For the case of farms, I suspect that it’d be better to use a different system to position a tractor. Like, set up poles with lights on them and have a 360 degree camera atop the tractor roof or something. If you undo the location of the light poles and the bearings to at least two, you have position.
Maybe make them use RGB light and have distinctive color-changing patterns so that they can be distinguished between.
We’re gonna need something like that anyway, for high-precision tasks like having self-driving trucks align themselves to loading bays.
Cameras are not too great. Ground-station-based radio positioning systems, some of which can be used to enhance GPS precision to centimeters, already exist. (The ground stations, in a mesh tens of kilometers apart, get their position to that precision by averaging GPS over several days.) I’m pretty sure there is already a system on non-GPS frequencies too.
For the case of farms, I suspect that it’d be better to use a different system to position a tractor. Like, set up poles with lights on them and have a 360 degree camera atop the tractor roof or something. If you undo the location of the light poles and the bearings to at least two, you have position.
Maybe make them use RGB light and have distinctive color-changing patterns so that they can be distinguished between.
We’re gonna need something like that anyway, for high-precision tasks like having self-driving trucks align themselves to loading bays.
That’s definitely possible, but is way more expensive than using an existing system like GPS.
Cameras are not too great. Ground-station-based radio positioning systems, some of which can be used to enhance GPS precision to centimeters, already exist. (The ground stations, in a mesh tens of kilometers apart, get their position to that precision by averaging GPS over several days.) I’m pretty sure there is already a system on non-GPS frequencies too.
I also suspect vibration from the motor may make camera/vision based triangulation more difficult.