A performance analysis of the MUSIC algorithm in the presence of amplitude and phase errors

This paper presents an analysis of the effects of amplitude and phase perturbation on the spectrum of the MUSIC algorithm. Theoretical expressions of the mean and variance of the spatial spectrum are derived, that is, we quantify the sensitivity of the MUSIC algorithm by first- and second-order statistical analysis. The expression reveals the relationship between the error of spectrum and the variance of the perturbation. The performance prediction is validated by computer simulation. The amplitude and phase perturbation decrease the peak values of the spectrum obviously and lower the resolving power, and the amplitude perturbation can result in spectrum error greater than that of phase perturbation for the case of the same variance of modeling errors.