Keyglove one-handed input device

I've made great headway on my Keyglove project (the code for which I got some help for on this very forum!). I'm sure it's going to work nicely as a finished product, but I'm still in the prototype stage.

The keyglove is a 34-sensor input device that relies on touch combinations (not pressure, just touch) to generate keyboard and mouse input signals. At the moment, it only supports the PS/2 interface, but I'm hoping for USB and, eventually, Bluetooth as well. I'm doing more development on it every day, and if you'd like to check it out or follow the progress, go to:

http://www.keyglove.net

... or follow @keyglove on Twitter at http://twitter.com/keyglove.

The Arduino is awesome! And, I might add, so are all the people who have written libraries for it for us to use. The ps2dev code has been invaluable.

This looks cool. Nice job. How are the contact sensors to work with?

Thanks!

Assuming I understand your question correctly, the answer at the moment is “I don’t know.” I haven’t had a chance to work with the material much, since it just arrived in the mail today. However, it is very thin and I don’t expect it to be disruptive, and I know it won’t add any significant weight.

My main concern is making sure I put them in ergonomically comfortable positions. Particularly the palm sensors and the thumb pads need to be just right. I will have more time this evening to actually mount the sensors to the glove and test it out a bit.

hey this is reallly cool im doing somthing similar right now. i made an animatronic hand controlled by servos and am currently working on the glove you wear wich has flex sencors in it. its for science fair for my school and its wireless with 2 xbee moguiles. but thats cool i would love to hear more baout the kind of sensors you use and all that stuff

Your project sounds cool as well!

For the sensors pads, I was going to use copper foil squares epoxied to the glove, but that has proven to be very cumbersome--nearly as difficult and messy as the plain silver epoxy approach.

I'm currently looking into conductive fabric as an alternative from this place: http://www.lessemf.com/fabric.html I've ordered the sampler pack to see which fabric might work best. Stay tuned for more info (or just follow the project on Twitter or via RSS).

For the sensors pads, I was going to use copper foil squares epoxied to the glove, but that has proven to be very cumbersome--nearly as difficult and messy as the plain silver epoxy approach.

Have you thought about using homemade "conductive wire glue"?

http://www.instructables.com/id/Make-Conductive-Glue-and-Glue-a-Circuit/

http://www.inklesspress.com/conductive_glue.htm

:)

PS - note that he details how the glue can also act as a resistive bend sensor/strain gauge (perfect for wired-glove applications!)...

Home-made conductive wire glue sounds interesting. My very first approach was to use wire glue, since it nominally sounds perfect. In reality, the stuff I got wasn't very conductive or strong (from a "glue" standpoint) and failed miserably. It was $5 off eBay though, so that could have been part of it.

Although it might not be quite as simple, I really like the idea of conductive fabric since it would end up being the most professional and safe kind of sensor. It might even be washable. It will bend and flex nicely. My only concern is if it's solderable.

I'll definitely keep those links handy if the fabric doesn't pan out. I'm not sure if the pressure sensitivity would be beneficial though, since it's my goal to have the pads require only touch and no real pressure. The I/O pins are all digital, not analog (for simplicity and because 34 analog pins are hard to come by on any of these AVR-GCC compatible chips). So having the voltage vary based on pressure might end up complicating the circuit.

But thanks for the input, and like I said, I will definitely keep those links close by.

For anyone interested, I got the conductive fabric sample pack from LessEMF today and did a bunch of testing. A few of the fabrics might do well as sensors (one in particular). My main challenge will be figuring out how to attach the wire leads to the fabric, as the one that otherwise has all the best qualities isn't solder-friendly.

You can see detailed info on each kind of fabric here:

http://www.keyglove.net/2010/10/15/conductive-fabric-performance

Does the Material manfacture offer any suggestion for attaching wires to the fabric?

Other wise you could just weabe the end of the wire into the fabric and mabee a dot of superglur to hold it on?

I sent an email to the LessEMF guys yesterday to ask if they had any tips for wire attachment, but so far I haven't heard anything back from them. It has been less than a day though.

The weave + glue dot is a good idea, and I was thinking of trying that very thing, but I haven't done it yet. If I figure out something that works reliably, I'll be sure and post it here and on the keyglove.net site as well.

Thanks for the recommendation!

Instead of superglue, some of that homemade conductive glue might work better. Or, if you want something more commercial, there are glues/epoxies available that are electrically conductive (usually silver-bearing - so they aren't cheap), used for PCB trace repair and similar uses. Either way, by bonding the wire to the fabric with something conductive, you might get a more reliable joint.

Hello,

I am very interested in your keyglove project. It looks like it will turn out very nice.

Here's a thought on the conductive fabric:

Instead of trying to solder the wires to the fabric, what if you did this...

If you got some stranded wire, and the strands were very fine. Almost like the type that they use in the wires on earbuds. Anyway, the strands are very fine.

Then you fan out the the strands, and lay the piece of conductive fabric over that. Then you sew the fabric to the glove, with the strands of wire sandwiched between the glove and the fabric. Also, I've never used it, but I picture using conductive thread to sew it together.

