Fast period estimation of X-ray pulsar signals using an improved fast folding algorithm

An accurate period is important to recover the pulse profile from a recorded photon event series of an X-ray pulsar and to estimate the pulse time of arrival, which is the measurement of X-ray pulsar navigation. Epoch folding is a classical period estimation method in the time domain; however, its computational complexity grows as the number of trail periods increases. In order to reduce the computational complexity, this paper improves the fast folding algorithm through segment correlation and amplitude accumulation, which is based on the post-order traversal of a binary tree. Compared with epoch folding, the improved fast folding algorithm can achieve a similar accuracy at the cost of a lower computational burden. Compared with the original fast folding algorithm, the improved algorithm can be applied to detectors with a much smaller effective area. The performance of the method is investigated by simulation data and observation data from the Neutron star Interior Composition Explorer (NICER).