On the blind detection of FRBs through spatial fourier transforms

This paper presents two methods for the blind detection of Fast Radio Bursts (FRBs). The methods are applied to filter bank data (also known as a spectrogram or (t, f) plane). The methods are designed to detect FRB signatures in the 1D Fourier Transforms (1DFFT) and in Spatial Fourier Transforms (2DFFT) of the (t, f) data. The signature of a FRB in the 2DFFT plane is a line with a slope between zero and 90° (but not zero or 90°) in the range of low spatial frequencies. In the sum of magnitudes of the 1DFFT of each frequency channel in the (t, f) plane, the signature of a FRB appears as a relatively strong and approximately Gaussian bell shaped pulse centered around zero frequency. The detection methods are different approaches to detect these signatures. To detect the line in the 2DFFT, we use a Hough transform. To detect the presence of an FRB signature in the 1DFFT approach, we evaluate the total power in the data and compare it to a decision threshold. The paper presents a performance analysis of each detection method.