Article ID: 14.20161053
α-stable distribution is a kind of non-Gaussian signal, and it is more likely to exhibit sharp spikes or occasional bursts of outlying observations than one would expect from normally distributed signals. So it is difficult to accurately estimate the parameters of chirp signal in α-stable noise. In addressing this problem, a novel method of accurate parameter estimation of chirp signal is proposed. Firstly, the characteristics of α-stable noise are analysed. Secondly, the chirp signal in α-stable noise is transformed into Gaussian-like distribution. Then, fractional Fourier transform is used to calculate the parameters of the initial frequency and chirp rate. By simulation, the method is verified to be effective and rational. The proposed method can effectively estimate the initial frequency and chirp rate of the chirp signal in α-stable noise.
