I. Introduction
Radio frequency interference (RFI) is a serious problem for passive spectrum users in remote sensing and radio astronomy. Time and frequency blanking, beam pattern nulling and spatial filtering, adaptive filters, and many other methods have been used for RFI mitigation [1], [2]. Traditional digital RFI mitigation schemes for phased arrays such as maximum signal-to-interference and noise beamforming and subspace projection [1] are inherently narrow band and are implemented for wideband signals using a fast Fourier transform or polyphase filterbank after sampling. The interferer must be identified and canceled separately in each subband, leading to significant computational overhead for real-time signal processing receivers. Analog phase shift beamforming does not require real time digital computation to cancel RFI, but is also limited to narrow-band signals and does easily allow computing many beams in parallel. It would be desirable if there were a compromise between the two architectures that could implement a wideband RFI cancellation beamformer in analog while preserving signals of interest for filtering by downstream digital beamformers.