Finger Actuation Technologies

Would this be used as a prosthesis or do you have rehabilitation in mind?

Size would have a different importance with usage.

How about pneumatic?

It would be used mainly as an assist device to move paralyzed fingers. Rehab may be a second option. In any case, it would fit over/on the fingers in some way to actively move them into a grasping pose.

What's your idea with pneumatic?

Perhaps something along the lines of what Butcher is suggesting. Pneumatic bladders are in use as conformal industrial robotic grippers.

The reasons being weight, size, noise and power dissipation. All this being remote to the hand.

Bladders have an elastic behaviour, like a balloon, and can conform to any shape.

Most cylinders have striction (mix of variable seal force, dynamic and static friction vs pressure) that makes them accumulate energy and release it abruptly. This behaviour may evolve over time and temperature as the seals wear. They also don't conform too well to human shape.

My experience with bladders them has been positive in exerting a controlled and continuous force to delicate devices. Well designed, they can have a lifespan in the 10⁵-10⁶ cycles or better.

With the recent developments in sensors and micro mechanics, there is a definite trend towards servo-pneumatics.

Cheap sensors are available to provide force, tactile and position data.

I am confident this would be a workable technology.

I know there is much advance in the ways to interface active prosthetics to one's will, but is is not my field of expertise.

Again, would this require tubing to run from the device to where a reservoir is located? On which side of the hand would you put the bladders? It sounds like you are suggesting placing them on the volar side to take advantage of their elastic behavior, correct?

I also intend to use force feedback for closed loop control. One cheap force sensor I've used in the past are the flexiforce sensors from Tekscan. I believe they have a design which includes an array. I figure I'll need to have sensors all along the finger, and not just at the tip.

There are also air muscles. Ever used those?

The idea as stated in your post #7 to house the reservoir and compressor in the wheelchair is the way to go for discretion, and noise reduction.

I would bring the valves close to or in the glove, so one tube is required. The tube/manifold can be considered and optimized a reservoir.

Same philosophy for the control, to interface many valves and sensors together, if some microcontrollers are placed in the glove, you need power lines and serial communication. This can be accomplished with 2-3 wires.

As a hand is partially closed at rest, the glove should promote this posture, weather using leaf, SMA or composite springs or the bladder shape, if you need zero power force or not, is to be experimented.

If I understand well, the volar side is the palm side. If possible the actuators should be on the other side, or in between fingers, to keep as much hand conformance as possible.

I have not used Air Muscles, but this is a good bladder functioning principle.

Tekscan have a good grip on the technology. A plus is the effects of bending and stretching are probably well characterized.

I would also suggest www.cui.com for sensors. Cheap as hell and available as samples or at DigiKey. Their electrolytic technology was initially developed for robotic artificial skin. They have a rubbery, gel like texture. I am not familiar with their QTC line.

Some 15 years ago I had to repair an array of 175 similar sensors used for corset study. I wished then the had not used individual sensors, with crimped and soldered wires, sown onto a mat...

I found this company for finger pressure sensors.

I need a device to help move my fingers, hand and arm.Any suggestion on who I could call to talk about this? sandyjb45@aol.com

I'm not aware of anything that integrates motion for the fingers, hand, and arm. For elbow rehabilitation you can check out Myomo. For finger and shoulder rehabilitation, check out Kinetic Muscles, Inc. Our hospital also has a completely passive device that supports the weight of the arm during grasping. Look at the last project here. Depending on your needs and resources, any of these may be a good start.

Let me know if you need any more help.

Check out this video. http://www.ted.com/talks/view/id/82

This is a full prosthetic arm, but the hand section should interest you. Note the recognition that you only need 3 fingers to grab something, instead of trying to fully duplicate a very complex mechanism.

I like the pneumatics idea, it would let you keep any heavy parts away from the hand and could be quiet. Maybe a glove that uses thin strips of spring steel or strong elastic to close the hand and give holding force, then pneumatics to stiffen a thin bladder along the top of each finger to straighten out? Might not work great, but could be an interesting thing to play with.

That sounds like a neat project, keep us (or at least me) up to date on your progress!

Thanks for posting the video link. Dean Kamen is a busy guy! I believe there are disparate design issues when discussing prosthetics vs. assist devices. There's just not a lot of room to put stuff if the hand is still there! If you read any of the comments after the video, I agree with the first poster that the technology to build a working prosthetic hand is just not affordable. There's a group where I work who built a 15 DOF prosthetic hand as their contribution to a DARPA directive that uses custom-made, proprietary motors. In my opinion, no amputee outside of the military will ever get to own one. However, touch bionics has built a prosthetic hand that uses only one actuator per finger. I like this idea, and it touches on the point you made regarding the number of fingers required to grasp an object. I was thinking to try to do something similar since the last two finger joints are coupled. During grasping, you don't need control of each joint. This reduces the need for either a complex mechanism or a complex control system.

