3 ms·

Ultra-Wideband Spread-Spectrum can be undetectable, as the signal can be well below the noise floor in any given frequency band, and only knowing the right key

Ultra-Wideband Spread-Spectrum can be undetectable, as the signal can be well below the noise floor in any given frequency band, and only knowing the right key will allow you to perform a convolution to lift the actual signal above the noise. Without the right key, all you'd see is noise, and you wouldn't even know there was a signal. see https://www.itu.int/itunews/manager/display.asp?lang=en&year=2004&issue=09&ipage=ultrawide&ext=html https://www.itu.int/itunews/manager/display.asp?lang=en&year...

sudosysgen 5y ago
Detecting these kinds of signal is an area of active military research. The key is that if the signal comes from one direction, you can use multiple antennas and filter by time-delay/phase to find it even if it is below the noise floor. To this extent, there are methods to reliably detect signals 12dB under the noise floor, and as I said this is an area of active research and militaries are probably even further ahead.
- dognotdog 5y ago
  That is quite interesting, and makes sense! The spread-spectrum signal is a psuedo-random, but same sequence for all receivers, while the noise varies and should cancel out. Though, if I am not mistaken, the laws of super-sampling hold in this case, and SNR is only improved by O(sqrt(N)) for N samples, so this approach will have practical limits as one has to place a very large number of receivers in range of the emitter (or do auto-correlations over a very long time window), with enough separation to make sure the noise isn't correlated anymore, if looking for a signal deep below the noise floor.
  - sudosysgen 5y ago
    I'm rusty on this but I do believe such a limit exists. However thankfully for this research it's about detecting radars, so thanks to the inverse square law there's a limit to how deep beneath the noise floor you need to look.
gus_massa 5y ago
I made in https://news.ycombinator.com/item?id=28003977 https://news.ycombinator.com/item?id=28003977 a back of the envelope calculation using the laser that NASA use to measure the distance to the Moon. It's a pulsed laser that fires in a very short time, and it use a very narrow band. If you use it to communicate with other star, the light of the Sun in this band is like 10^8 times stronger. For a wide spectrum signal, you have an even higher noise level.

sudosysgen 5y ago

Detecting these kinds of signal is an area of active military research. The key is that if the signal comes from one direction, you can use multiple antennas and filter by time-delay/phase to find it even if it is below the noise floor. To this extent, there are methods to reliably detect signals 12dB under the noise floor, and as I said this is an area of active research and militaries are probably even further ahead.

dognotdog 5y ago
That is quite interesting, and makes sense! The spread-spectrum signal is a psuedo-random, but same sequence for all receivers, while the noise varies and should cancel out. Though, if I am not mistaken, the laws of super-sampling hold in this case, and SNR is only improved by O(sqrt(N)) for N samples, so this approach will have practical limits as one has to place a very large number of receivers in range of the emitter (or do auto-correlations over a very long time window), with enough separation to make sure the noise isn't correlated anymore, if looking for a signal deep below the noise floor.
- sudosysgen 5y ago
  I'm rusty on this but I do believe such a limit exists. However thankfully for this research it's about detecting radars, so thanks to the inverse square law there's a limit to how deep beneath the noise floor you need to look.

gus_massa 5y ago

I made in https://news.ycombinator.com/item?id=28003977 https://news.ycombinator.com/item?id=28003977 a back of the envelope calculation using the laser that NASA use to measure the distance to the Moon. It's a pulsed laser that fires in a very short time, and it use a very narrow band. If you use it to communicate with other star, the light of the Sun in this band is like 10^8 times stronger. For a wide spectrum signal, you have an even higher noise level.