I hope that is pretty easy to visualize, I will try and get some pictures up later of what I mean.

Joshua

I’ve found that wire glue + epoxy works great. The wires can’t be pulled off or out of the fabric without significant force, and the wire glue under the epoxy bead maintains great conductivity.

The “very fine wire weave” idea isn’t bad, PianoKid, but I would worry a little about the wire coming out. You would have to sew very, very tight stitches, and then if the wire did happen to come out, it would be very difficult to attach it again.

Anyway, here’s the latest update: http://www.keyglove.net/2010/10/16/working-sensor-material-and-wire-attachment

New updates on the Keyglove project, particularly of interest to cr0sh: I'm going to try that glue you recommended. It will hopefully be useful in simplifying the build process. What I had before worked okay, but it made assembly very difficult because of the order in which everything had to be done (specifically, sensor wire attachment before sewing). I hope that by using this new glue, I can sew the sensors on all by themselves, which is way easier, and then attach the sensor wires afterward.

I have also discovered a more optimal sensor size across the board. I had been making all of them rather large, and this turned out to be a bad idea. Accuracy is far too difficult. It works much better if the main thumb/fingertip sensors are large, and virtually everything else is small. I will be building a second prototype to thoroughly test this as soon as the new gloves show up.

I'm also having excellent success with an ADXL345 accelerometer, which I will be using for mouse control. I haven't got too far on that just yet, but the I2C interface is working great, which to me was the hard part.

I've put together a rather basic (and not fully complete) set of instructions on building your own now, available here: http://www.keyglove.net/build

I have made great progress overall, and more good stuff is happening almost on a daily basis. I was also lucky enough to be picked up by the Makezine blog this afternoon. Awesome. Go to http://www.keyglove.net/news for the latest news, as usual.

Looks really great! I hope the conductive glue works out for you. Your blog link is now a part of my VR collection of bookmarks; I'll definitely be following your progress!

:)

Jeff. I love your project (but I believe I may have posted something like that on your website.)

Have you taken a look at what I am fooling around with? (fkeel.blogspot.com)

Once my overall design is finished (and I still have a far way to go) I will take a closer look at what you came up with and possibly extend my setup to incorporate your glove.

Eventually (time-wise this will be sometime next year, when I’m writing my Bachelor thesis) I hope to use this a an input device of a context aware, wearable computing device… :slight_smile:

I love the fact that lots of high-tech stuff has moved to outside of MIT Media-Lab and similar places into the garage workshop of various tinkerers…

just imagine the possibilities.

a Keyglove on your left, a mouseglove on your right hand for explicit input

body motion capture & biofeedback for implicit input

webcam and microphone for environmental input

eyeTap and earphones for output

pandaboard for computing and internet connectivity

*

put all that on and use it and you might find out that you are just a character in a novel written by william gibson...

@fkeel,

Indeed it is starting to sound like a William Gibson novel (also a bit of Snow Crash as well). Your scenario is very interesting, as is your motion-sensitive glove. I prefer more precise, exact motions (like touch combinations) for keystrokes, compared to the more fluid, analog detection of flex and rotation, but that's just for typing. For moving around in a 3D space, your glove looks awesome.

Jeff

yeah - we are doing completely different things. yours is for exact user controlled input (what I referred to as explicit input). mine is for registering body language (what I referred to as implicit input... )

(just to get things straight, my goal is not to develop a glove, but rather a suit or something. these two videos should give you a better idea of what I'm doing

http://www.youtube.com/watch?v=vT3CkJ0n9gM (rather old)

and

http://www.youtube.com/watch?v=Hk62nHSvSFM) (just recorded a couple of minutes ago)

the point where they meet is that they both can be turned into very low profile input tools (and true niel stephensons description of the gargoyle in snow crash might be the closest fiction wise.)

*

More on topic:

Why do you use the glue to connect the sensors? Why not use the glue as the sensor?

I may be missing something obvious, but it seems like you really don't need the conductive fabric if you can just have conductive glue... you can also cut back on sensor wires to a certain extent by just "drawing" the wire with the conductive glue. (of course I have never worked with it, so this may just exceed what the material is capable of, but it is worth checking out, imho.)

cheers

p.

Those videos are very interesting. Such a project seems a lot more ambitious to me than the relatively straightforward glove I am building! :-)

I will definitely test what you have suggested with using the glue as the sensor as soon as the materials arrive and I can make a batch. I am afraid it might not work simply because the conductivity between two surfaces (i.e. touching them together) may not be nearly as good as the conductivity inside the material. That was my trouble with many of the fabrics I tried as well as the original wire glue I've been using. It works to solidify electrical connections between two points, but the dried glue touching another bit of dried glue is not nearly so conductive.

However, it's definitely worth an experiment, since if I could eliminate the fabric, that would cut down the cost a little and the build difficulty a LOT (no cutting, pinning, gluing, or sewing). I'll let you know how it turns out.