Better images, fewer samples; optimizing array configuration for compressed sensing in radio interferometry

As with studies of geophysical features, some of the most important tools in astronomical studies lie in imaging techniques. However, the size of a telescope lens sets a limit on image resolution, and it would be impractical to extend lens sizes indefinitely to resolve sources of longer-wavelength emission, such as radio galaxies. The solution, radio interferometry, employs arrays of many antennae and is distinguished by its application to incoherent radiation sources. By sampling an incoherent radio signal with every possible pair of antennae, where the greatest distance between antennae is the interferometer's aperture, we can collect phase information that is inaccessible with a lens. The technique has provided us the highest-resolution images we have in astronomy. Reliable image reconstruction, however, can depend on a hefty amount of interferometric data, prompting such ambitious up-and-coming projects such as the 64-antennae Atacama Large Millimetre/submillimetre Array (ALMA) and the Square Kilometre Array (SKA), which will use several thousand antennae. Meanwhile, deconvolution algorithms such as CLEAN are used to remedy the effects of incomplete sampling.