自适应交互式多模型滤波在被动制导中的应用  

Adaptive Interacting Multiple Model Filter and its Application in Passive Guidance

作　　者：刘毅[1] 

机构地区：[1]北京电子工程总体研究所,北京100854

出　　处：《现代防御技术》2009年第2期41-45,共5页Modern Defence Technology

摘　　要：首先给出了一种更加简单的关于BOM问题的变增益函数的推导,新的增益形式比原先的形式更具有数字稳定性。然后考虑当模型具有不确定性时的被动制导仿真。当模型具有参数不确定时,一般的单模型滤波器已经不能满足制导的性能要求。采用交互式多模型算法,与修正增益扩展卡尔曼滤波器结合,并使用能实时估计量测噪声的Sage-Husa估值器,设计出一种新型自适应交互式多模型修正增益扩展卡尔曼滤波,将其应用到被动制导中,仿真结果表明该方法的优越性和实用性。First a much simpler derivation of the modified gain function is given for the BOM problem. The new form of the gain is more numerically stable than the original form. Then the passive guidance simulation is considered when the model has uncertainty. When the model has parameter uncertainty, generic single model filter can not satisfy guidance＇ s performance require. Using interacting multiple model algorithm, combined with modified gain extended Kalman filter （MGEKF）and using the Sage-Hu- sa estimator which can estimate the measurement noise in real time, a new adaptive Interacting multiple model was designed modified gain extended Kalman filter （IMMMGEKF） , then these are applied to passively guidance, simulation results demonstrate this method＇s advantage and practicability.

关 键 词：修正增益扩展卡尔曼滤波器 自适应交互式多模型 被动制导 

分 类 号：TJ765.3[兵器科学与技术—武器系统与运用工程] TN713[电子电信—电路与系统]