Excising radio frequency interference using the discrete wavelet transform

Abstract

Radio frequency interference (RFI) excision methods have been developed using discrete wavelet transform (DWT) representations and wavelet selection based on optimizing entropy or energy by varying orthonormal wavelets and using a cost-based tree-search algorithm to compute the optimal sample shift. These methods have been applied to typical RFI waveforms obtained from Stanford's 150-foot parabolic antenna, and have succeeded in excluding the RFI while leaving background radio astronomy signals intact.