@david moloney yeah this is a real pain point with 2.4GHz and you're right to be cautious — multipath is genuinely brutal at short to medium distances because you're getting constructive and destructive interference from reflections all fighting each other, so RSSI can swing 10-15dB just from someone walking between the nodes.
The good news is that at longer distances the multipath situation actually gets a bit more predictable, not less. Once you're out past a few hundred meters in open terrain, the reflected paths attenuate enough that you're increasingly dominated by the direct LOS component, so the RSSI tends to settle and you get less of that rapid jitter. On the MultiNav Pro+, the receiver sensitivity is tuned pretty aggressively which means you're still getting clean reads at ranges where a commodity module would be in the noise floor or just dropping packets entirely — and a dropped packet is way worse for ranging logic than a slightly noisy RSSI.
That said, for your 3-4 tier approach I'd still recommend smoothing — a rolling average over maybe 5-10 samples with outlier rejection will clean up the occasional spike, and make your tier thresholds wide enough that you're not on the edge of two zones constantly toggling. Something like "under X dBm = far, over Y dBm = close" with a dead band between them rather than a hard cutoff. The firmware repo is open so you can add that logic directly —
https://github.com/RFOXiA/MultiNav-Pro-Long-Range-BLE-Module-Firmware-STM32WB07 — and the AI firmware builder in the Club can actually scaffold that kind of RSSI-with-hysteresis logic pretty fast if you don't want to write it from scratch.
What's the deployment environment — open field, urban, airborne? That changes the answer a fair bit.