I like your spring steel idea, but how is it actuated? Same question about the elastic - how is it actuated? My current prototype has a cable that runs on the volar side of the finger to flex it, and then uses an elastic band on the dorsal side for passive extension. It only extends the last joint, however. Here's a picture of it:

The spring steel seems similar to the idea of shape memory alloys (SMAs) which could both flex and extended the fingers. In addition they could perform both functions from the back of the hand, and, therefore, not interfere with grasping. SMAs are being used in prosthetic hands, but there are concerns about heat dissipation, the energy requirements, and the actuation time.

My lab does much research in stroke rehabilitation, and we have a pneumatic glove with the bladders on the palmer side of the hand to extend the fingers. In this case, would a large reservoir be needed. My initial thought was to house as much stuff as I could on the user's wheelchair, but this would require long runs of tubing, in the case of pneumatic, or cable, in the case of DC-motors.

I want something inconspicuous enough that the user wouldn't mind being seen in a restaurant with it. I'd also like it to be fairly modular, so it could fit quickly on different sized hands. In addition, they would need assistance donning the device, so it should be easy to put on and take off.

I'm not trying to box anyone in with some of my comments. I appreciate everyone's thoughts and ideas. That's why I put it out here. They are good, and keep 'em coming! I'll keep everyone posted.

I was picturing the elastic/spring steel to be placed on the palm side of the fingers so that the hand would be closed or gripping passively. The active part of the movement would be the opening, which you might be able to do with a pneumatic bladder on the back side of each finger. That picture looks like the same idea I was getting at, but with the passive vs activated positions switched.

I haven't really thought about this kind of thing much, but I would expect that it would be safer, or at least more reliable, to have the grip force be passive. That way a failure doesn't cause you to drop something or crush something. Unfortunately, that means your grip force would not be adjustable on the fly.

To clarify on the spring steel, I was thinking it would work like a constant force spring (clock spring) to curl the finger in, and would only be 1/4 to 1/2 the width of the finger. I suppose that would work on either side of a finger. I don't know if a bladder would work better to extend a finger by pulling on the back side or pushing on the palm side.

SMA's work very nicely in well controlled ways. If you want this for someone to use in everyday life, including outside, I might be concerned with reliability at different temperatures.

I think piezoelectric actuators are, by nature, small displacement devices and mechanically amplifying the motion would result in a bulky/heavy mechanism.

How about small, NeFeB linear motors, as opposed to electromechanical solenoids? Force would be linear and controllable over the mechanical range.

Some piezo motors are in fact quite compact:

http://en.wikipedia.org/wiki/Inchworm_motor

http://www.nanomotion.com/

http://www.newscaletech.com/

The force may not be the strong point for these configurations.

Also some solenoids have a quasi linear force/displacement response.

Hi Bill,

I believe you commented pretty extensively on a question I posted awhile back regarding a tmote device. I never got around to thanking you. So, thank you!

Can you tell me more about NeFeB motors? Starting with what NeFeB stands for?

Neodymium-iron-boron magnet. (should be NbFeB, Ne is neon). They are (I think) the highest strength commercially available magnets. Using that type of material, you can get a higher torque for a given motor size. I've never used them, so I can't be of much help on the details.

Ah, ok. A neon motor sounded a little far-fetched to me.

I never considered passive flexion. In that case, can it be varied, to accommodate different sized objects?

I doubt you could, at least in a way that wasn't overly complicated. You might be able to use the pneumatic (or whatever) opener to lower the grip force, but I don't see a simple way to increase the force.

Passive gripping probably wouldn't be the way to go for something looking to really duplicate the flexibility of use of a hand. The real time force adjustment that you are probably going for by adding feedback would be impractical to say the least.

The user could do some basic things like vary the number of fingers they grabbed something with to limit the grip force.

It sounds like using a passive grip might not be the way to go in the long run to really get a good replacement, but it might allow fairly reliable function in a cheap or early model.

That would be NdFeB . . . Specifically, Nd₂Fe₁₄B. Wikipedia has good info.

I didn't think Ne etc. looked right when I typed it, but it was early . . .

I don't understand the connection. How would these magnets be used?

Used in a linear motor such as:

http://www.copleycontrols.com/motion/motors/ServoTube/

http://www.aerotech.com/products/motors/linear_motors.html

The stronger the magnet, the more powerful the motor.

The magnets would not be used independently. NdFeB magnets could be used in linear motors to miniaturize the size and weight of a device similar to what is shown in your post #7 so that it is not objectionable to use. (I have envisioned such a device strapped to the arm of the person.)

