Fast Direction Finding Using Modified Pseudocovariance Matrix

A modified pseudocovariance matrix is described which can improve the performance of Kim \´s ("Fast DOA estimation algorithm using pseudocovariance matrix," IEEE Trans. Antennas Propagat., vol. 53, no. 4, pp. 1346-1351, Apr. 2005) algorithm to realize the fast direction finding with single snapshot in coherent or uncorrelated case. By adding the backwards formulation of pseudocovariance matrix, the rank of pseudocovariance matrix is enhanced. Then a set of pseudocovariance matrices are constructed by circularly shifting the matrix. By weighted averaging, a modified pseudocovariance matrix is reconstructed whose noise is reduced. The proposed matrix has less array element loss than that of the original matrix. Moreover, this algorithm can be applied in the lower signal-noise-ratio (SNR) through noise preprocessing. The new matrix has expanded the application of high-resolution direction finding to nonuniform linear array (NLA), since traditional methods usually assume the array is uniform linear when handling coherent sources. Simulations are conducted to compare the performance of the modified method and the original method. The results illustrate that the high performance of DOA estimate could be obtained rapidly