中文摘要：

针对扩展卡尔曼（EKF）类递推方法用于固定单站无源定位时存在依赖于初始值和不稳定性等缺点,本文提出了一种利用角度和频率变化率无源跟踪的网格搜索方法,只要网格取得足够密,最终估计趋近于全局最优的最小均方估计（MMSE）。为了减少搜索的计算量,提出了利用多级搜索法将计算量控制在合理可实现的范围。仿真表明,该方法的滤波不受状态初值误差的影响,滤波收敛速度接近无初值误差时的EKF、修正协方差扩展卡尔曼滤波（MVEKF）方法。

英文摘要：

Due to the drawbacks of ＂erratic＂ and depending on initialization when extended Kalman filtering（EKF） method used in fixed observer passive location, a new passive tracking method which searched grid of bearings and frequency changing rate was put forward. If the grid space was small enough, the last estimation would approach the global least minimum mean square error （MMSE）. To minimize computation load of searching, a multi-level searching method which controls computational complexity to a reasonable and feasible level was put forward. Simulation results reviews that the filtering process of this method did not affected by state initialization error,and it＇s convergence speed reached EKF and modified covariance EKF（MVEKF）.