Force Engineering has a blurb on linear motors, although they don't mention the advantage of using NdFeB magnets. Wikipedia also has info here and here.

Tape or wire extensions of the linear motor moving part can be laid over the back of the finger and attached to a flexible thimble-like device stuck on the end of the fingers to cause them to extend and retract. For people who have limited use of their hands, one tape or wire extension (actuator) could be used to create a signal to make all other extensions grasp or extend. For people who have very little or no use of their fingers, signals can be generated to grasp or extend by moving a foot, for example. Force (or current for permanent magnet DC motors) feedback can limit the amount of force applied.

I'll be glad to go into more detail if you wish. (I actually worked on things like this many years ago because my mother-in-law had rheumatoid arthritis and had very little strength in her hands and fingers.)

How about flexinol wire. Applying small currents causes it to shrink in length.

Here is a link to a FAQ page with strength and current needed info. they can be used in parallel to increase strength

http://www.mondotronics.com/mwfaq.shtml

and here is a source for buying small quantities

http://www.jameco.com/Jameco/catalogs/c273/P243.pdf

Mature SMA actuators:

http://www.migamotors.com/Products.html

How about stimulating the existing muscles?

I'm assuming it's the nerve pathway that is broken.

This may be 'safer'? As applying mechanical force could damage joints?

There are loads of thes TENs machines ... muscle stimulation thingies sold as sliming/toning aids.. could be easy to play/experiment with.

Yes, you're right, and I'm considering that too. There are issues associated with transmitting through the skin, and some people don't like them because a "novice" would be responsible for getting the stimulator pad in the right place --- stimulate the wrong muscle, you get a different movement than you were expecting.

I like them because they are reconfigurable; to do something new, all you have to do is move the pads around. I also like them because I'm squeamish , and don't like to implant things, like electrodes, in the body.

For compactness about the hand, maybe use two opposing stainless wires, about .010 or .012 inch dia. These could be guided by passing through tiny holes in a plastic band, one fitting band between each joint. The band at the finger tips would terminate the wires. Pulling on one wire while simultaneously slacking up on the other would create the torque to curl or open the fingers. The finger joints would be in compression.

These wire pairs (one for each finger) could be guided to a set of double-acting cylinders placed just above the wrist. (A couple of 10-hole bands would be needed to guide these wires across the back of the hand.) Each cylinder would operate a wire pair in "push/pull fashion. These 5 cylinders could be each about 1 inch long and maybe .25 inside dia. at 20 psi they could output about 1 lb to a digit.

The operating fluid could be either air or water. A reservoir could be located somewhere on the person's body, and the pressure could be developed either by an electrically driven pump or by a pump designed to be driven by some still-functioning part of the person's body ...say foot, leg or arm motions.

For dexterity and fine motor control, use separate miniature ball screws on the inside of the wrist with straps going to each finger tip and pivot points at the knuckles. (think brake cable on a bicycle.) On the back of each finger have a stretchable material to help open the fingers. Bi-directional control will be controlled by the motor driven ball screws on the wrist. And the power is during the closing motion. Also wrist control may be achieved in this method.

The thumb should be the opposite with the motor on the back of the wrist. and the control cable going between the first two fingers.

We have a similar design to open the fingers of stroke survivors, except we used a DC-motor instead of a ball screw. Are you aware of ball screws small and strong enough to fit on the wrist? One reason we went with the DC-motor was because we couldn't find any linear motors that could provide the required force and still fit on the forearm.

Another issue with mounting motors, ball screws, etc. is that they tend to slide around a lot, and the main solution to this point has been to strap them down as tightly as possible. Any ideas?

In stead of strapping them down tight which will affect circulation, Have stiff strapping going to a point like the elbow, shoulder or the opposite side of the neck.

Another option is to have stiff plastic (or other material ie bowling wrist support) that has been formed to the wrist yet only goes part way around (then buckled together). The stiff plastic will have pressure points but will also have points where pressure is less to allow circulation. Thus circulation is not strangled.

The use of a ball screw with the DC motor will amplify the power of the DC motor (more turns less movement)

try

http://www.npa.nsk.com/public/enu/info_1001.asp

Update for those interested:

So far I've been able to embed 2 wire ropes, one for each of the first two digits, along the glove. I've used a longer piece of elastic that runs proximal from the metarcarpal joint to the finger cap. This worked ok on a healthy hand, but the elastic tended to slip off the middle knuckle. So, I stitched the elastic at the end of the glove's finger hole. Of course, this increased the force at the latter section which caused my last finger joint to hyperextend. I'm still looking for a linear actuator drive the ropes. I'm also going to use the ropes to aid in extension. I'm not sure how hard I'll be able to push on them without them "bow stringing."

