基于箱粒子滤波的再入弹道目标跟踪  

Ballistic Reentry Target Tracking Based on Box Particle Filter

作　　者：倪鹏[1] 刘进忙[1] 刘昌云[1] 李振兴[1] 

机构地区：[1]空军工程大学防空反导学院,陕西西安710051

出　　处：《现代防御技术》2016年第4期72-80,100,共10页Modern Defence Technology

基　　金：国家自然科学青年基金项目(61102109)

摘　　要：弹道目标在再入段运动方式的不确定性和复杂性导致了跟踪问题呈现非线性、不精确性。为此,引入一种"广义粒子滤波"算法——箱粒子滤波算法对再入段的弹道目标进行跟踪。该算法有别于传统点量测和误差统计模型,采用基于区间分析的箱粒子来处理这类不精确性,更加符合实际系统的工作情况,且克服了粒子滤波因需大量粒子拟合带来的实时性差的问题。仿真实验将新算法与粒子滤波算法和无迹卡尔曼滤波算法进行了对比。仿真结果表明,在确保了跟踪精度的前提下,新算法计算效率更高,是个很有效的跟踪再入目标的非线性滤波算法。The problem of ballistic reentry target tracking appears non-linear and inaccurate is caused by the uncertainty and co problem, the box particle mplexity of ballistic target movement in reentry phase. In order to resolve this filter （BPF） as a ＂generalized particle filter＂ is adopted to track ballistic reen- try target. Different from the traditional measurements and error statistical model, BPF uses the box parti- cles, based on interval analysis, to deal with this inaccuracy problem. As a result, it is more consistent to the working condition of the practical system. What＇s more, it avoids the real-time problem of particle filter （PF） using lots of particles to fit posterior probability distribution. Simulation of the performance of BPF, PF and unscented Kalman filter （UKF） proves that BPF computes more efficiently with a high tracking accuracy. Therefore, BPF is an effective non-linear filter algorithm for the ballistic reentry target tracking problem.

关 键 词：目标跟踪 箱粒子滤波 再入弹道 区间量测 粒子滤波 区间分析 

分 类 号：TJ761.3[兵器科学与技术—武器系统与运用工程] TN953[电子电信—信号与信息处理]