Close
0%
0%

Origaime (Cardware-bis)

An educational system designed to bring AI and complex
robotics into the home and school on a budget.

Similar projects worth following
Cardware is an educational system designed to bring AI and complex robotics into the home and school on a budget. It leverages cheap off-the-shelf electronics and kinematics - eg Raspberry Pi Zero and hobby servos plus a cheap webcam - and templated cardboard parts.

Origaime is a base technology; it enables many others including Cardware because it uses planar geometry to make precise 3 dimensional shapes. Cardware is an exoshell system for a specific robot but Origaime includes forks to use the system in other ways.

All work contained in this project is subject to Creative Commons Licensing.

The Concept

Modern robotics hardware for the hobby market has been reduced to a series of modules, from servos to processors. The only things connecting these together are wire and chassis, and I've been experimenting with using easy to obtain materials to build those chassis. Polycarbonate and polystyrene, polyethylene, sheet aluminium etc were natural choices.

Mark's IO made me think again about the validity of cardboard as a working material, and we began a dialogue that culminated in this collaboration.

The idea was to use a single sheet of craft card and a printed template to create chassis units that connect the electromechanical building blocks together.

We have since iterated over a few changes to the design, incorporating better geometry to give the parts strength, room for cabling and better overall appearance.

First generation limb built by Mark N during the initial prototyping.


Origaime - The next generation

The updated parts built in exactly the same way as the first set, they fold up from a single piece of cardboard to make a modular piece for the robot. Mark calls this Origaime, after the original robot that inspired everything and the art of folding. :-)

Other MaterialsWe've
looked at other materials besides cardboard after realising that
self-recycled materials were also a possibility. The initial discussions
for the project touched on the fact that paper and card were already a
recycled material besides their other qualities, but the inclusion of @markgeo
to our collaborators brought this to the front quite literally as he
attempted a build of the early generation limb parts using cardboard
scavenged from cereal packets. This was necessary as he is developing
Cardware in Thailand where card stock and other things we take for
granted are not so easy to obtain. MarkG is using this ability of the
system to promote robotics in education there, as other resources needed
to teach the subject are not easily available.

Other ways of using the materials available

Exoskeletons are one way of using planar geometry to build 3 dimensional shapes. Planes can also be applied equally to laminar technology to make solid forms.

Once the cardboard has soaked up a generous amount of PVA glue it becomes a composite more in keeping with building technologies like MDF. Laminar technologies are also extremely strong and accurate, using averaging to spread loads and imperfections in their design to achieve military precision.

Origarmy

Extending a finger in the direction of solid forms. Repeated accuracy has always been a problem with hand built machinery. Those days are gone with laminates.

Not just fingers either, the AIME hand mechanism was deliberately scalable even though it never really saw the light of day. Inspired by Mark's #Uglytech - The cardboard relay I went back to the drawing board for the limb parts and thought about hybridisation.

Still a printed template but uses much crappier materials. MarkGeo highlighted that good cardstock is not always available and Cardware Exoshells require a narrow window of weights to work properly due to the card being too thick and stiff for the small tabs, and problems like de-lamination crop up as well.

I looked at how Mark Nesselhaus was stacking the material and realised that garbage corrugated card was actually fine, just real ugly and required special considerations.

It cuts out cleanly and quickly with a craft knife. You can be lazy about this, providing the holes for the bearing pins are punched accurately, you could hack the rest out with a butter knife and it would still work. This is a deliberate response to the immediate reaction to Cardware - because it looks pretty, it must take monumental skills to build - which is in error. Yes, there is technique, I have built it into the paradigm so it teaches you as you go. Providing you dont jump in with the hardest part of all - Hip Yokes.

These are still a pain to construct, but are easier laminated.

The Sensor...

Read more »

  • 1 × Corrugated Card Rough as you like, it doesnt have to be good. Just dry and undamaged.
  • 3 × Servos Standard mount 9G hobby servos
  • 1 × Foil Tape Or plain foil
  • 1 × Sticky-back plastic I've used clear, I didnt want love-hearts all over my robot
  • 1 × Pack of small screws For the servo lid.

View all 9 components

  • Installation

    Morning.Star05/06/2017 at 11:10 0 comments

    Now all the infrastructure is built, I can start putting the systems together. First thing is run all the power and signal cables a bit more neatly so they dont interfere with the mechanisms.

    Plenty of room to bring cables down through the neck joint.

    Thats a pair of H-Bridges I built ages ago to experiment with. The resistors on the left are a voltage divider for a series of bump-switches, they arent being used.

    Both the motors are wired into the bridges for testing.

    Add the neck and eye servos, tidy up all the wiring - make sure that all the grounds are commoned and add a strap for the lid. All it needs now is a host computer and a power supply for the motors and its ready to roll.

    Wire in the cameras and its ready for testing - once I find some decent axles. I'll have to glue these into the track hubs and put a circlip on each to stop them from coming back out again as I really dont want to use those kebab sticks...

  • Skunkworks

    Morning.Star05/05/2017 at 14:28 0 comments

    Yeah I know, I've been howling a little about lack of funds and therefore lack of hardware. Good news, I have some servos on the way! Another 8 x 9MGS, none of the plastic rubbish that caused me most of the grief.

    In the meantime, I still need a mobile platform to test the 3D vision system with. I've been looking at a tracked base because I have all the components laying around :

    A pair of 1:70 gearboxes and 5v DC motors. This one is a piece of a kit but this mechanism could easily have been chopped out of a toy. It doesnt have any way of detecting the position. More on that later...

    There is something raw about grabbing materials and a rule and prototyping out of your head. It leaves CAD completely cold.

    Those kebab skewers are temporary, I'm trying to find something a little more substantial...

    Interface it to the existing Origaime hardware and I'm now nearly ready to proceed where I left off due to the servos.

    Life imitates art, a Wall-E made of trash :-)

    The neck pitches and turns, the eyes focus and it will move on its tracks when I get the motors wired in.

    Because they are DC motors I'll need to find a way to sense their position for basic metrics. There is no room in the gearbox for a magnet near the motor output shaft or any other way to read the motor revolutions. I could rip up an old mouse and use the encoders, but I've come up with an interesting alternative.

    As the motor has a commutator it throws out a lot of EMF and needs suppressing. While I'm in there, I may as well sense the back-EMF using one of the MCU's analogue inputs and use that to determine rotation. I havent tried this yet, but I'm optimistic.

View all 2 project logs

Enjoy this project?

Share

Discussions

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates