Self-Localizing Sensors

Human vision is good at finding contours. Can a simple algorithm do as well? Yes!

Here I used a biological algorithm called “lateral inhibition” to find the edges & corners of a sensor array. All that is needed is that each sensor’s output has an inhibitory effect on its neighbors which falls off as the square of the distance (which is what a radio signal coming from a sensor does).

LEFT SIDE – A regular array and a jittered array.

RIGHT SIDE – What the algorithm decided:

  • Red indicates “corner”
  • Orange indicates “edge”.
  • Green indicates “inside”

This work was performed while employed at Sun Microsystems and formed the basis of a patent entitled “Sensor Localization Using Lateral Inhibition”. See patent details in my RESUME.

Read the full paper HERE