Nested Array Processing for Quasi-stationary Distributed Sources with Reduced-Dimension Generalized ESPRIT

This paper focuses on the direction-of-arrival estimation problem for quasi-stationary coherently distributed sources with nested arrays. To tackle this issue, we propose the reduced-dimension generalized ESPRIT (RD-GESPRIT) algorithm based on nested arrays for quasi-stationary signal sources under local scattering. Firstly, we exploit the second-order statistics of the received signals and construct a virtual array. Then, features of the quasi-stationary signals are leveraged to recover the virtual output signal matrix rank. A generalized ESPRIT estimator (2D-GESPRIT) for nested array processing is introduced in this scenario. Due the fact that 2D-GESPRIT requires a computationally complex two-dimensional peak-searching procedure to estimate the parameters, we propose the RD-GESPRIT with a special virtualization strategy to lower its complexity. Through theoretical analysis, we know that the algorithm is able to obtain higher direction finding accuracy and lower complexity. Numerical simulation results prove the fact that our method outperforms other competitors with the same parameters.