A Nyström-Based esprit algorithm for DOA estimation of coherent signals

Sample covariance matrix (SCM) and its eigenvalue decomposition (EVD) are needed in the conventional subspace based algorithm. Under normal circumstances, the computation burden is really heavy, especially for the estimation of the coherent signals which also needs to use the spatial smooth technology. In this paper, we firstly use the Nyström-Based method to obtain the signal subspace, which results in reducing the computational complexity greatly. To further reduction of the computational complexity under the completely coherent condition, we use an improved SVD method to calculate the equivalent forward and backward covariance matrices. Then we use the ESPRIT method to estimate coherent signals' direction of arrival (DOA), and we also compare the performance among the Nyström-Based algorithm, the SCM-EVD method and the classical space smooth algorithm. The simulation results show that this new method achieves almost the same accuracy with SCM-EVD based method and classical spatial smooth method with a much less computation burden.