Centimeter accurate triangulation over a desk

I've found many triangulation/location projects, but my settings have several restrictions that might make the solution easier, so I want an opinion about the best approach.

I'm working with a very small ESP8266 based board (WeMos) to create small self-contained blocks that are put over a table (or any unobstructed small plane region with at most 2m on each side) and I want to set 3 sensors as reference points and make every block on the the table to be able to find its own position ideally with centimeter precision. The boards naturally have wifi but I can add other sensors or components, as long as it's reasonably cheap. Is there a better solution? What kind of signal/transceiver should I use? I believe the board can measure time with microsecond precision and +-4 average error.

I would do that by putting an LED at each of the triangulation points and getting the Arduino to rotate a shielded direction-limited photo-transistor and note the angles at which the lights of the LEDs are detected - similar to how ships used to navigate by taking bearings from light-houses and hills (and stars).

Somehow I suspect you have something different in mind.

...R

Robin2:
I would do that by putting an LED at each of the triangulation points and getting the Arduino to rotate a shielded direction-limited photo-transistor and note the angles at which the lights of the LEDs are detected - similar to how ships used to navigate by taking bearings from light-houses and hills (and stars).

Somehow I suspect you have something different in mind.

...R

No, I think this is a great idea and could work for me, but I never used this kind of component before. What kind of keyword/partnumber should I be looking for if I try this one?

Thanks for the help!

I don't have any part numbers. You could probably use any bright LEDs (maybe Infra Red) and most any photo transistor that is sensitive to the same wavelength as the LEDs. Perhaps use a servo that can rotate at least 360 degrees (probably a sail-winch servo) with position control. Mount the transistor at the end of a narrow black tube (rolled paper?) so it can only "see" over a narrow angle. Maybe squash the tube so the vertical angle is wider than the horizontal angle.

Rotate the servo back and forth through about 380 degrees in small steps and check the output from the photo transistor after each step.

Great fun.

...R
PS a more complicated but probably more accurate alternative to a servo would be a simple geared DC motor and a high precision rotary encoder on the output shaft.