Learning about IMUs and such used in VR tracking remind me of some research I did a few months back for a pen that could trace what you’re writing and directly upload it into notes on a computer. One of my roommates brought up the idea while hiking, and we got really jazzed up about it. Half-baked product ideas are always stupidly fun to jabber about…
The idea for our pen was that all you’d have to do is buy the pen, use it on any surface, and it’d automatically sync to a predetermined section in your notes. This would be great for taking notes during meetings and whatnot (bad handwriting, eliminate distracting devices).
Existing pens generally require expensive “paper” notebooks that go with it or a smart pad under the paper to record. I can’t find the link, but I also remember seeing a pen that used an elaborate camera system to record position etc.
In order for the pen to track the writing without the aid of special paper/ tablets, it would *probably* need a hypersensitive inertial navigation system (INS), presumably placed at the tip. A combination of accelerometers, gyroscopes, and maybe magnetometers would allow the pen to track.
The main reason I stopped thinking about it is I found myself really, really deep in the science/ math/ engineering of INS’s and didn’t see enough plausibility at the time to want to think about it more. The main problem the pen would face is accuracy at the tiny size – and variability – of handwriting. INS’s have trouble with integration drift, which is basically the accumulation of tiny errors into much greater errors. This happens because new positions are calculated from previous positions, so the errors compound as more new positions are calculated. So the pen may be precise at measuring local movements but totally inaccurate in what it’s actually recording after a certain point.
Anyways, once it became clear that an INS would likely have to be supplemented with, say, a super good camera system, the idea got orders of magnitude more complicated. Fitting those systems plus a (Bluetooth? Didn’t even get that far) system to send the data plus the actual mechanical elements of a pen (ink tube, spring, etc.) onto one single, sleek looking little guy seemed like a task that wasn’t worth exploring at the time.
But now that I’ve found myself coincidentally drawn back to the technologies that could enable this kind of thing, I guess I’ll keep it in the back of my head…