fBLS -- a fast-folding BLS algorithm

TL;DR

fBLS introduces a fast-folding algorithm for efficient detection of ultra-short period transiting planets in lightcurves, significantly reducing computational complexity compared to traditional methods.

Contribution

The paper presents fBLS, a novel algorithm that accelerates the search for short-period transiting planets using a fast-folding approach adapted from pulsar astronomy.

Findings

Successfully identified all known ultra-short period planets in Kepler data.

Discovered three new ultra-short period planet candidates.

Demonstrated computational efficiency over traditional BLS methods.

Abstract

We present fBLS -- a novel fast-folding technique to search for transiting planets, based on the fast-folding algorithm (FFA), which is extensively used in pulsar astronomy. For a given lightcurve with $N$ data points, fBLS simultaneously produces all the binned phase-folded lightcurves for an array of $N_p$ trial periods. For each folded lightcurve produced by fBLS, the algorithm generates the standard BLS periodogram and statistics. The number of performed arithmetic operations is $\mathcal{O}\big(N_p\cdot\log N_p \big)$, while regular BLS requires $\mathcal{O}\big(N_p\cdot N\big)$ operations. fBLS can be used to detect small rocky transiting planets, with periods shorter than one day, a period range for which the computation is extensive. We demonstrate the capabilities of the new algorithm by performing a preliminary fBLS search for planets with ultra-short periods in the Kepler main-sequence lightcurves. In addition, we developed a simplistic signal validation scheme for vetting the planet candidates. This two-stage preliminary search identified all known ultra-short planet candidates and found three new ones.