On a paralyzed hand, however, things didn't go so well. It is very difficult to put on the glove since the fingers are flexed so much to begin with. I tried making a little shoe horn device, but there wasn't enough room inside the glove to push the last section into extension.

So, still plenty of things to work on. Here's what it looks like so far:

It looks to be coming along nicely. I have just two comments at this time:

1) a possible guide for the "ropes" might be Teflon Tubing. There are different formulations giving different stiffness and abrasion resistance, but all will allow easy sliding. If the "ropes" are only used in tension, they could be made of very small diameter stainless wire ...say .008 inch. You can find small ID Teflon tubing fairly easily, and if you need to go smaller, just stretch the tubing until you get the size you want. It will retain that shape indefinetly, unless heated considerably.

2) an alternative to moving the stitching up on the elastic band, to prevent slipping sideways might be to pass it through a sewn-in rectangular loop. Or you might cut two slots in the glove and pass the elastic in one and out the other.

Ray,

Do you mean to use the tubing as guides on the forearm portion? I'm pretty sure the ropes will have to do some pushing. Their fingers are very tight and I don't think elastic alone will be enough to extend them. I may have to think about adding an additional spring element.

As you might notice from the pictures, I didn't sew the material too well. The ropes were supposed to come out perpendicular to the wrist. I'm trying to borrow a linear motor to see how well I can pull on the rope.

I like your ideas for the elastic. I wish I had a better way to connect/mount all this stuff so I could try different ideas faster. I was originally thinking about using a glue gun to glue some tubing to pass the rope through. I don't know if I'd be able to remove the tubing and clean the glove. Anyway, I went with stitching the fabric. Which wasn't easy by the way. If it were up to me, we'd all still be wearing fig leaves!

This seemed to be the best for hiding the ropes as much as possible.

I was thinking of using tubing as guides from the point of actuation, all the way back to the motive driver. If you also want to Push with the wires, just use a larger diameter ...maybe .011 to .018 inch.

Here are some links to the wires. Notice a wide selection of sizes and materials:

http://www.smallparts.com/products/descriptions/swx%2ECFM

http://www.smallparts.com/products/descriptions/stt%2Ecfm

Sewing to secure the tubing or wires may turn out to be the best because it allows the fabric of the Glove to flex. I had a project where I applied a bead of "grip" material to gloves. It created significant stiffness because it locked the stitchwork together. This, even though the bead material was only about 55 A durometer.

Sewing loops to guide the elastic may work out good too.

My initial thought was to attach tubing to the glove and run the ropes through it. However, as you mentioned, I was concerned about adding stiffness to the system. One thought was using corrugated tubing. However, since the ropes run between the fabric and the glove's surface smoothly enough, I saw no reason to add tubing.

It is nice that the fabric flexes, but it's also unfortunate because I'm looking for some place to mount force sensors as well. My boss has been developing some small, 3-axis force sensors which will fit on a fingertip. I hate to have to put anything on the finger though.

I hadn't thought about the trouble it would be to get your glove on a paralyzed hand...

Would it be an option to make a mitten instead of a glove, as an easier starting point? Use all 4 fingers together against the thumb, then maybe move to using 2 fingers together rather than 4 to effectively have 2 large fingers and a thumb (like the ninja turtles or the 'live long and prosper' hand gesture :) ) A mitten would probably be much easier to work with.

Starting with less elaborate movement might help the size constraint problems you seem to be running into and let you get a lower level of functionality to people much sooner.

Another way that might make getting the glove on easier would be to put a small zipper up the entire pinky and/or thumb side to increase the open area you are working the hand into without increasing the size of the glove.

Taking a sewing class might help too

Butcher,

I hadn't either. The major problem is just that their joints are so tight, the fingers tend to be quite flexed. The therapists I work with say there's no good way to don gloves for this reason.

The mitten is an interesting idea. Similar to what you've suggested, my next steps are going to be to try to open up the glove like a clam shell by cutting the seams on the pinky side and also around each finger. They would be reattached with velcro. Many of the gloves we use for stroke survivors have a zipper that runs from the pinky side of the wrist up and across the palm. However, they usually have at least some ability to extend their finger joints.

I think what we're going after right now is minimalistic. We're just trying to do 3 grasps, and 2 might even be more reasonable, and trying to keep complexity down to a minimum by using a simple actuation. The reason for that is, as you mentioned, to get something useful to the population as soon and as cheaply as possible.

Hi DomerDoc

I have also seen a need for this type of thearpy and have discovered that a lot of other people have been working on this issue both inside the body and outside of the body.

A couple of issues that you will discover is called stiction or friction of what ever type of actuation your are using, the second is user safety for the skin and also not ripping off of digits or fingers with over powered actuation devices.

Good luck and I hope you create something that helps the human race.