异步马尔科夫跳变系统的分布式融合估计  

Distributed fusion estimation for asynchronous Markov jump systems

作　　者：张蕊 林红蕾 ZHANG Rui;LIN Honglei(College of Electronic Engineering,Heilongjiang University,Harbin 150080,China)

机构地区：[1]黑龙江大学电子工程学院,哈尔滨150080

出　　处：《黑龙江大学自然科学学报》2024年第1期116-126,共11页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金资助项目(61903128);黑龙江省自然科学基金资助项目(YQ2022F016);中国博士后科学基金资助项目(2020M670938);黑龙江省博士后科学基金资助项目(LBH-Z19091);黑龙江省普通本科高等学校青年创新人才培养计划项目(UNPYSCT-2020001);黑龙江大学杰出青年科学基金资助项目(JCL202101)。

摘　　要：对带有相关噪声的多传感器异步马尔科夫跳变系统,提出了线性最小方差意义下的按矩阵加权分布式融合估计算法。系统中不同传感器间测量噪声相关,并与系统噪声同时刻相关,同时不同传感器以不同的采样速率均匀采集观测数据。在状态更新点利用伪观测法,并引入Dirac函数,通过状态增广,将异步马尔科夫跳变系统转化为状态更新点的单速率系统。进而利用Kalman滤波理论,提出了线性最小方差意义下的最优局部状态滤波器、系统噪声估值器和两传感器间的估计误差互协方差矩阵,提出了多传感器按矩阵加权分布式最优融合估值器。通过跟踪系统数值仿真实例,进一步验证了所提算法的有效性。For multi-sensor asynchronous Markov jump systems with correlated noises,the distributed matrix-weighted fusion estimation algorithm is investigated in the linear minimum variance sense.The measurement noises from different sensors are cross-correlated,and correlated with the process noise at the same time step,while the sampling rates of the sensors are different.By using the dummy measurement method and introducing the Dirac function,the asynchronous Markov jump systems are transformed into single rate system at the state update points by state augmentation.Based on the Kalman filter theory,the optimal local filter,the system noise estimator and the estimation error cross-covariance matrices between any two sensors are proposed in the linear minimum variance sense.The distributed fusion estimator is proposed based on the matrix-weighted fusion estimation criterion.The effectiveness of the proposed algorithms is further verified by numerical simulation of the tracking system.

关 键 词：分布式融合 马尔科夫跳变系统 异步采样 噪声相关 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]