弹道目标的变结构多模型数据滤波方法

Data Filtering Method of Variable Structure Multiple Model for Ballistic Target

针对单模型滤波方法不能实现对弹道目标全阶段的数据滤波的问题,提出一种变结构多模型(VSMM)弹道目标数据滤波方法。该方法在模型集切换算法的基础上设计了模型集合自适应的决策规则,该规则可有效地避免虚假切换,并采用变结构交互多模型递归算法作为模型集序列条件估计方法。根据弹道目标不同飞行阶段的受力特点建立相应的运动模型和模型集。该方法可用于弹道目标任意阶段的数据滤波。仿真结果表明,基于变结构多模型算法与单模型算法的位置均方根误差(RMSE)的比值在主动段、自由段和再入段分别为0.803、0.601和0.536,速度均方根误差的比值分别为0.787、0.654和0.740。与单模型数据滤波相比,该方法有效提高了各阶段弹道目标的估计精度和稳定性,并且在弹道目标飞行阶段转换时能够更快地适应。

To solve the problem that single model can't track the whole ballistic target data consecutively,a variable structure multiple model(VSMM)filter is proposed.The model group switching algorithm is chosen to solve the filtering of entire trajectory,the rule can effectively avoid false switching.The variable structure interactive multiple model is used as the model set sequence con⁃dition estimation method.The method establishes relevant motion models based on different phases of the ballistic target.The algo⁃rithm can be used to filter the data of ballistic target in any phases.Simulation results show that the ratio of VSMM filter and single model filter RMSE of position are 0.8350,0.601 and 0.536 in the boast,coast and reentry phases,respectively,and the ratio of the RMSE of speed are 0.787,0.654 and 0.740.Compared with single-model data filtering,the method has better estimation precision and stability effectively,and it can adapt faster during the switch phase of ballistic target.