Statistical discrimination of RFI and astronomical transients in 2-bit digitized time domain signals

We investigate the performance of the generalized Spectral Kurtosis (SK) estimator in detecting and discriminating natural and artificial, very short duration transients in the 2-bit sampling time domain Very-Long-Baseline Interferometry (VLBI) data. We demonstrate that, while both types of transients may be efficiently detected, their natural or artificial nature cannot be distinguished if only a time domain SK analysis is performed. However, these two types of transients become distinguishable from each other in the spectral domain, after a 32-bit FFT operation is performed on the 2-bit time domain voltages. We discuss the implication of these findings on the ability of the Spectral Kurtosis estimator to automatically detect bright astronomical transient signals of interests, such as pulsar or fast radio bursts (FRB), in VLBI data streams that have been severely contaminated by unwanted radio frequency interference.