Pulsar signal detection and recognition

Abstract

Pulsars are fast rotating neutron stars that emit radio waves. As it concerns received signals ranged from -90dB to -40dB, the signal-to-noise ratio (SNR) is very low. This is the main limitation regarding the use of pulsar signals in practice. In the current paper, we offer an innovative complex algorithm for detection and recognition of a pulsar signal, which in turn contains several basic algorithms: signal denoising algorithm using Wavelet transform, jumping average filter, Constant False Alarm Rate (CFAR) detection, parameter estimation, epoch folding and recognition algorithm. We have achieved partial verification, by considering a real pulsar signal from pulsar B0329+54, obtained by the Radio Telescope in Westerbork, The Netherlands. We hence argue that the combination of those algorithms can improve the signal-to-noise ratio, detection and recognition probability, which is the most important finding and contribution of the current study.