Hey guys, I'm looking into feasibility/best method of a touch interface that would work through glass, namely a car's windshield.
Background: I like very much having a keypad on my car. So much so that I added one to my last one the hard way. I added a Ford keypad to my Dodge by cutting a hole in the door and handling the in-between with an Arduino. http://www.arduino.cc/cgi-bin/yabb2/YaBB.pl?num=1236735887/0 Unfortunately, I lost that car to a careless driver and now I need to start over on my new car. My first thought was to just cut another hole, but honestly, I don't think the keypad would look quite right on the new car, too many curves.
So I was thinking about making a touch interface that I could mount on the inside of the windshield. I would only need to simulate 5 buttons. The two possibilities for this that came to mind are capacitive and ir sensing. From what I've read, some of the capacitive touch sensors can sense through a small layer of glass, but I have my doubts about using it with the thickness of a windshield. Also, I've read that not having a true 'earth' ground causes problems with these also.
My other option is (one or more) ir leds around an ir receiver, so that when you place your finger on the glass, your finger reflects back the ir, picked up by the receiver and interpreted as a button press. The only downside to this is that whenever the windshield ices over (or just gets dirty) it will look like constant pressing.
A knock sensor would be neat (shave and a haircut, two bits!), but this feels too insecure for me.
My car already has an rf key fob. What I am looking for is a way to get into my car without having anything on me, be it a key, card, ring, remote, or whatever. As I said, I can always cut the hole again and implant the ford keypad into the door, but I was looking into other solutions for aesthetic reasons. That, and re-doing my last project identically wouldn't be as interesting or as much of a learning experience.
I don't see why the ir method wouldn't work. I might have to modulate the light in some way to reduce false positives from outside sources, that and the aforementioned frost issue.
How about magnetically? You mount reed switches inside the window, read by arduino, then a ring on your finger with a magnet attached is used to enter the code sequence needed to unlock it. Keep the magnet on your belt or something.
Digikey has a wide variety of them.
neonpolaris sounds like an intriguing project.
Can we assume this is a daily entry method, or is it just when you lock yourself out.?
There are IR Tx, Rx paired at an angle. We used one that was 1mm distance for paper out sensing.
The testing for false positives could be as simple as 1, then 2, then 3, then 4, then 5, have it acknowlwdge you, then enter the code. That way you would know you have to wipe it with some part of your clothing....(you said you weren't carrying anything, not that you weren't wearing anything )
If its an emergency only application, could the grille area allow for a clear top box to house it in.
I've just had another thought.
What about using LDR to detect the light level. They would be relatively self calibrating, and if spaced far enough the ice shouldn't cause the darkness to be sensed by adjacent LDR's.
Try out this site they make a high power switching and low power switching capacitance module, that allows you to make your own key pads through glass or plastic, or wood, pretty cool.
Canoeman
)