Abstract: To achieve the two-dimensional direction of arrival (2D-DOA) estimation of mixed signals with colored Gaussian noise, an algorithm based on fourth-order cumulant and oblique projection operator is proposed. Firstly, with the received data the algorithm constructs a fourth-order cumulant matrix slice instead of the traditional covariance matrix to suppress Gaussian color noise, and estimates the elevation and azimuth of the incoherent signal through propagator method (PM). Then a specific oblique projection operator based on orthogonal triangular (QR) decomposition is constructed, so that the received signal only contains the information of coherent signals. Finally, the elevation and azimuth of coherent signals are estimated through the combination of matrix reconstruction and PM. The algorithm can realize the 2D-DOA estimation of mixed signals in the presence of Gaussian colored noise with the estimated elevation and azimuth of the signals paired automatically. Simulations demonstrate the effectiveness of the algorithm.

Key words: direction of arrival (DOA) estimation, four-order cumulant, oblique projection operator, mixed signals, Gaussian color noise