Adaptive Correntropy based Correlation for DOA Estimation by using Co-prime Array in the Presence of Impulsive Noise

Finding directions-of-arrival (DOAs) of targets by using co-prime array, extensively investigated in the scenario of Gaussian disturbances, faces issues with impulsive background noise. Existing solutions, like fractional low order moment (FLOM) and correntropy, aim to construct a statistic from the received signals but they require knowledge of background noise and under-perform in highly impulsive noise environments. To address these limitations, in this paper, we develop an adaptive correntropy based correlation (ACRCO), which can operate in all impulsive noise scenarios and do not demand the knowledge of the noise characteristics. The proposed scheme utilizes the adaptive kernel size function to process the impulsive noise and constructs correntropy based covariance matrix as statistical measures of the received signals. Within the scope of the kernel size function, we create numerous pseudo-snapshots from the co-prime virtual array with different kernel size function values to enhance DOA finding accuracy and resolution. Afterwards, multiple signal classification with de-correlation preprocessing is applied for computing the DOAs. Experiment results demonstrate that the proposed ACRCO provides superior DOA estimation in the presence of impulsive noise, compared with existing algorithms.