Skip to content

Detector

Colm Talbot requested to merge detector into master

I added bits of functionality to the Interferometer class so that it interfaces with the PowerSpectralDensity object.

The Interferometer has new attributes:

  • power_spectral_density: an instance of the PowerSpectralDensity class
  • power_spectral_density_array: an array of values of the PSD for a user-specified frequency array
  • amplitude_spectral_density_array: as above for the ASD
  • data: as above but h(f)
  • whitened_data: data/amplitude_spectral_density_array This is what goes into the likelihood

and methods:

  • set_spectral_densities(frequency): set power_spectral_density_array and amplitude_spectral_density_array using the provided frequency array
  • set_data(frequency_domain_strain): set data given an array with frequency domain strain

While doing this I updated a bunch of other bits of code to be consistent with the new detector setup.

  • I moved some of the code from the Interferometer class into utils.py so that the methods are shorter and I think the code will have more general use.
  • I changed how the PowerSpectralDensity is initialised, you now pass it either an ASD or PSD file and it sets itself up.
  • This changed the specific interferometers we have at the bottom of the file.
  • I added a new likelihood function that interacts with the new setup of the detector.
  • The new likelihood is just a method, rather than a class.
  • The new likelihood has a few significant changes, @rory-smith can you check these make sense:
    1. I changed an overall sign in the logL.
    2. I changed an np.sum to sum, the old version was treating the model as a number rather than a frequency series.
  • @gregory.ashton I updated the SimpleSinusoidSource to follow the convention I described earlier.

Finally, I added a script that goes through:

  • initialising a set of interferometers.
  • generating noise.
  • injecting a signal.
  • calculating the likelihood.

This has wound up being quite a bit more extensive than I anticipated, so I'm not going to just push or merge the changes. Does someone want to look over this and give me a thumbs up/down.

@paul-lasky @gregory.ashton ?

Merge request reports