Below-the-knee prosthesis

Hi all. I am designing / prototyping a below-the-knee prosthesis for my final-year project in college. I am keeping my design repository at [indigo.uk.to] (It's a bit of a mess at the moment - and there is still a lot to do). The plan is for an active prosthesis with some form of human control (electromyography or mechano-myography ideally - but I'm not sure if that will be viable). I will be printing the required parts using my Reprap Prusa Mendel i1.

• Software tests so far have been done an oLinuXino iMX233 Nano.
• The mechanical parts are being designed in FreeCAD / OpenSCAD.

I am experimenting with a bespoke linear actuator design (because those commercially-available which could provide the required torque seemed either very expensive or the wrong shape) - the design files for which are here and images of which will be uploaded here. I am hoping to print initial iterations of some of the more crucial parts over the next few days (so will hopefully have a demo of some sort).
The idea is inspired partly by the Open Hand project.

Part of the reason I mention it here is because I am thinking of setting up a fundraiser / bounty for it. I've been thinking about starting my own "business" when I graduate (I put business in quotation marks because I'm not commercially-minded - but need to have some sort of income). This is one of several ideas I have.

I have provisionally licensed the design under CC-BY-NC-SA - i.e. there is a non-commercial clause in there. I might reconsider a more liberal license at some stage.

Feedback welcome. Is this something a bounty / fundraiser might work for?

---

---

Semi-commercial engineering | Reprap diary

Here are some photos (not all the parts are fully printed yet). That's a potentiometer mounted on the ankle joint - to provide feedback on the ankle rotation [1]. I have proportional control working but it's pretty crude (I haven't tuned the gain, for example). Ideally it would be controlled eventually by an MMG (mechanomyographic) signal.

I am using a plain DC motor rather than a servo - as the potentiometer is providing feedback and might be adequate. I'd love to experiment with shape memory alloys but time is against me and resources are limited. (And I'm not sure they would have sufficient torque to move even the toes!)

It only has one degree of freedom. I'd like to implement more, but only have a few more weeks to finish this so don't expect I'm likely to have time.

Design files and software (only test programs so far) are available in the repository (link in previous post).




Feedback welcome. I'll post some videos if time permits.

[1] I have to credit my project supervisor for that. I was thinking of using a relatively complicated flexion sensor based on a strain gauge - the potentiometer is much simpler.

Edited 1 time(s). Last edit at 04/12/2014 04:41PM by dslc.

Check out this website to browse other cool products: [www.medicalexpo.com]

I have updated the repository for this - link.




Report (including Laplace-domain analysis of actuator-ankle-joint complex): [indigo.uk.to]
(There are a few shortcomings of the analysis there, but it's a starting point for controller design, simulation and evaluation at least).

Presentation slides: [indigo.uk.to]
It includes results of simple closed-loop control experiments using different controller configurations (proportional, proportional-derivative and on/off control) - although the results should be taken with a grain of salt because the experimental setup was not rigorous.

More pictures attached.