Direct Position Determination of Non-Gaussian Sources for Sensor Arrays via Improved Rooting Subspace Data Fusion Method

The improved rooting subspace data fusion (IR-SDF) method of non-Gaussian (NG) sources for sensors with augmented coprime arrays (SACA) is introduced into the direct position determination (DPD) problem in this article. The higher-order cumulants used in parameter estimation for NG sources are the fourth-order cumulants which help to expand array properties. The proposed IR-SDF method which combined the weighted method and rooting method solves the problems of poor stability and high complexity of spectral peak search for subspace data fusion (SDF) method. Sensor-augmented coprime arrays enhance the degree of freedom (DOF), and the spatial smoothing method is employed to handle the augmented coprime array for DPD. The emulation results demonstrate that the performance of the IR-SDF method is superior to rooting SDF (R-SDF), SDF, Capon, maximum likelihood (ML), and two-step method. The proposed method uses the cost function of rooting the SDF method to obtain spectral peak searching estimation position so its complexity is lower than the SDF, weighted SDF method, and Capon method.