Description of a RFI mitigation pipeline for the FAST 19-beam receiver

Radio Frequency Interference (RFI) is currently a common problem faced by radio telescopes with high sensitivity, and it is getting worse with the development of other man-made radio services. Spatial filtering techniques are developed by Kocz et al. (2010b) to filter out the common RFI from the data taken by a multi-beam receiver. In this paper, we describe a spatial filtering based RFI mitigation pipeline with the Asymmetrically reweighted Penalized Least Squares smoothing (ArPLS) algorithm (Baek et al., 2015), and apply it to the data taken with the Five-hundred-meter Aperture Spherical radio Telescope (FAST) 19-beam receiver. Firstly, we use ArPLS to estimate and remove the baseline from the input data for each beam. Secondly, we construct the spatial filter with the RFI projection matrices. Thirdly, we flag the contaminated data with an optimized threshold. In the end, the results are encapsulated in a mask file, a format for RFI masks in PRESTO . The effectiveness of our method has been confirmed by the PSR J0528+2200's data. Our method outperforms rfifind on true positive rate, it also supports parallelization and GPU acceleration. Our method is used to mitigate RFI in some FAST pulsar survey programs.