I have a sensor tracking question. I’m working on a project where we have a need to measure where a small handheld device is moving on the surface of human skin (it’s for a non-invasive breast cancer probe!), starting in 2D, but perhaps in 3D at some point.
We have a few ideas to do this (our current choice an optical mouse type sensor and a 9dof).
Accuracy on the order of 1mm would be great.
It occurred to me to ask, If one of you sharp engineers at SparkFun were asked to do something like this, how would you go about it??
An optical mouse sensor makes images of the underlying surface and tries to see how much they moved. This means that each frame has some position uncertainty with regards to the previous position. Over a long distance across the skin this error adds up destroying your desired accuracy.
Consider the older mouse technology instead. A wheel for linear distance or a ball for 2D position in contact with the skin. Which is measured with an IR/light gate encoder. Only wheel/ball slip, and rotation of the device along the path, would put a damper on your accuracy requirements.
TL/DR: Open up an old fashioned ball-mouse and get inspiration from that.