不敏Kalman滤波(UKF)算法可以广泛用于各种目标运动的非线性估计中,传统的UKF滤波算法对于时间更新(即 一步预测),一般采用对目标运动方程进行离散化或线性化处理,其结果不可避免地产生离散化误差,当目标运动非线性 较强时,会导致跟踪误差增大,甚至无法给出正确的预测结果。文中提出的基于阿当姆斯(Adams)预估校正的UKF算法 (即Admas-UKF),很好地解决了弹道目标过顶点的跟踪外推问题,仿真结果显示,与传统的UKF算法相比,此算法提高了 跟踪外推精度,而计算时间远少于龙格库塔不敏Kalman滤波(Runge_Kutta-UKF)算法。

The unscented Kalman filter has been widely used for nonlinear estimation of many different kinds of targets ,discretization and linerization of targets'motion equation are usually been used to deal with the time updating(one-step prediction) in conventional UKF. However, this may produce discretization error inevitably and will increase the tracking error. when the nonlinearities become severe, it can not give a exact forecast result. An unscented Kalman filter tracking algorithm based on Adams predictor-corrector method (Admas-UKF) is presented, and it has been good to solve the vertex tracking of an artillery target . The results by simulation show that the algorithm has better tracking precision than conventional UKF algorithm while it costs much less time than Runge_Kutta-UKF algorithm.