Parameter Estimation of Polynomial Phase Signals Using Spectrum Distinction

The spectrum phase unwrapping algorithm is a recently proposed algorithm for estimating parameters of polynomial phase signals (PPS). It can handle special PPSs known as strictly monotonic PPSs (SMPPS) with a low computational load. However, due to the use of phase differentiation, this method cannot be used for non-SMPPS when signal-to-noise-ratio (SNR) is low. We analyze the spectrum of SMPPS and non-SMPPS. By considering the spectrum distinctions between them, we propose to extend the spectrum phase unwrapping algorithm by replacing the phase differentiation procedure with a one-dimensional search algorithm to handle general PPS and non-polynomial phase signals. Simulation shows that this modification can improve the algorithm's SNR threshold for general PPS and non-polynomial phase signals without considerably increasing the computational complexity.