Efficient Imaging Strategies For Next-Generation Radio Arrays

The performance goals of the Square Kilometre Array (SKA) are such that major departures from prior practice for imaging interferometer arrays are required. One class of solutions involves the construction of large numbers of stations, each composed of one or more small antennas. The advantages of such a “large-N” approach are already documented, but attention has recently been drawn to scaling relationships for SKA data processing that imply excessive computing costs associated with the use of small antennas. In this paper we examine the assumptions that lead to such scaling laws, and argue that in general they are unlikely to apply to the SKA situation. A variety of strategies for SKA imaging which exhibit better scaling behaviour are discussed. Particular attention is drawn to field of view issues, and the possibility of using weighting functions within an advanced correlator system to precisely control the field-of-view.