An affordable Exoskeleton Arm (ExoArm), that will help elderly, disabled people and workers complete everyday tasks with less exhaustion.
Already have an account? Log in.
To make the experience fit your profile, pick a username and tell us what interests you.
Update that is also included in Project Logs, but I'll put it here non the less.
First prototype is complete!
How to make it:
Measurements are not defined because until the arm is "universal size" the lengths vary from person to person. So the best act will be that you measure your own arm and use those measurements.
A schematics. I never got to make a PCB, but that doesn't mean you shouldn't. It's actually very easy, as you can see that there are no resistors, diodes etc.
For components and links go to PARTS LIST.
Code? Well if I knew how to make a good program that runs it.. I spoke to a friend of mine that is extremely good in electronics and programming, and he said that the most optimal solution to intuitive control would be to insert strain gauge in your bones. But in this world that is not possible and probably never will be. So there's that.
See, the problem is that the ExoArm must read the weight of object you lift before it even tries to assist you in lifting. And that is so hard. So hard. Weight changes depending on the angle of fore arm, how your arm is angled in respect to earth, etc. I'm not saying it's impossible, but it would really take some time to make this happen.
SO if you are looking to make one yourself, and are not that good in programming you should just use 2 buttons. It's astronomically easier.
MUSCLE SENSOR:
Short answer. No. Why? Because.
Just kidding. I tried using it and it did not work as I thought it will. Not even a bit. See the problem is that when you put the electrode on your biceps, you can trigger it just by stretching your skin. it causes so much electricity that it false triggers and device thinks that you send a muscle signal. So when the Exo-Arm is fastened to your arm, the belt will stretch your skin, especially at the biceps. What happens? Exo-Arm flexes, but does not let go, because it thinks you're still sending signals to the biceps when the only signal is your stretched skin, if that makes any sense.
Where to put all electronics?
I spend so much time trying to come up with a decent place. Perhaps a good starting point would be a little "backpack thingy" that sits on your shoulder or maybe even a small backpack. See the problem are batteries. They are rather heavy and make things difficult. We must also keep in mind that this should be as low cost as possible, so no fancy ExoBacks or similar things.
This are all the wooden and aluminum parts you'll need, minus the mounts that hold wood plates (check out the videos). Please keep in mind that there are no holes. But why Kristjan? Because I'm 99.9999% sure that every motor you'll find has different geometry, every arm is different and it goes on. So in order to make this arm, you'll need to be creative. You can follow my instructions as far as you can, but I can promise you that you'll have to figure things on your own too. But hey! That's all part of experience and that is the most valuable thing in life next to health and knowledge.
Here is a little test with PID control. As you can see it can hold a position pretty well. The PID is not tuned. That's why you can hear that beeping sound coming from the motor.
Now all I have to do is to join PID control with intuitive movement, so that I can actually lift something.
This video is without any regulator/power:
Here you can see PID regulator in action. It is designed so that it holds a setpoint (angle), and tries to do that no matter whether there is a force on it or not)
Let's start off with the MOTOR. I'm using windshield wiper motor, but it is a stupid idea. Well not stupid but a beginners mistake. It's heavy. So heavy in fact that it pulls the whole ExoArm from your arm. I used it because I had ideas of using this ExoArm to lift 10kg and more. What I did not account for was that this ExoArm needs back support if one would wish to lift this kind of loads. But then it wouldn't cost $100 but more like $200.
So if you are going to make this arm, I suggest you try the motor that lifts and lowers the car glass window. It's lighter, it has lower torque, but for our applications it's good enough.
Something like this: https://www.ebay.com/itm/Power-Window-Motor-Front-Left-ACI-MAXAIR-83098-fits-98-11-Lincoln-Town-Car-/253215691421?epid=75818760&hash=item3af4d4d29d:g:DrgAAOSwno1ZpYq8&vxp=mtr
But I strongly suggest that you visit a local car junkyard and buy one there. Who would spend $50 on a new motor if you can get used one for $20.
Second, how to fix your arm to the ExoArm. If you look at my design, you will see that the only supports that I used are on the bottom and on the side (wood plates and foam). I encourage you that you place a support on the upper part as well. It will make your arm fit tighter to the ExoArm and therefore make it more intuitive.
I'm back and it's two days until the deadline. Which is funny. This summer was hectic and I honestly didn't have much time to sit behind computer and stare in the code while sipping 6th cup of coffee. What little time I had I've been experimenting with arm and I'm getting better and better responses from it. I also resorted to 3D printing because I'm at collage and don't have access to a workshop. All the parts will be designed in a way that they can also be made from wood or any other material.
