Typically you do this the other way around.
Moving devices are beacons, fixed devices scan for them and report via something else, eg, WiFi. ESP32 development board is around $7.
What you’re describing is more the case of when a phone tracks itself relative to fixed beacons. But a phone already has connectivity.
It’s not really clear that there’s a role for LoRa (and especially not LoRaWAN) in your system, as that is typically too small a “pipe” for detectors to use to report their results.
The examples I recall seeing, of using RSSI for the application, seemed to rely on a fair bit of averaging, as in lots of TX RX cycles, to produce an average and reduce errors such as people or objects moving around in the building.
Therein is part of the problem as I see it; use of TTN significantly limits the potential for lots of transmissions and hence averaging.
It is a ‘solution’ I would like to try for myself (as a sceptic) and with a few Arduinos, it would be easy enough to do so.
At the moment however I am under local lockdown.
As the Welsh Assembly has to out-do Holyrood you’re about to confined to a single room in your home …
Hi @markdegroot
I’m afraid your idea would fail because a phenomenon called multipath fading causes the RSSI to vary wildly indoors. So wildly that it would not be possible to draw any conclusions (at all) as to the whereabouts of the transmitter…
Hi @markdegroot,
not really adding anything new to the thread, but confirming what other experienced people have said above, and given good reasons for:
this can and will not work, because (essentially) … physics.
there are good reasons why indoor location systems, specifically in hospitals, turn to RFID as reliable and cheap solution, possibly complimented by Wi-Fi/BLE tracking of people, maybe some UWB and ultrasonic for specialized systems.