Robust Spectral Analysis of Multi-Channel Sinusoidal Signals in Impulsive Noise Environments

Robust spectral analysis of the sinusoidal signals corrupted by impulsive noise poses a big challenge in signal processing community. In this paper, we address issue robust for multi-channel signals, including order detection and parameter estimation. The successive low-rank decomposition is firstly designed to extract common subspace from data matrix. Subsequently, number poles determined with model selection criterion, based on so-obtained subspace. With information, parameters outliers are jointly estimated according maximum a posteriori criterion. To find initial guess parameters, an estimator weighted linear prediction developed. Additionally, performance provided, which includes computational complexity, convergence verification estimation, asymptotic consistency detection. Simulation results demonstrate advantages proposed framework compared state-of-the-art schemes.