网络化随机不确定系统鲁棒矩阵加权融合稳态估值器  

Robust Matrix-Weighted Fusion Steady-State Estimators for Networked Stochastic Uncertain Systems

作　　者：刘文强 李爽 陶贵丽 高志军 沈忱[1] LIU Wenqiang;LI Shuang;TAO Guili;GAO Zhijun;SHEN Chen(School of Information and Electronic Engineering,Zhejiang Gongshang University,Hangzhou 310018;College of Media Engineering,Communication University of Zhejiang,Hangzhou 31001S;Computer and Information Engineering College,Heilongjiang University of Science and,Technology,Harbin 150022)

机构地区：[1]浙江工商大学信息与电子工程学院,杭州310018 [2]浙江传媒学院媒体工程学院,杭州310018 [3]黑龙江科技大学计算机与信息工程学院,哈尔滨150022

出　　处：《系统科学与数学》2021年第8期2077-2100,共24页Journal of Systems Science and Mathematical Sciences

基　　金：黑龙江省自然科学基金项目(LH2019F035);浙江省自然科学基金项目(LQ18F030003);国家自然科学基金项目(61803148);浙江省新型网络标准及应用技术重点实验室(2013E10012)资助课题。

摘　　要：对于一类在状态转移阵和系统观测阵中带相同的状态依赖乘性噪声、带噪声依赖乘性噪声、一步随机观测滞后、丢包和不确定噪声方差的多传感器网络化系统,文章研究其鲁棒融合稳态滤波问题.应用增广方法、去随机化方法和虚拟噪声技术,系统被转化为仅带不确定噪声方差的多模型多传感器系统.根据极大极小鲁棒估计原理,并应用按矩阵加权最优融合算法,提出了鲁棒局部和矩阵加权融合稳态卡尔曼估值器(预报器,滤波器和平滑器).应用增广噪声方法、非负定矩阵分解方法和李雅普诺夫方程方法,证明了估值器的鲁棒性.给出了鲁棒局部和矩阵加权融合稳态估值器之间的精度关系.应用于弹簧-质量-阻尼系统的一个仿真例子验证了所提出方法的有效性和正确性.For a class of multisensor networked systems with same state-dependent multiplicative noises in state transition and system measurement matrices,noisedependent multiplicative noises,one-step random measurement delay,packet dropouts,and uncertain noise variances,the robust fusion steady-state filtering problem is investigated.Using the augmentation approach,de-randomization approach and fictitious noise technique,the system under study is converted into a multi-model multisensor system only with uncertain noise variances.According to the minimax robust estimation principle,and applying the optimal fusion algorithm weighted by matrices,the robust local and matrix-weighted fusion steady-state Kalman estimators(predictor,filter,and smoother) are proposed.Their robustness is proved by using the augmented noise approach,decomposition approach of non-negative definite matrix,and Lyapunov equation approach.The accuracy relations among the robust local and fused steady-state estimators are given.A simulation example with application to spring-mass-damper system verifies the effectiveness and correctness of the proposed method.

关 键 词：矩阵加权融合 鲁棒稳态估值器 一步随机观测滞后 丢包 极大极小鲁棒估计原理 

分 类 号：TN713[电子电信—电路与系统] TP212[自动化与计算机技术—检测技术与自动化装置]