First post on the fourms
Probably 4th real project using Arduinos but by far the largest in scale.
I designed an Arduino controlled shoulder assembly that fits around a human arm. The concept is closer to the Elysium exo-skeleton with Mat Damon than anything else. Are there any linear actuators that would be small enough (around 6" retracted), strong enough (at this point 20-40lbs of force will do for testing), and fast enough to not restrict movement completely (30mm/s at least hopefully)? I've spent a lot of time looking and from what I've seen I can usually get 2/3 categories checked off if I'm willing to spend minimum $150 a piece, which I reluctantly am.
I'm using 4 actuators per shoulder, 3 to control a point in XY&Z much like a 3D printer with a 4th to mostly fix rotation front and back while using its length to abduct the arm.
Any advice on both design and parts would be greatly appreciated.
Thanks,
Duke
To answer your question first, no, I don't know of a linear actuator which would meet your criteria. Second, I'm thinking you're encountering what everyone who first looks at exoskeletons encounters, the problem that off the shelf electromagnetic devices aren't comparable to natural muscle when it comes to power/weight ratio. The HAL exoskeleton uses some kind of electromagnetic motor and seems to have decent performance, so you might look into that. But all of the DIY exoskeleton projects that I've seen which use EM devices move at a snail's pace, which would indicate they've given up speed for torque.
You've said you want to use 3 actuators to control a point in xyz, but is that actually what you want? That would be an exoskeleton where the wearer is kind of a passive passenger. Generally with an exoskeleton you're measuring the user's intent to exert force in a given direction, not where it will end up. So you use force sensors or electromyographic sensors or even muscle hardness sensors. The user and not the system determines where things end up. The exception to this might be someone with a spinal cord injury and my apologies if my assumption was incorrect.
Right now pneumatic actuators seem to be most popular for DIY exoskeletons. These might be conventional pneumatic cylinders or air muscles. A problem here is that proportional control valves are more expensive than the linear actuators you mentioned. I've heard people mention that they were going to try to use a PWM signal to a conventional valve, but I have no idea how that worked out.
Who knew creating an exoskeleton to give yourself super strength would be so hard? I will try to beat everyone else here telling you to first just look at augmenting one joint, generally this is just one side of the joint (assisting the biceps muscle and not the triceps). This will let you look at the whole chain from sensor to actuator and evaluate. So at this point you're not heavily committed to a particular technology.
Hope this has been helpful. Good luck!
Seems like I'm not the only one having this issue. Any actuator that would allow at least sloth speed isn't strong enough by a long shot.
The 3 actuators would essentially allow the ability to roll your shoulder taking the place of some of the duties of the pectorals, traps, and rhomboids. The 4th actuator with a bracket that would only allow rotation on the lateral end in one way would take the place of the deltoids. 4 force sensors would determine the direction the shoulder was moving and adjust the lengths of the actuators accordingly. The good thing about the 3 actuators is that the range (min-max length) required to get a full range of motion is very limited. The 4th actuator (allows for abduction) is the only one that truly needs the speed and power and would be controlled primarily by force sensors just above the elbow.
Is my best bet just to use 2 weaker actuators to take the place of 1 imaginary strong light and fast one? I would rather not carry a battery and an air pump plus I'm less than proficient with pneumatic systems.
I've done a bit of testing on a powered gauntlet that uses a wormgear motor and a cable system that is controlled by 2 force sensors on the middle finger just to work through having a motor augment grip strength using sensory feedback. I'll put up photos of it at some point.
I don't know that there is a good solution with EM actuators. To some extent, pneumatics are just moving weight from a number of distributed points to one central point, but it also allows a lot of flexibility, you can fill a pressure bottle and leave the compressor at home, you could drive the compressor with a piston engine or even a gas turbine to improve power/weight ratio. But I'm not really sure I should be advising you, as my own experience with exoskeletons is limited to what I've put on paper. I'm still stuck working on the actuator problem.
Pneumatic actuators are probably a poor idea for several reasons. Hydraulic actuators might be an option if the electric linear actuators can't be used. Have you built an unpowered working mockup of the shoulder joints to see how the joints will move?
I've made a mockup with thermo plastic that worked relatively well and pointed out some placement issues. How difficult/expensive would a hydraulic system be?