Deep, Broad Null Formation for Canceling Moving RFI in Radio Astronomical Arrays

Radio frequency interference (RFI) is rapidly becoming a major issue for many applications. It is especially problematic for radio astronomy, where signal-to-noise ratios (SNRs) are less than unity. Antenna array systems such as phased array feeds (PAFs) and interferometric imaging arrays are able to cancel RFI through adaptive projection-based spatial notch filtering techniques. Current methods for formulating these projection operators suffer from an unfortunate trade-off. They must sacrifice integration time when calculating the sample spatial correlation matrix in order to track RFI motion, but this consequently increases sample estimation error and reduces null depth. In this work, we propose a new way to process spatial correlation matrices to form a broad null that reliably cancels moving RFI without increasing sample estimation error due to insufficient integration. Additionally, when assisted by an RFI-tracking auxiliary antenna, this approach also reduces the total data rate coming out of a spatial correlator, thus making broad-null-based RFI cancelation more computationally efficient and practical for real-time active array-based RFI mitigation systems.