I'm using another vendor beacons that use the nRF52832 as the basis of them and am seeing things that I don't understand on them:
Using a tablet (Lenovo M10) and nRFconnect to watch the RSSI values of beacons in the area I see
a) quite a difference (like 10-20 dB) difference in RSSI values between different beacons at the same distance.
b) some beacons that have quite variable RSSI values (like 25dB) over time, particularly at 4-6' distance.
I'm trying to get distance from these values or relative distance. Basically I want to see the difference between a beacon that is say 12" away and one that's 24" away. Is this a completely impossible quest? What is a reasonable distance difference to be shooting for? what could be the thing creating the variable RSSI values? Background noise?
The antenna gain varies largely depending on the near field characteristics of the antenna. The near field characteristics are prone to changes when there are obstacles, like a person, a piece of furniture…
Variation in RSSI can be caused by any interferences like any noise picked up by any external 2.4 GHz device (like the wifi you use), regardless of protocol it uses (wifi, classic BT, proprietary, etc). It can also vary even if a person moves or the location of the device changes. A proper and steady RSSI is a bit unachievable (in a completely stationary environment) but you can calculate the distance by other means which involve filtering and averaging of your values taken over multiple readings.
basically low pass filter those readings, yes? Assuming we can do that, what would you figure is a reasonable distance resolution that is possible?
The antenna gain varies largely depending on the near field characteristics of the antenna. The near field characteristics are prone to changes when there are obstacles, like a person, a piece of furniture or even anything small. So basically you need to be in full control of your transmission and reception antennas which is very unlikely. Using a low-pass filter would also not give good results. Taking multiple readings and then averaging is better even though the result will still be inaccurate.
I believe that handheld devices are not suited for this application because the user will affect the measured RSSI. Due to the signal noise, it's not usually a good idea to estimate the distance using RSSI measurement. It can at best be used to realize the proximity of the beacon and not the exact distance.
Ah, this link helps a great deal.... Thank you for that one.