多速率马尔科夫跳变系统的序贯融合估计  

Sequential fusion estimation for Markov-jump systemswith multi-rate sampling

作　　者：张洋洋 林红蕾 田甜 ZHANG Yangyang;LIN Honglei;TIAN Tian(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》2023年第6期730-738,共9页Journal of Natural Science of Heilongjiang University

基　　金：国家自然科学基金(61903128);中国博士后基金(2020M670938);黑龙江省自然科学基金(LH2020F045);黑龙江省博士后基金(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(JCL202101)。

摘　　要：对带噪声相关的马尔科夫跳变系统,提出了在多传感器多速率均匀采样情形下的序贯融合估计算法。系统中状态均匀更新,不同传感器以状态更新周期的正整数倍进行观测采样,同时不同传感器间的观测噪声不仅同时刻相关,而且与过程噪声相关。根据状态更新点处的传感器观测采样情况,建立状态更新点的单速率状态空间模型。对转化后的单速率马尔科夫跳变系统状态进行扩维,进而将对原系统状态的估计等价地转化为对扩维后状态的估计,在此基础上,提出了在线性最小方差(Lineor minimum variance sense,LMSE)意义下的最优集中式融合估计算法和最优序贯融合估计算法。所提出的序贯融合估计算法通过对传感器观测数据的实时处理,可获得状态的实时融合估值。同时证明了集中式融合估计和序贯融合估计的等价性。仿真实验进一步验证了所提算法的有效性。For Markov-jump systems with correlated noises,the sequential fusion estimation algorithm is proposed in the case of multi-rate uniform sampling.In the system,the system state is updated uniformly,and different sensors sample observation at a positive integer multiple of the state update period.Moreover,the observation noises between different sensors are correlated at the same time,and are correlated with the process noise.A single-rate state space model at the state update points is established according to the sensor observation sampling at the state update point.The state of the transformed single-rate Markov-jump system is augmented,and then the estimation for the original states is transformed into the estimation for the augmented states.The optimal centralized fusion estimation algorithm and the optimal sequential fusion estimation algorithm in the Linear minimum variance sense(LMSE)are proposed.The proposed sequential fusion estimation algorithm can obtain real-time state fusion estimators through the real-time processing of sensor observation data.At the same time,the equivalence between centralized fusion estimation and sequential fusion estimation is proved.Simulation results further verify the effectiveness of the proposed algorithm.

关 键 词：马尔科夫跳变系统 序贯融合 集中式融合 多速率采样 相关噪声 

分 类 号：TP274.2[自动化与计算机技术—检测技术与自动化装置]