Software defined radio RF tracking option

[CofC_HighBall]

A tracking solution that relies on direct RF interface between the ground station and payload via a Software defined radio. There are already current examples which can send live coordinates too, see link. Could this be used for our system?
The FAA already tracks all balloons at the designated 400-to-406 MHz and 1675-to-1700 MHz bands per international agreement.

UKHAS wiki Link

[leviw]

Over the last week, we've been working on adjusting the dish position based on RSSI between the ubiquity modems, which is available as JSON on it's <ip of modem>.status.cgi page. We're using GPS coordinates to find an initial angle, then 'fine tuning' it by making small adjustments to the elevation and altitude and checking to see if the RSSI improves or declines.

Our early tests look promising but it's a tricky thing to test over short distances on the ground because we see similar signal strengths over a wide range of angles. (Our dish often gets the best signal by pointing at the wall behind it.) We're trying to improve our algorithm to narrow the search window to a smaller number of degrees, based on the initial gps positions.

If there's interest, I can post our code to github sometime in the next few days. It's maybe 200 lines and runs in Spyder, no extra libraries to install.

[CofC_HighBall]

Apr 2, 2017 8:02:55 GMT -7 leviw said:

Over the last week, we've been working on adjusting the dish position based on RSSI between the ubiquity modems, which is available as JSON on it's <ip of modem>.status.cgi page. We're using GPS coordinates to find an initial angle, then 'fine tuning' it by making small adjustments to the elevation and altitude and checking to see if the RSSI improves or declines.

Our early tests look promising but it's a tricky thing to test over short distances on the ground because we see similar signal strengths over a wide range of angles. (Our dish often gets the best signal by pointing at the wall behind it.) We're trying to improve our algorithm to narrow the search window to a smaller number of degrees, based on the initial gps positions.

If there's interest, I can post our code to github sometime in the next few days. It's maybe 200 lines and runs in Spyder, no extra libraries to install.

Great idea, please share the code when you find some time!

[David MSGC]

We tried a similar approach last summer, trying to point based on rssi, but were never able to spend enough time to get an idea of how easy or difficult it would be to implement. I would be curious to see your results/code. I have seen rssi based algorithms work very well in hidden transmitter hunts with the local ham radio fox hunt antennas.

[leviw]

So after a little tinkering here's what we've learned about tracking via RSSI: it's not as easy as we expected.

What we're seeing is that, at least in ground tests, the best RSSI signal comes from unexpected directions. The following code works for us and will rotate the dish around to find the best signal, but it's often coming from the nearest wall. We took the payload outside and walked it around a track maybe 1/4 mile away from the tracker, which had a tendency to either point towards a row of trees behind the payload, or the brick wall behind the tracker. We could see from the RSSI that these two locations did indeed provide the 'best' signal, but that we can't expect those strategies to work during a flight.

So now we're shifting gears to another aspect of our project and leaving this alone for now. We plan to come back and try something similar again using a narrower search window, say a couple degrees away from the expected targeting solution. I.e. if we expect the balloon is at pan=42, tilt=110, then we want to search from pan=40:44, tilt=108:112. That way it finds the best signal within a narrow range from where we expect the best signal to be.

Anyway, here's the code that we've been experimenting with. The most useful part is probably how to scrape the RSSI signal from the ubiquity modem.