Bluetooth triangulation is very hard (don't ask me how I know) and won't get you that kind of resolution. Apple's AirTags won't either. Dots painted on the floor at regular 1 yard or 10 foot intervals, in view of the cameras so that calculations can be inferred? Now we're getting somewhere.
The good news is security cameras have existed for a long time, and you can almost certainly find someone to bulk purchase from on the cheap.
If you insist upon using a wearable you may be able to get a light source on everyone you can calibrate towards. I'm not sure whether that's very helpful in practice.
With that caveat, to answer your question, you'd probably use Bluetooth LE devices that everyone carries. If everyone had wi-fi devices, you could use wi-fi triangulation software. This is usually an add-on package to existing enterprise wireless solution.
You could also do active NFC tags and have choke-points with scanners everywhere.
I have my own Python script at work, appropriately named creep.py, that locates people by wi-fi by logging into our enterprise wireless system, checking what AP their MAC/DHCP client ID/username combo is connected to, gets the RSSI, then runs that through some if/else statements to determine what room they're in to a high degree of accuracy. The script also gets readings off the adjacent APs, as well, for better triangulation. I keep quiet about this script, lest it be used for police state purposes.
The above is a simplification of the script, but if RSSI is approx <=55 or so, then the person is likely in the same room as the WAP, given normal US construction practices. ~65 or so, there's probably a drywall wall between the person and WAP, etc etc. This is an extreme oversimplification, but gives you an idea.
tl;dr:
Your tracking options are:
- Bluetooth LE
- Wi-Fi
- NFC
- GPS devices
None of these will get you to a couple inches of accuracy. Think more like "room-level accuracy".