that this would be useful in some cases and I did not know that this exists. So, thanks for the post.

      1. The trouble is the longer each cable becomes the more stout the anchors must be. Once you pass 120* between the gondola and the anchors, 100% of the load is passed to the anchor. So depending on the building (assuming you’re using it’s frame as anchors) you may cause a very dangerous scenario.

        Not technologically insurmountable, but it may become cheaper to just install a more conventional gantry crane.

      2. The lifting capacity diminishes rapidly the longer the cables become.

        It takes an infinite amount of tension on the cables to pull them straight with a load in the middle, which is why all rope bridges sag, so the amount of force required to actually lift anything increases exponentially towards infinity the further away the anchors are.

        That puts a practical size limit on the system. Because of the exponential relationship between cable tension and lift, it may be just a couple meters difference between lifting a ton, and lifting none.

    2. As an industrial fabricator, I could really see something like this kicking major tail in high volume high mix shops, even if you didn’t combine it with a robotic welding system. A few of these with overlapping areas could really change the game on assembling complex fabs of almost any size or shape. It would certainly make fitting less onerous.

  1. This doesn’t require the inverse kinematics. Just basic algebra. Even with the sag of the cables, it is pretty straightforward.

    Determining location and orientation from the cable lengths requires inverse kinematics, and is an interesting problem amenable to a number of approaches, depending on whether you have a known motion history or are starting cold.

  2. Make one to operate a pattern torch. There’s a few videos on youtube but none show the pattern above the torch. They have a variable speed motor above the torch head with the shaft in line with the torch tip. On the shaft is a knurled, magnetic roller, usually 1/4″ diameter. The steel patterns are mounted onto arms that hold them over the worktable. The patterns have to be 1/8″ smaller than the desired cut out. The arm is built like a SCARA but without motors at the two joints.

    To CNC one, attach the four cables to rings which fit over the drive motor shaft. Build a frame around the top, with stepper or servo motor and cable drum at each corner.

    Then there’s the software. That needs to be written. Or perhaps a driver/plugin/module/whatever for Linux CNC.

    There are a lot of these pattern torches out there, often cheap after a shop has upgraded to a CNC gantry plasma cutter. They are also simple to build. Some high strength cable and angle iron are much cheaper than long gear racks or screws, and a couple of pivot joints should be easier to get level than a large gantry table.

  3. I drove forklift for years, I can’t see how that thing could operate in a fifty foot warehouse with five tiers. Three cable suspension cant dip down into a two hundred foot long row that is only six feet wide. Perhaps it would be good somewhere very specific like moving nuclear fuel rods in a reactor or explosives you don’t stack but a day to day commercial warehouse? I can’t see it. Still very cool.

  4. Hi all.
    I just discovered your post and nice interest in this technology!

    I’m leading the robotics department of Tecnalia, where we are developping and promoting this Cable-based robotics technology.
    Tecnalia is a Spanish Private, Non profit, Applied research center whos aim is to develop and mature technologies before transferring them to our clients. The cable based technology has been developped in our french subsidiary, in close cooperation with the CNRS-LIRMM public research laboratory.

    As an example, Tecnalia is now transfering this cable-based technologies to small and large companies in various industrial sectors: from traditional crane manufacturers, to nuclear field specialists, via civil-works companies…
    Don’t get too close to the Techology demonstrator showcased in this video… The applications are yours to invent! and we’ll help you make sure that the required technologies are available and applied correctly!

    So, this technology (cable-based robotics) implies lots of software and control to manage both the actuation redundancy and the overconstrained aspect. 8 actuators for 6 degrees fo freedom – it thus allows to have an autonomous crane where you control both position and orientation.

    Some comments also point out that the higher the load, the higher torque required on the motors. Yes this is one of the limitations. But we have developped specific tools and simulators to allow to design the cable robot specifically for this operation… So of course, it’s not good for eveything, but still for a very large number of operations.

    Some other comments point out to the Fraunhoffer IPA research center. We know them very well and have some collaborations with them, through European research projects.

    As Peterthinks comments, the various cables would interfere with the traditionnal shelves in a warehouse… unless you bring new ideas and ways to use this techno to this specific case, and accept some limitations… other times it just won’t be the correct techno for this application…
    As for any new technology, you need to find the good match between technology push and market need/Return on investment…

    If you have brilliant ideas (and willingness to invest and develop your product), don’t hesitate to get in touch with Tecnalia:

    Thanks for interest!!


  5. I did some work with a winch robot around 2005. We used it to tell a floor covered in small wireless sensors what the location of each sensor was by broadcasting position as the robot moved over the sensors. So a winch robot might be of use for interacting with distributed sensors on a factory floor. Research paper is here:


    Video is here (the one labeled Crane Robot):


Leave a Reply

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.