A robust DOA estimation algorithm for impulsive noise based on improved phased fractional lower-order moments

Traditional DOA estimation algorithms frequently exhibit a lack of robustness, as they are typically based on the assumption of Gaussian noise, which struggles to account for the non-Gaussian characteristics and intensity of impulsive noise, leading to significant performance degradation. In order to address this limitation, this paper proposes a direction of arrival (DOA) estimation algorithm based on sign phased fractional lower-order moments (SPFLOM). This algorithm combines the sign function with phased fractional lower-order covariance estimation methods. First, the sign function normalizes the received data, and then the PFLOM operator applies, effectively suppressing impulsive noise while preserving useful information. Compared to other MUSIC algorithms based on fractional lower-order statistics (FLOS), this method exhibits stronger robustness in heavy-tailed noise environments and low signal-to-noise ratio conditions. The use of non-uniform arrays can also improve the accuracy of DOA estimation under heavy-tailed alpha distributions.Thus, to further enhance the algorithm's performance, this paper extends the SPFLOM algorithm to the nested arrays and proposes the SS-SPFLOM-MUSIC algorithm. Simulation experiments validate the broad applicability and effectiveness of this algorithm across impulsive noise environments.