Joint Estimation of the DOA and the Number of Sources for Wideband Signals Using Cyclic Correntropy

In this paper, a new direction of arrival (DOA) and the number of signals of interest (SOIs) estimation method is proposed for wideband sources in impulsive noise environments. By evaluating the cyclic correntropy function (CCF) of the received signals at a certain cycle frequency, the impulsive noise and all co-channel interferences with different cycle frequencies can be suppressed. In this approach, a theorem of the CCF is proposed, and a linear prediction (LP) model of the CCF of the array data, which is important for both DOA and number of SOIs estimations, is built by the theorem. Additionally, we introduce multiple regularization parameters into the LP model and derive an analytical expression of the maximum correntropy criterion (MCC) of the parameters for obtaining the optimal regularization parameters expression. Furthermore, by the expression of regularization parameters and the estimation error of the CCF in a limited number of snapshots, an iterative algorithm is proposed for joint estimation of the DOA and the number of SOIs with only knowledge of the cycle frequency of the SOIs.