Slashdot Mirror
Building a Pressure-Sensitive, Multi-Point TouchScreen?
sonamchauhan asks: "I'd like advice on building (yes, building) my own low-res touchscreen. The reasons for 'build' instead of 'buy' are: 1) to have it sense pressure (pressure sensing is quite expensive) and 2) to have it sense multiple points of contact simultaneously (which is a useful thing). Back in 1985, researchers at U.Toronto built (PDF file) a touch-tablet (not a touch-screen) that fulfilled both requirements (pressure-sensitive and multi-touch) and used only basic electronics: lots of diodes, A/D convertors, etc. Some 17 years later, it should be possible to build a touch-screen using the same techniques (possibly using layers of transparent conductive and insulating paint for the sensor paths.) Any comments? Some other links: a Microsoft paper (PDF file) describing a touch sensor painted onto a mouse, a basic FAQ on current touchscreen technology, and a slashdot thread that discussed building touchscreens (these links don't address pressure or multi-touch though)."
I would think that the time and effort that went into developing something like this, basically from scratch, would outweigh the "real" cost savings of not buying one off-the-shelf. I suppose if this is just for fun, than that's not really a concern, but if your time is worth anything at all, it ought to be considered.
1- have the slashdot community provide plans for a pressure-sensitive multi-point touchscreen
2- ??????????
3- profit!!
First, is it ok to use a stylus? If so, you're problem just got easier - you simply use the same technology that wacom uses, except you'll need to track several pens instead of one.
I'm betting, though, that you want a real 'touch' screen which registers anything hitting the surface.
Here's what you do: Use a transparent metalization process to put squares of transparant, conducting (but resistive) metal on two opposing pieces of plastic. Take a wire from each square and use a microprocessor with a lot of i/o (switched A/D converters, most likely) to detect which metal plates are being pushed together. The size of the plates determines resolution.
That's the easy, no brainer way.
The cool way (with infinite resolution ) is to use the normal method modified a bit. Normally you have two sheets covered with resistive material, one which has conductors on the vertical sides and one with conductors on horizontal sides. Increase the conductors (say, four shorter conductors on each side, and make them points instead of lines).
Send a signal to one conductor, and listen to all the other conductors for the wave front of the returning signal. Through a ton of signal processing (and sending signals from the others and measuring the response) you'll be able to detect an arbitrary number of points on the screen. You'll need to do a lot more processing to measure surface area (pressure), but you'll get there eventually.
Alternately, you get rid of it entirely and make a video system that can sense the position and movement of your hand and get rid of the 'touchyfeely' altogether. This (IMHO) would be easier and faster to develop.
-Adam
From a whole lot of research I have been involved in, for a while we worked with a company, TacTex Controls Inc.
They have a pressure sensative pad and system you can use, the MTC Express (http://www.tactex.com/prodMTC.htm), however, that is not where you would be interested directly.
Their technology uses fiberoptics in a rubber pad with a small controller. There is a lot of research in integrating their technolgies into other things (a prototype of a complete pressure sensative suit was discussed at one stage).
It fufills your requirements quite well.
256 levels of pressure for each region of the pad independently.
The MTC Express pad had only a DB9 serial connection, and needed a power input as well. However, the OEM material and controllers were a lot more flexible in usage.
ICQ# : 30269588
"I used to be an idealist, but I got mugged by reality."
You know about these guys:
http://www.fingerworks.com/
right? They've been mentioned on Slashdot before. Their products can detect multiple fingers at once.
They are selling tablets and keyboard replacements, not touchscreens, but I think I read somewhere (on their website maybe) that they believe it would be possible some day to use this tech in a touchscreen.
The cool thing is it sees your whole hand via infrared, so it can tell which finger you're using. For the first time, you can "right click" on a touch surface, or use other finger combinations/gestures for various operations (drag, pan/scroll, rotate, etc.)
Been done (like 20 years ago and still in use today) but that does not fulfill either of the specifications.
1) No pressure sensitivity
2) When 2 fingers are pressed at once you get 2 X values and 2 Y values (X1, X2, Y1, Y2). So if they both happen at the same time are the fingers at (X1, Y1) and (X2, Y2)? or are they at (X1, Y2) (X2, Y1)? You can't tell. So multiple touches also doesn't work. (If the fingers come in one at a time then you can tell. But the fact that they MIGHT come in simultaniously and that there is now way to tell means you really can't support it)
#2 is a major obsticle to designing a touch system that can actually handle multiple touches. A touchscreen with hundreds of tiny etched "wires" across the inside of two surfaces that "connect" when you touch can get around this and is also widly in use today (Though the "wires" are usually spaced pretty far apart, giving you no more than 10 points per inch resolution or less). But there is no pressure sensitivity. You might have to combine two different technologies to get both.
Contrary to popular belief, coding is not all free blow-jobs and beer. Those things cost MONEY!
If not, you might think about using a standard touch screen system, then building (or buying, if they exist - probably don't) pressure sensitive "gloves". For the sensors (at the finger tips), buy thin ribbon plastic tubing (multiple tubes joined in a ribbon - used in many continuous ink kits for ink-jet printers, so you should be able to buy it somewhere by itself), spread one end and seal the ends of each tube, apply gentle heat and blow "bubbles" at the end of each tube (this may be difficult, but not impossible). Then, attach these bubbles to the fingertips of the gloves, and use a T-connector on the other end of the cable to hook the tubing to a very low pressure air supply and the other side of the T-connector to a air-pressure transducer. Hook the transducers up to a PIC or Stamp (or other ADC system) to measure the pressure change (the positive pressure keeps the bubbles inflated at all times, and reinflates them after being depressed - a closed system might work too, with no active pressurisation - you might have to experiment here).
Other cool things to do with the gloves:
1. Touch gestures (like a chorded keyboard)
2. Use water instead of air, then warm or cool the water with a peltier to provide feedback.
3. Pulse the pressure for other feedback needs.
Hope this helps...
Reason is the Path to God - Anon