I wish to build a glove that recognizes when my (specifically my four) fingers are in certain positions. I need some sort of sensor that will identify when adjacent fingers are close together. I want a sensor (block) along sides of the fingers between each knuckle. So two on the inside of the pointer, the middle and ring would have two sensors per side, and the pinkie would have two more on the side that rests against the ring finger. I'd also like to be able to differentiate between left and right hands, so that I can tell if the right middle finger is pressed against the right pointer, or left pointer (or any finger on the left hand, I don't need this to be specific), or if the fingers are spread and there is no contact.

Any suggestions as to a type of sensor that matches my description? I need to know if I am pressed against one of two possible contacts or apart.

I've tried to do some googling, but I have not had luck with finding good keywords. I've found a glove that the designers say can translate sign language, but the parts I can read about the glove refer to sensors that are inherently more sophisticated than what I need.

Thanks!

 

http://en.m.wikipedia.org/wiki/Wired_glove

 

I'm sorry, I've read several wired glove articles and all are focused on a different problem. These solution are too complex - I dont need to know if fingers are bending or if the wrist is twisting.

Perhaps I should rephrase my problem. I need to find a "sensor" (for lack of a better word) that can identify if it is pressed against/close to another sensor.

I am looking for a solution similar to the window alarm sensor, which knows if the window is closed by proximity. I will keep searching google on my own, I was hoping someone here might have a suggestion based on experience/knowledge. I hope I didn't expect too much.

 

Perhaps a hall effect sensor and magnet can be used.

 

I need some sort of sensor that will identify when adjacent fingers are close together.

By "close" you mean close but not necessarily touching?

If that's the case, I can't think of much other than a magnetic field to "communicate" across the gap. I'm thinking that you could use small, matched coils to send and receive. By oscillating a signal on the transmitting coil, at a different frequency for each finger, you could later determine which receiving coil was getting a hit.

I think an oval shaped coil, like a race track, with the long axis perpendicular to the finger bone, would allow for some misalignment of the fingers. You would choose how much sensitivity you need by setting the threshold strength of the received signal.

You have other options if you're only detecting touch, not proximity.

 

Well, like I said in the first post, I'm interested in making a glove that can recognize specific hand positions based on whether the sensors are touching-close. By "close" i meant that perhaps the sensor can read through the cloth of the glove rather than be forced onto the exterior of the glove. I believe that touching is likely better, because I don't want to have to worry about interference from the sensors on gloves other fingers.

I am looking up hall effect sensors as suggest by KJ6EAD. Is that what you are describing, wayneh? This seems like a good solution, but I really would like to sense two on positions and one off position. It seems like this method can only detect the sets of paired coils; just one on and one off position.

Thanks again!

 

Is that what you are describing, wayneh?

Yes, but I was thinking of custom sensors wound as flat, or "pancake", coils that wouldn't take up much space compared to a generic sensor, if there is such a thing.

What do you mean about 2 "on" positions? Close and touching as 1 & 2? A Hall sensor might not detect touching per se but would give a signal proportional to closeness.

 

It would be expensive and complicated but one thing you could use are accelerometers. I don't know if you can get them the size that you would nee though.

The "fingers" would probably need to be initialized from a starting position then a mcu would keep track of where the fingers are after that. If you could get them small enough they could go inside, actually probably you would have an inner glove for the hand and an outer glove to cover the sensors.

 

I figured that there were two options, one was the double glove as you suggested, altho I'm not sure how accelerometers could tell if my index finger was pressed against my middle finger. The other was to put like square contacts on the outside of the gloves that would complete a circuit when touching. Still not sure how to differentiate between hands.

 

Is there any way you could us the "bend sensor" listed on this page:
http://www.karlssonrobotics.com/shop/flex-sensor-2-2/?gclid=CK-J0qeb_LMCFQWe4AodD00A3w

Somehow connect one end of the sensor to one finger, and the other end to another finger, and thus have it respond to the angle between them.

 

Still not sure how to differentiate between hands.

Pass a signal frequency instead of just power. Or have a different voltage for each hand.

 

I figured that there were two options, one was the double glove as you suggested, altho I'm not sure how accelerometers could tell if my index finger was pressed against my middle finger. The other was to put like square contacts on the outside of the gloves that would complete a circuit when touching. Still not sure how to differentiate between hands.

It wouldn't really but it could be calculated. You would just keep track of each finger's position in space.

Then again I could be totally off base.

 

That sensor seems hard to use for my purpose, but the karlsson site seems helpful. Perhaps my use of the word sensor was wrong or gave the wrong implication. I think if I can just complete a circuit? I would run wire up up the fingers and have something on the exterior of the glove that would complete the circuit, a small square of conductive material on on each side of the glove. I guess I dont really need it to differentiate between hands, that would just be a plus.

 

I'm still confused about the requirement. Is it proximity sensing , contact, both or other?

 

Well I suppose at the risk of someone stealing my idea, I will be as specific as possible. The intent is to create a "Light glove" like the product found at http://www.emazinglights.com/glove-sets/emazing-glove-sets/matrix-glove-set-trippz.html

Currently this is their most advanced option. It uses their eLite Matrix Chip: http://www.emazinglights.com/rave-microlights/elite/elite-multi-mode/elite-matrix-chip.html

These are used on each of the fingers of the most expensive glove, and they have a potential to display 3 out of 20 color options. Then the user chooses from 35 effect options that change the strobing effects and how many of the three colors to display. You can set up some presets, which are typically: 1. your main effect mode. 2. blank, to transition smoothly to 3: your secondary effect mode and 4: another blank.

I'm not sure how easy it is to change all your fingers, because you have to press a button for each finger and I don't have experience with this particular glove to have practiced.

So I'm interested in making glove set that can tell if my fingers are squeezed together. And I want to know results for individual fingers and preferably individual knuckles (or rather the midpoint between the knuckles) This requires a sense of squeezed or not squeezed and preferably same or opposite hand info. I do want to include an accelerometer because i would love to use that data to affect the lights as well. However, I'm not sure how accelerometers would have to be aligned to get all the right points, and I feel like if I can just complete a circuit or I dunno, depress a pad with pressure, then I would have a simple and effective method of telling my arduino or whatever what position my hand is in. orientation is space doesn't matter for me, just are they pressed together or not.

Thanks again!

 

Maybe just a thin tactile switch with the right button could work.

 

That sounds good actually. I really just need to know when the area is pushed on, I guess, so it seems that just a switch would be sufficient. I think I got too excited when I thought i'd like to recognize opposite hand connections, and that just recognizing when there is pressure is enough information.