I'm looking at the possibility of making a portable device to help me go in the field and find a lost node.
I read an interesting paper about AoA ToA and TDoA methods, but they seem complex and expensive
So I was making the analogy of a node being a screamer in the wild, so why not use the same as our ears, and have two directional antennas at an angle, with a listening node each. By reading the same message at the same time, there would be a different RSSI on the left and right, and after a bit of experiment and algorithmic manipulation, we could get an estimate of the azimuth and distance of that message.
Couple that azimuth with a magnetometer would yield the bearing of the last message, and help me align my heading toward the node.
In the radio amateur world, quite a lot of stuff has been done over many years on 'fox hunting' and radio direction finding.
If using RSSI in the manner suggested was practical, then I would have imagined hams have done it years ago, I dunno maybe they have.
Back to LoRa ...... received RSSI does vary quite a bit and the resolution is only 1dBm, so most likely you will not see a significane difference in RSSI, most likely the two readings will be very similar.
And of couse, RSSI levels in themselves are not much of an indication of distance, unless the area being searched is clear of all obstructions and flat, like in a desert.
However, you can do RDF with a single directional antenna by making the LoRa device emit a FSK carrier, you can then use a FM handheld to listen to the signal and the background noise level can be used to get a fairly accurate directional fix, especially at very long distances.
Just one antenna works fine, since you are holding it and can move it all around, any direction. Watch video of people trying to track an animal with a tracking device on it's body. Identical process to your proposal.
Thanks all, thanks for the step back & pointing to well developed fields.
Unfortunatly I can't do basic directional find since my signal is scarce, it would take way too long to get multiple messages and get a decent directional fix. Hence the idea of multiple antennas that would "catch" the signal simulteanously. Resolution being limited by the RSSI steps here, in theory very noisy, but once the system scales up to a half dozen high gain antennas?
I'm probably going to keep it cheap & order a couple of LORA boards (LilyGo LoRa32 V2.1 Q211 915Mhz) plus a high gain yagi (KM-041031-915), fix the reciever on a rotary table, and do some experimentation with a board fixed on a pole around 1.5km away sending pings at regular interval. Plotting the RSSI vs antenna angle might get me an idea of the feasibility of a RDF in that specific application, sensitivity to weather, repitability etc...
If anyone has ever done such experiments, I'd love reading about your findings!
These kind of things are to be found under "plane finder". In earlier days you'd just hit a switch on you rc transmitter and the device would start screaming, so you knew where it was. Nowadays it's just GPS and your transmitter/phone shows you the map. Then there are "avalanche beepers" (also around for ages) that just do what you describe.
I have have doubts as to whether RSSI alone will give an accurate indication of direction.
However if the transmitter is analogue its normally quite easy to get an accurate direction fix by ear, point the antenna for the loudest signal etc. I did on one ocaision get an accurate direction fix on a 100mW 434Mhz transmitter that was some 1,200km away in low Earth orbit.
You can make a LoRa device transmit simulated FM audio tones which you can hear quite easily on one of the many low cost UHF handhelds, like the BAOFENGs, so standard audio based RDF with a high gain antenna is possible and something I have tried.