Measuring Co-Channel Interference: How Can It Be Done?

[dervast]

Dear all,
I am trying to understand something for SNR,

I know that is the ratio of the Power of a signal divided by the Power of noise.

According to wikipedia

Signal-to-noise ratio is defined as the power ratio between a signal (meaningful information) and the background noise (unwanted signal):

    \mathrm{SNR} = \frac{P_\mathrm{signal}}{P_\mathrm{noise}}, 

where P is average power. Both signal and noise power must be measured at the same or equivalent points in a system, and within the same system bandwidth.

This is something that I was expecting to read.
MY question though is "how do we measure the noise"?
If you have a receiver and a filter you can measure how much energy is received but how can you measure how much of "noise" is in it?

I would like to thank you in advance for your help

Best Regards
Alex

 

[sophiecentaur]

Noise measurement is a very touchy subject. One way is to use a noise source of known power and then attenuate it by a known amount. Looking at the output of your receiver / whatever, you add in a known amount of noise until the output doubles. Then you know that the system noise is the same as what you have added. But you also need to know the input bandwidth of the system so that you can tell how much of your injected noise you are actually putting into the system.
Hope that helps - but you need your calibrated noise source!

 

[Studiot]

Hello SC I like your strapline motto - It should be a compulsory catechism on all science courses.

 

[dervast]

Hmm thanks for the replies ...
In telecommunications we use a lot this SNR or C/I which means that I measure the received power source against the noise. In telecommunications this 'noise' is usually what comes from the outside 'world'. How I can define that noise 'in-real-time'?

I would like to thank you in advance for your help

B.R
Alex

 

[marcusl]

You can characterize the noise of the receiver by disconnecting the antenna or source and terminating the receiver input with a resistance equal to that of the source. Now the output contains only noise. SC's procedure is one way to measure that noise, but others also exist.

This approach leaves noise coming from the sensor itself, however. (Antennas pick up thermal noise from the warm atmosphere, for example, magnetic read heads have Barkhausen noise, etc.). Sometimes doing this is easy by simply arranging for no signal to be present (point your satellite dish slightly away from the satellite, for instance, to collect sky plus antenna plus receiver noise). In other cases (radiometry, e.g.) it is difficult and requires specialized techniques.

EDIT: I was typing while you posted your reply above. It is usually sufficient in comms systems to characterize receiver noise once, since it generally doesn't change. From that point on, subtracting the receiver noise power from the total measured power gives you signal power, and then you know SNR.

C/I (carrier to interference ratio) is different, and more complicated. You need to measure an interval where the desired signal (carrier) is absent--perhaps between frames--to characterize the interference. If you know the waveform to expect in advance, you can use correlation techniques to see how much of the signal is present, then subtract it off so you are left with an estimate of interference and noise.

As a practical matter, you can often infer C/I or SINR by observing the BER (bit error rate).

 

[dervast]

You can characterize the noise of the receiver by disconnecting the antenna or source and terminating the receiver input with a resistance equal to that of the source. Now the output contains only noise. SC's procedure is one way to measure that noise, but others also exist.

This approach leaves noise coming from the sensor itself, however. (Antennas pick up thermal noise from the warm atmosphere, for example, magnetic read heads have Barkhausen noise, etc.). Sometimes doing this is easy by simply arranging for no signal to be present (point your satellite dish slightly away from the satellite, for instance, to collect sky plus antenna plus receiver noise). In other cases (radiometry, e.g.) it is difficult and requires specialized techniques.

and what if I have cochannel interference? For example in GSM networks due to the frequency reusability I might have some extra interference from an adjacent cell. I think in the C/I (SNR with no noise) tracks that. How I can find SNR or C/I in that case?

 

[marcusl]

Please be careful with C/I and SNR, they are not interchangeable. The co-channel interference signal due to reuse looks similar to your signal of interest, making it difficult to separate. Of course system performance suffers (that's why co-channel interference is so tightly controlled in wireless systems.) I don't know how or even if co-channel C/I is measured during routine operation in commercial wireless networks. Perhaps you can do it by correlation, as I mentioned. Each mobile user, once connected, broadcasts certain information that is then known to you a priori (header symbols for each frame, e.g.). You might be able to measure the properties of those symbols (essentially build an adaptive signal processor) and estimate the interference levels from there.