An Efficient Maximum-Likelihood-Like Algorithm for Near-Field Coherent Source Localization
Cheng Cheng, Songyong Liu, Hongzhuang Wu, Ying Zhang
Abstract
This communication presents an efficient iterative approach for locating the near-field (NF) coherent sources. In each iteration, the covariance matrices only containing single source information are constructed by using alternating oblique projection. Then based on the principle of vector dot products, a new iterative direction-of-arrival (DOA) estimator is proposed. After the DOA of each separated signal is estimated, the paired ranges are obtained from the 1-D maximum likelihood (ML) estimator. The proposed algorithm avoids high-dimensional spectral searches, subspace extraction, and any preprocessing such as spatial smoothing, leading to low computational complexities and high estimation accuracy. Comparative simulations show the efficiency and merits of the proposed method.