Pseudo-Power Spectrum Estimation Method Based on Fractional Low-Order Covariance Matrix

摘要

Minimum covariance spectrum (MCS) estimation method is not robust in alpha stable distribution because of no finite covariance. A pseudo-power spectrum estimation method based on fractional low-order covariance (FLOC) matrix is proposed in this paper to improve MCS estimation method. The pseudo-power of a frequency point can be obtained by using FLOC matrix, and its pseudo-power spectrum (PPS) can be calculated by filter banks. Simulation shows that the proposed method is robust and has the capability of resolution better than MCS estimation in alpha stable distribution signal or noise.