Its a hard problem. Using multiple IR cameras and IR beacons on the moving objects can work well
(for instance its been successful in drone-swarm research), but it involves CV and some decent processing
power.
The UWB time-of-flight techniques I refered to above seem to be the way of the future for position
triangulation indoors (GPS obviously doesn't work). Triangulation requires multiple base stations
(at least 3, 4 gives 3D position).