Hi Joachim, Thanks for the answer. I'm aware of that but am very specific about not using IR based tracking. Thanks for the quick response.

Well, it's not capacitive nor magnetic, but seems to be working very well and the input device is freely available and not really expensive:

http://www.youtube.com/watch?v=Fxa43qcm1C4

http://www.ics.forth.gr/~argyros/research/kinecthandtracking.htm

Try on ADI's website - there you can find some physical solutions also - I know they had an article this or last year.

This was for finger (like on a touchpad). Unless, maybe some acceleration, position and GPS sensors with wireless connectivity may do - but this is not cheap at all. You can try finding this also on ADI's or Microchip's website. Maybe you can find some theses.

See the web site of: Leap Motion, Point Grab, primesense, enviosion and many more

The problem with all these systems is that they are either computer vision systems which pose non-line-of-sight detection problems or are off-the-hand stationary tracking systems i.e. they are not a part of the hand.

I suggest separately systems:

- For each hands a systems built with 3 wires connected to simple wire-spring resistors, from which you can establish the positions elaborating the three relative distances;

- For each fingers, optic fibers inserted in the gloves, having Fiber Bragg Gratings (deformation sensors) in the junctions.

Every components is EMI resistant, you can use the system also in radiation environment (shelding only the cases of wire-potentiometer....).

One can make a glove with adequate sensors, which signals are then preprocessed by an adequate microprocessor and are sent then via wireless to the receiving device. Because of the processing unit (and the additional power supply), the system would be quite uncheap due to the hardware and software component. The glove can lack and the sensors are stuck (glued to the hand). Some additional types of sensors (acceleration and GPS can improve the application).

These gloves do already exist. Look for CyberGlove Systems for Instance, they use strain gauge sensors for measuring the hand pose. But they are very expensive, though.

Exactly my point. I do not want expensive gloves. There a ton of such expensive gloves in the market and in academia. Ideas include magnetic dipole calculations, IR, ultra sonic sound and doppler effect, multiple accelerometer and gyroscopes on each finger etc. But all these options are really costly.

To track trajectories of each finger you can use the spherical triangulation algorithm and 3 Cable Extension Transducers (http://celesco.com/faq/cet.htm) for every finger.

May be the cheapest solutions. Only the fingers can't cross each others.....

This is a good beginning I believe, though I dont have any cost data.

http://www.edn.com/electronics-products/electronic-product-reviews/other/4401197/Microchip-elevates-HMI-design-with-world-s-first-e-field-based-3D-gesture-recognition-chip?cid=Newsletter+-+EDN+on+Systems+Design

This would never have the precision of a glove or of a vision based system. It works similary to a touch pad. Would it cope with two hands' movement. A third hand or some other moving object would make it very missunderstanding. And this is also not cheap at all.