Performance and Design of Radio Interferometers with large eXtension – PerDRIX

The recent development of new radio telescopes such as LOFAR or SKA, consisting of several thousands of elementary antennas, leads to consider new design approaches, especially regarding their antenna configuration. These new instruments will make it possible to achieve unprecedented imaging accuracy. The objectives of this project are twofold: (i) derive relevant imaging performance criteria for radio interferometers, and (ii) devise original configurations of antenna arrays for these instruments, based on the proposed criteria. To this end, we will use a parametric model of radio astronomical data, and we will derive lower bounds of the imaging (mean-squared) error. The Cramér-Rao bound (CRB) for the imaging problem in radio astronomy has been only little studied, whereas it is usually used as a design criterion for multi-sensor systems. We propose to fill this gap, and to extend the CRB for this problem to other version like bias-corrected, or Bayesian. In addition, the CRB appears to be insufficient to describe the estimation performance in situations of low signal-to-noise ratio and/or small amount of data. We will thus resort to other bounds than the CRB, which have wider applicability conditions in these challenging situations, commonly encountered in radio astronomy. We will then focus on antenna placement methods, e.g., a grid-based antenna selection method formulated as the minimization of a convex function. Finally, we will compare these solutions to those commonly used in radio astronomy, and we will use them for observation design using NenuFAR data.