Fourier domain excision of periodic radio frequency interference

Abstract

The discovery and study of pulsars and fast radio bursts (FRBs) in time-domain radio data is often hampered by radio frequency interference (RFI). Some of this terrestrial RFI is impulsive and bright, and relatively easy to identify and remove. Other anthropogenic signals, however, are weaker yet periodic, and their persistence can drown out astrophysical signals. Here we show that Fourierdomain excision of periodic RFI is an effective and powerful step in detecting weak cosmic signals. We find that applying the method significantly increases the signal-to-noise ratio of transient and periodic pulsar signals. In live studies we detected single pulses from pulsars and FRBs that would otherwise have remained buried in background noise. We show the method has no negative effects on pulsar pulse shape, and that it enhances timing campaigns. We demonstrate the method on real-life data from a number of large radio telescopes, and conclude that Fourier-domain RFI excision increases the effective sensitivity to astrophysical sources by a significant fraction, which can be even larger than an order of magnitude in the case of strong RFI. An accelerated implementation of the method runs on standard time-domain radio data formats and is publicly available.

Chromatic periodic activity down to 120 MHz in a Fast Radio Burst

“…All the LOFAR data were also searched offline for FRBs and periodic emission. After subbanding and RFI cleaning 44 , data were dedispersed and searched for single pulses. Candidates were clustered in DM and time, visualized, and examined by eye.…”

Low-frequency wideband timing of InPTA pulsars observed with the uGMRT

“…Using the known pulsar ephemeris (IPTA DR2: Perera et al 2019), this pipeline partially folds the data from all the sub-bands into subintegrations of 10-second duration to archive files in the TIMER format (van Straten & Bailes 2011). Radio-frequency interference mitigation was performed by RFIClean (Maan et al 2021). Before our analysis, all the reduced data were further collapsed in time to a single integration, with 64 subbands.…”

Magnetar XTE J1810-197: Spectro-temporal evolution of average radio emission

“…As mentioned above, observations after November 2020 already recorded data with 1024 coherently dedispersed sub-bands. The data from the individual epochs are then subjected to size reduction by down-sampling from 16 bits to 8 bits per sample using digifil, and excision of radio frequency interference (RFI) using RFIClean 2 (Maan et al 2021) and rfifind from the pulsar search and analysis toolkit PRESTO (Ransom et al 2002). The resultant data are folded using prepfold from PRESTO and the timing parameters from Levin et al (2019), with 1024 bins, 128 sub-bands and typically 64 sub-intervals.…”