非线性离散系统基于扩展Kalman滤波的鲁棒融合状态估计  

Robust fusion state estimation based on extended Kalman filtering for nonlinear discrete-time systems

作　　者：马光鹏 孙小君[1,2] MA Guang-Peng;SUAN Xiao-Jun(School of Electronic Engineering,Heilongjiang University,Harbin 150080,China;Key Laboratory of Information Fusion Estimation and Detection,Heilongjiang Province,Harbin 150080,China)

机构地区：[1]黑龙江大学电子工程学院,哈尔滨150080 [2]黑龙江省信息融合估计与检测重点实验室,哈尔滨150080

出　　处：《黑龙江大学工程学报》2021年第1期75-81,共7页Journal of Engineering of Heilongjiang University

基　　金：国家自然科学基金项目(61104209)。

摘　　要：非线性系统的状态估计问题是现代控制理论研究中不可或缺的一部分。提出一种基于扩展Kalman滤波的鲁棒融合状态估计算法,它是在已有的扩展Kalman滤波算法的基础上,通过引入CI状态融合来实现的。融合算法的引入能有效填补扩展Kalman滤波器在模型线性化时所带有的精度损失;CI融合算法因为避免了局部估计误差互协方差的计算而变得简单易行,尤其便于实际工程应用。仿真实验结果表明了所提出的多传感器非线性离散系统的CI融合状态估计算法可有效减少估计误差,具有一定的有效性和可行性。State estimation problem for nonlinear system is an indispensable part of modern control theory.A robust fusion state estimation algorithm based on extended Kalman filter is proposed.It is implemented by introducing CI state fusion algorithm on the basis of existing extended Kalman filter algorithm.The introduction of fusion algorithm can effectively fill the accuracy loss of extended Kalman filter in model linearization.CI fusion algorithm becomes simple because of avoiding calculation of the local estimation error cross covariance,especially for practical engineering applications.The simulation results also show that the proposed CI fusion state estimation algorithm for multi-sensor nonlinear discrete-time systems can effectively reduce the estimation errors,and has certain effectiveness and feasibility.

关 键 词：信息融合 鲁棒融合 CI融合 扩展Kalman滤波 非线性系统 

分 类 号：TP271.72[自动化与计算机技术—检测技术与自动化装置]