粒子滤波在机动飞行器轨道确定中的应用  

Orbit determination based on particle filtering for maneuvering target

作　　者：林健[1] 林晓辉[1] 曹喜滨[1] 

机构地区：[1]哈尔滨工业大学卫星技术研究所,黑龙江哈尔滨150080

出　　处：《控制理论与应用》2008年第5期811-814,共4页Control Theory & Applications

摘　　要：对于机动飞行器轨道确定,由于状态误差及测量误差的分布不是高斯分布,传统的滤波方法受到了一定的限制,为了解决这个问题,本文采用"采样–重要性–采样"粒子滤波算法,在考虑只有目标位置信息而需对目标速度信息进行估计的情况,对含J2摄动项机动目标轨道信息进行确定.滤波过程中,结合交互式多模型方法,由多模型输出误差判断当前模型类型,通过当前模型的快速切换,提高滤波精度.最后给出应用此方法进行轨道确定的仿真实例.The non-Gaussian distribution of state error and measurement error is an important element for accurately modeling the dynamics in satellite orbit determination problem, causing limitations in applying the traditional filtering methods. The sampling-importance-resampling（SIR） particle filter is used in this paper to solve the mentioned problem. The dynamics including the J2 perturbation is considered in the design. To follow the target＇s abrupt motion, a multimodel method is used. The residual error is also applied to detect the maneuvering of the target. The model of the target is switched to the current thrust model when the residual error intersects the model aggregates. Finally, a simulation is given to estimate the orbit of a maneuvering target.

关 键 词：粒子滤波 机动目标 多模型 轨道确定 

分 类 号：V412.41[航空宇航科学与技术—航空宇航推进理论与工程]