• force output parallel to the surface of the human fingerpad;
• fingertip tracking across a flat surface.
Algorithms using the system to implement forms of touch are generalizable, as they can deal with input in millimeters (mm) and output in newtons (N).
During development, a lot of time was spent on custom signal conditioning in the input chain.
As a result – despite being based on noisy, off-the-shelf optical mouse sensor input – algorithms running on top of the hardware can induce aspects of fingertip surface texture perception at high resolution.
One example of this is when extending turntable scratching with friction display. How this can be done, at millisecond resolution, is described here.