GSM Direction Finding

[Psykeman]

Hi everyone,

I’m currently using the KrakenSDR for DF purposes in the GSM bands (880 - 960 MHz) and I’ve a couple of question that I’d like to ask

1. In the UCA setup, which antenna is the reference for the DoA? Is it the ch0 antenna?
   (Like if the ch0 antenna would face straight to the base station without any obstacles through the path, would the DoA be 0°?)

2. Does the multipath affect the accuracy of the setup very bad? I’m using it in an urban environment, and the results that I do get aren’t very convincing (Like a DoA of 320° NW when the ch0 faces the North, while the BTS is at 167° SE from the north (In a clockwise sense) )

3. Supposedly I put my setup in a rotating platform, how the setup would read the angles? In a clockwise or anti-clockwise sense?

4. Is there any way to increase the resolution with MUSIC? (Can this super-resolution algorithm increase the resolution beyond 10 times?)

I’m sorry if some questions aren’t that clear or may sound dumb, english is not my native language and I’m a student .

Anyway I’d like to thank you for any answer.
Best regards,
Masanni

[krakenrf_carl]

@Psykeman

• In the UCA setup, which antenna is the reference for the DoA? Is it the ch0 antenna?
  (Like if the ch0 antenna would face straight to the base station without any obstacles through the path, would the DoA be 0°?)

Yes CH0 is the reference antenna, and ues the direction of CH0 is defined as 0 degrees.

@Psykeman

• Does the multipath affect the accuracy of the setup very bad? I’m using it in an urban environment, and the results that I do get aren’t very convincing (Like a DoA of 320° NW when the ch0 faces the North, while the BTS is at 167° SE from the north (In a clockwise sense) )

Yes multipath is always an issue. The idea with the app is to drive around and take many readings at different locations so that multipath is averaged out. If you only take one reading at one location, then that reading will be skewed by whatever multipath corruption is going on at that location.

@Psykeman

• Supposedly I put my setup in a rotating platform, how the setup would read the angles? In a clockwise or anti-clockwise sense?

I’m not sure what you mean by a rotating platform? The software expects CH0 to be the forward direction, so you’d just need to compensate for whatever your external rotation is.

@Psykeman

1. Is there any way to increase the resolution with MUSIC? (Can this super-resolution algorithm increase the resolution beyond 10 times?)

Not really, not without adding more elements in the array and using a receiver with more inputs (ie not the KrakenSDR). And those probably don’t exist to the public as then you start entering ITAR territory.

[Patrike]

When i was working for sweden FCC (PTS) i was on a lot of interference cases with mobile phones. And my experiance was that best for locating signal on mobile bands was a directional antenna and a real time spectrum analyzer. Test was done with DF equipment. But result was poor. So fastes and easier was portabel spectrum analyzer in real time mode and handheld direction antenna.

[Psykeman]

Thank you so much for answering my question , now everything is more clear.

It happens that I’ve another question now : How does MUSIC improve the Rayleigh resolution by a very approximate factor of 10?

I’ve checked it out on the KrakenSDR wiki but it wasn’t furthermore explained.

Best Regards,
Masanni

[krakenrf_carl]

That’s really a topic for a textbook on the theory and math behind the MUSIC algorithm. Factor of 10 is just a very rough rule of thumb guess.