基于初距划分的纯方位目标跟踪的EKF滤波器群

Group of EKF Based on First-range Partition for Bearings-only Target Tracking

利用EKF对纯方位目标跟踪的工程算法进行了探讨。目标初距范围被划分成若干个小单元,每个小单元形成预估初距;在观测器匀速直航下,EKF滤波器群分别估计目标-观测器相对速度与初始距离比值;然后每个滤波器利用新量测方位估计目标参数,同时依据Bayes-ian公式更新滤波器概率密度;观测器匀速下预估状态初始值,降低了误差影响,提高了推广卡尔曼滤波方法(EKF)收敛率;多滤波器算法便于并行计算和实时处理,符合工程要求。

Extended Kalman filter （EKF） is applied to design the engineering algorithm for bearings-only tracking. The prior range region is divided into a number of smaller cells treated as initial- range. The ratios of relative velocity between target and observer to initial-range are estimated by the group of EKF filter while observer travels with constant velocity. Utilizing a sequence of new measurements collected by a manoeuvre observer, the group of filters estimate the target motion parame- ters. Their probability of density function is updated according to Bayesian rule. The influence of error can be alleviated and the ratio of convergence can be increased due to measurements collected by the fixed observer in initial value of state. The algorithm accords with the need of engineering that can calculate real time and possesses parallel.