To keep this short, in the time of making this project I came to few conclusions and what should be done to make this ExoArm better. I will make another Project Log where I'll talk about that.
Decided to make the stand little more functional.
Not much going on at the moment regarding the ExoArm. I'm still throwing my time out of the windows trying to make it work with a weight attached to it and I can say with all of my heart that it's a lot harder than it might look.
Stay tuned for updates, it is bound to happen anytime now! :)
The idea turned to reality! It actually works, and I'm super hyped!
Next thing is to make it useful for lifting heavier things and to somehow detect when something is suspended on it.
The SG (strain gauge) is mounted on the forearm. I placed 2.5kg weight on the arm, the SG did the job and everyone is happy forever. Well not really. Not yet. When I applied pressure upwards, the sensor value decreased, which is a good sign so big props to the theory from june 20 log. The battle is on!
Next step is to create a holder for the arm, so that I can further test it (putting it on and off from my arm is agonizing especially if I do it 40 times a day).
I'm still waiting to receive the strain gauge that will be placed in the middle of fore arm. It will be used to determine if there is any object suspended on the ExoArm.
So far my idea is as following (it's just the idea and may have big flaws, but until I receive the part I can't be sure)
- There is no weight on the ExoArm:
This one is simple. Strain gauge will deflect when my arm is moved up or down, so controlling this movement shouldn't be so hard.
- An object is suspended on the ExoArm:
The strain gauge sensor will detect a mass of the object, therefore we know the force and also torque. Torque changes with the angle (theta) of the object but if we say that the upper arm is parallel to the body, then the Phi angle is the same as -Theta angle.
So now we have a good estimate of the torque. Here is the part where I lose my "insight" into the mechanics. If an object is suspended on the ExoArm when I try to move my arm up, the torque on the ExoArm changes. And if torque changes, the output of strain gauge also changes, therefore the ExoArm will follow your arm depending whether the strain gauge output it bigger or smaller. So this could maybe be an effective way to control the ExoArm. I'm still thinking and trying to come up with a 100% effective way, but at the moment this is as furthest as I am.
Would love to hear your ideas and suggestions!
Thanks! It's good to hear that this project gives ideas and inspirations to other people :)
Don't forget safety features, you don't want it to break someones arm when a sensor or bad wiring gives faulty data.
It's not said in the videos directly, but if you look at the junction of fore arm, upper arm and where they are fastened to the motor, you can see that the upper arm limits extention of the fore arm. (Video).
Have you tried using a force gauge for controlling the arm instead? Here's an example:
You could put this in between your arm and the device, so when you push up with your arm, the arm follows. You could base the speed of the arm on how hard you push on the sensor.
I do foresee potential problems with getting good contact between your arm and sensor, and with pushing the sensor hard enough to register a change. So you could use a long and wide "lever" that would multiply the force on the gauge, while also increasing the area you can push on. Similar to how a microswitch works.
I nearly used this kind of control method, but then I realized; what about when there is weight on the exoskeleton? The weight of the object would pull down, and trigger the sensor that is on the upper side. I guess I could try to compensate this in the code, but after some thinking I came to a conclusion that it would be better to use strain gauge and play with torque on the ExoArm. I will make a video/post soon because I just received this strain gauge sensor and will try to use it. So stay tuned! :)
You actually want the motor to get more current and therefore output more torque the heavier the load is, correct? If you think about it, the motors just amplify the force by your arm, so it would make sense to supply them more power when you are carrying something heavy. I think the body would intuitively adapt to this as well.
As far as triggering the sensors too soon, all you would need to do is tune the cutoff level at which you don't want the motors to come on at all, and the ratio of force to output voltage or PWM duty cycle. In fact that looks strangely like y=mx+b... turns out math is useful :).
I can't respond to your latest comment, so I'll reply here (I hope you see this).
Yes, you just described PID control :) But this still doesn't solve the flex/extend arm problem.
Also, I have a bad feeling that torque to current plot is more of exponential nature then linear as you said. But for my goal, I probably can assume that it's in fact linear. Will do some testing in these days so I'll keep you updated!
I don't think it would be too hard to compensate, but it's up to you :).
Can you add a cherry picker... I learned 3 pulleys two on ceiling and 1 on floor divides the tensions so that makes lifting easier. Maybe keep that idea in mind I wish we made pulleys like this tiny.cc/faradayrim
Kristjan Berce
WalkerDev
Karl Bielefeldt
Dan DWRobotics
YodaLogic
Beautyful idea. I was searching around for some Input. Im a Nurse and i am about to build a similar device for a Patient. Great Inspiration.Ty<3