中文摘要：

提出了一种高精度的近场和远场混合信号源定位算法．此算法利用混合源阵列流形的对称性特点，从阵列流形里分离出到达角（directionofarrival，DOA）信息，并实现对所有近场与远场信号源DOA的估计．基于得到的DOA估计值，根据近场与远场源距离参数位于不同区间的特点实现对近场及远场源的分类，以及对近场源距离参数的估计．此算法由于充分利用了数据协方差矩阵的信息，并且基于多项式根值方法形成了一个统一的DOA估计器，所以获得了一个高精度的DOA估计性能，且进一步提高了近场源range参数的估计精度．此外，此算法不需要构造高阶累积量，不需要进行二维搜索，不需要进行参数配对；所有的实现过程仅需一维搜索，计算量小，实现简便．数值及与现有算法的对比实验验证了所提出算法的有效性及优越性．

英文摘要：

A high-accuracy localization algorithm for the mixed near and far-field sources was proposed in this paper. Making the most of the symmetry of array steering vector, the proposed algorithm first separates the DOA informa- tion from the array steering vector so as to finish the DOA estimation of all near- and far-field sources. Then, based on the estimated DOAs, the near-field sources and far-field sources were distinguished from the mixed sources ac- cording to the different range intervals of the near-field sources and far-field sources, and the range parameters of the near-field sources were estimated. Because the proposed algorithm, which is based on the polynomial rooting method, makes full use of the data covariance matrix information to form a united DOA estimator for the near-field and far-field sources, it can obtain high estimation performance for both the DOA and range parameters. Besides, the proposed algorithm requires neither high-order statistics nor two-dimensional search（only one-dimensional search）. It requires no pairing parameters, either. Therefore, the proposed algorithm only requires low computational cost. Numerical experiment and comparison with other existing algorithms verify the effectiveness and superior per- formance of the algorithm proposed in this paper.