中国空间科学技术

中国空间科学技术 ›› 2018, Vol. 38 ›› Issue (4): 11-19.doi: 10.16708/j.cnki.1000-758X.2018.0041

• 研究探讨 • 上一篇    下一篇

基于事件驱动的容积卡尔曼滤波定位算法

林浩申1，李思嘉2，刘刚1，*   

  1. 1. 火箭军工程大学，西安 710025
    2. 火箭军驻7103厂军事代表室，西安 710100
  • 收稿日期:2017-12-20 修回日期:2018-02-07 接受日期:2018-05-15 出版日期:2018-08-25 发布日期:2018-06-07
  • 通讯作者: 刘刚（1964-），男，教授，723174670@qq.com，研究方向为控制科学与工程、空间应用工程
  • 作者简介:林浩申（1982-），男，博士研究生，linhaoshen1@163.com，研究方向为空间目标跟踪与信息融合

Cubature Kalman filter based on event-triggered mechanism

LIN Haoshen１, LI Sijia２,LIU Gang１,∗   

  1. １. Rocket Force University of Engineering, Xi′an 710025,China
    2. Rocket Force Military Representative Office in Factory 7103, Xi′an 710100, China
  • Received:2017-12-20 Revised:2018-02-07 Accepted:2018-05-15 Published:2018-08-25 Online:2018-06-07

PDF (PC)

149

摘要： 以网络化非线性滤波系统为研究对象，为了平衡系统的通信率和滤波精度之间的矛盾，引入随机事件驱动（stochastic event-triggered）的思想，并在此基础上建立了基于残差检测的事件驱动（detected event-triggered）模型。针对系统的强非线性，将线性随机事件驱动滤波系统中的更新结论推广至非线性系统，推导了两种事件驱动机制在容积卡尔曼滤波（CKF）算法框架中的滤波更新过程，得到了检测事件驱动CKF（DECKF）和随机事件驱动CKF（SECKF）两种算法。最后，通过天基平台空间目标跟踪问题对算法性能进行检验。仿真结果表明，当通信率下降20.64%时，DECKF算法的位置跟踪精度和速度跟踪精度相比标准CKF仅下降了5.50%和7.74%。此外，在通信率相同的情形下，DECKF比SECKF的精度高40%以上，证明检测事件驱动模式优于随机事件驱动模式。

关键词: 网络化状态估计, 随机事件驱动, 检测事件驱动, 容积卡尔曼滤波, 空间目标跟踪

Abstract:

In order to balance the contradiction between the communication rate and the filtering precision of the networked nonlinear filtering system, two kinds of event-triggered mechanisms based on stochastic event-triggered and detected event-triggered were proposed respectively. The framework of cubature Kalman filter (CKF) was designed under two kinds of event driving mechanisms on account of nonlinear of system, and the mathematical description of the update process was deduced. Two algorithms of detected event-triggered CKF (DECKF) and stochastic event-triggered CKF (SECKF) were obtained. Finally, ECKF was applied to the space-based spatial target localization to examine the performance of itself. The results show that the position and velocity tracking accuracy of DECKF are 5.50% and 7.74% lower than traditional CKF when the communication rate had declined by 20.64%. On the other hand, the accuracy of the DECKF was above 40% higher than the SECKF under the same communication rate, which demonstrated the effectiveness of the DECKF.

Key words:

networked state estimation, stochastic event-triggered; detected event-triggered, cubature Kalman filter, spatial target tracking