L_p-norm based iterative adaptive approach for robust spectral analysis

摘要

Recently, the iterative adaptive approach (IAA) has been shown to be an effective nonparametric methodology for high-resolution spectral analysis. Its main idea is to reformulate the nonlinear frequency estimation problem as a linear model whose parameters are updated iteratively according to weighted least squares. Since the derivation of the IAA is based on l_2-norm, it cannot work properly in heavy-tailed noise environment. In this paper, a generalized version of IAA is derived to provide accurate spectral estimation in the presence of impulsive noise, which replaces the l_2-norm by the l_p-norm where 1 ＜p＜2. Simulation results are included to demonstrate the outlier resistance performance of the proposed algorithm.