Parameter Joint Estimation Based on Generalized Cyclic Correlation Spectrum in Alpha-stable Distribution Noise

Abstract

Due to the fact that the parameter estimation method of traditional second-order statistics fails under alpha-stable distribution noise, this paper proposes a two-parameter joint estimation algorithm based on an optimized search strategy and the generalized cyclic correlation spectrum (CCS). At the same time, a type of normalized compression function which is similar to “S” is designed in this paper because of its ability to suppress impulse noise. Firstly, the cursory estimation of the carrier frequency is obtained by the improved frequency domain centering method based on Welch, and two thresholds are set on the frequency axis corresponding to the cursory estimation result. The thresholds are used to divide the search range of spectral peaks in the cyclic spectral section about carrier frequency and symbol rate estimation. And finally, the generalized cyclic discrete spectral line extraction algorithm is used to estimate the accurate parameter in the search range. Simulation results show that the proposed algorithm can suppress the influence of strong impulse noise, and has strong adaptability to the changes in mixed signal-to-noise ratios (MSNR), noise characteristic indexes, and the number of sample points.