Parameter estimation for chirp signals using the spectrum phase

Abstract

In this study, the authors propose for the parameter estimation of chirp signals at low signal-to-noise ratios using the spectrum phase instead of the traditional temporal phase. First, the spectrum phase expression of chirp signals is derived, which demonstrates that the spectrum phase can be approximated as a quadratic function of the frequency. However, the spectrum phase is a wrapped phase modulo 2π, like the temporal phase. To obtain the unwrapped spectrum phase, the authors develop a robust phase unwrapping method and then perform linear regression on the unwrapped spectrum phase to obtain the estimates of the parameters. Finally, the estimates are refined by the O'Shea refinement strategy to reduce bias. The simulation results show that the proposed technique has a lower signal-to-noise ratio threshold and is less complex than the existing methods.