An interesting project to replicate – or at least to try to

While reading through the project blog of all famous wiimote master, Johnny Chung Lee, I have found out about Sensitive Object – a company that developed a quite simple yet powerful technology to make any object touch sensitive. Basically it they use few (two) microphones or more likely piezo-electric vibration detectors and a pattern matching technology. After the sensors are placed the process of gathering samples takes some time but after that, you are ready to go – you can tour make invisible light switches, a keyboard directly on your desk, or tourn a vase into interactive, touch sensitive object.

Sounds nice indeed. So, it came to my mind, why not to try to duplicate it? My goal is to make a prototype of wooden or plexi board that would be touch sensitive – if I would get 5 cm of precision on 50 x 50cm sourface I would be extreamly satisfied.

So far here is what I think I will need:

Hardware:

  • At least 2 elctret microphones (piezoelectric sensor.s are hard to find where I live Ok, not really if figure out how the sales people name it…)
  • At least 2 cheap USB sound cards to have the microphones connected to.
  • Plexi or wooden board approximately 50 x 50 cm, not more than 1cm thick

Software:

  • App to gather the samples and map them to the points on the boards, maybe noise removal would be nice .
  • Self-learning neural-network that will be trained using the gathered samples to match the ‘taps’ to the points on the board.

Well, I guess this may be not so easy to get any results, as there are many problems easy to foresee, that can make this a total failure:

  • How the sound samples should be presented to the neural network? Should a time delay, phase shift, or anything be used to distinguish the sources of sound?
  • Will 48kHz or 96kHz sample rate be enough?
  • Will the cheap USB sound cards be enough?
  • Will the elctret mics do the job?
  • How the neural network should be designed?

I hope the share you with some initial results any time soon. If any could point me to some resources that could help please do so. :)

0 Responses to “An interesting project to replicate – or at least to try to”


  • No Comments

Leave a Reply