Application of robust Capon beamforming to radio astronomical imaging

Radio telescopes based on large phased arrays form an interesting application area for array signal processing. LOFAR is a large low frequency (10-240 MHz) array consisting of 13,000 antennas grouped into 50 stations, currently under construction in the Netherlands. Data from a 60-element test station for LOFAR is available to evaluate the performance of calibration and imaging algorithms. We apply the robust Capon beamformer (RCB) to make images of the sky from measured data, and compare them to classical Fourier-based images. The RCB takes uncertainty in the calibration into account. Instead of the usual spherical uncertainty sets, we have also derived a more constrained uncertainty set specifically for imaging with the RCB. The results are images with a higher dynamic range than classical or Capon beamforming. Additional simulations confirm that the images are more accurate.