基于分量信息融合的扩展卡尔曼滤波算法  

The Component Fusion Extend Kalman Filtering Algorithm

作　　者：耿越 雷旭升[2] GENG Yue;LEI Xu-sheng(Chengdu Aircraft Industrial ( Group) Co.,Ltd,Chengdu,610031 China;School of Instrument Science and Opto-Electronic Engineering,Beihang University,Beijing 100191,China)

机构地区：[1]航空工业成都飞机工业(集团)有限责任公司,成都610031 [2]北京航空航天大学仪器科学与光电工程学院,北京100191

出　　处：《计算机仿真》2019年第7期239-243,共5页Computer Simulation

摘　　要：研究非线性系统的最优线性化问题,借助误差传播方程中各分量信息融合的极大似然方法,得到最优线性化的分量信息融合扩展卡尔曼滤波算法,相比经典扩展卡尔曼滤波等其它非线性滤波方法,分量信息融合扩展卡尔曼滤波算法增加了状态转移矩阵和量测矩阵的更新方程,能够自适应调节参数,具有跟踪原系统的能力,能有效地提高非线性系统状态估计的精度、快速性和鲁棒稳定性。同时,分量信息融合扩展卡尔曼滤波算法的计算量与经典EKF算法接近,易于在实际工程中应用。In this paper, the Component Fusion extended Kalman filtering ( CFEKF) algorithm is proposed to obtain the optimal linearization of nonlinear systems through the maximum likelihood information fusion method of each error propagation component. Compared with the classical extended Kalman filtering (EKF) and other nonlinear filtering method, the CFEKF algorithm can increase the updating process of state transition matrix and measurement matrix, and the linearized model can adjust parameters adaptively. With the ability to track the original system, it can fundamentally improve the precision, rapidity and robust stability of nonlinear state estimation. The calculation amount of CFEKF is close to the classical EKF algorithm, so it is easy to be applied in practical engineering.

关 键 词：分量信息融合 误差传播 最大似然 扩展卡尔曼滤波 

分 类 号：TP13[自动化与计算机技术—控制理论与控制工程]