I proposed a research project a decade or more ago at a university to use a simple light sensor to build an ultracheap location device. It would use light inte

I proposed a research project a decade or more ago at a university to use a simple light sensor to build an ultracheap location device. It would use light intensity to figure out when sunrise and sunset were along with an internal clock. I have the details archived somewhere, but while it was super cheap it was not fast. It would take a day or several days to get a location approximation, and it could possibly be fooled by mountains or hills causing sunset or sunrise (or both) to occur later (although I think that as long as it had some exposure to outdoor light it would probably still work, since sunset behind a mountain is not the same as actual sunset). It needed a fair amount of storage to store some lookup tables. Even so the idea was it would be able to calculate approximate location anywhere in the world at low cost. The use case I imagined was package and equipment (cars, bikes, tractors, gear) tracking, in part for theft recovery. If the system had a radio it could use opportunistic connections to send its estimates occasionally (every few days perhaps) to a remote system. While it had accuracy limitations it might have been useful to get an approximate idea of where anything with a small tag attached to it was in the world. Once you had a rough idea of where it was, that narrows the search for whatever is missing. When you get close to it the radio could help refine the search (set out some sniffers and wait for the once every few days ping to try to triangulate it.) It didn't really fit with our research groups expertise though and we never pursued it. The opportunistic network connections could have also helped zero in on the location.