量测随机延迟下带相关乘性噪声的非线性系统分布式估计
Distributed estimation for nonlinear systems with correlated multiplicative noises and randomly delayed measurements

本文提出了乘性噪声和加性噪声相关下的量测随机延迟非线性系统分布式状态估计. 在所考虑系统中, 相 关状态被多传感器簇构成的传感器网所观测. 所得理想量测被传送到远程分布式处理网, 并伴随服从一阶马尔可 夫过程的随机延迟. 在此基础上, 本文提出了分布式高斯信息滤波(distributed Gaussian-information filter, DGIF), 来 实现估计精度与计算时间的折中. 在单处理节点/单元中, 以估计误差协方差最小化为准则, 设计了相应的高斯递推 滤波, 并实现了延迟概率的在线递推估计. 进一步地, 在分布式处理网中, 基于非线性量测方程的统计线性回归, 结 合一致性算法, 给出了一种分布式信息滤波形式, 有效实现了分布式融合. 分别在单处理单元和分布式处理网中仿 真验证了所提算法的有效性.
This paper presents the distributed state estimation for nonlinear systems with randomly delayed measurements under correlated additive and multiplicative noises (NSAMD). In the considered problem, the interested state is observed by multiple sensor clusters, and the corresponding measurement data is sent to the remote distributed processing network via data transmission, along with the random delay obeying the first-order Markov chain. Then, the distributed Gaussian-information filter (DGIF) is presented to pursue a tradeoff between estimate accuracy and computation time, including a novel Gaussian filter for NSAMD with the estimated delay probability online (abbreviated as GAMDF) in the sense of minimizing the estimate error covariance in the single local processing node/unit, and a distributed information filter form to give an efficient distributed fusion via consensus strategy based on the statistical linear regression applied to nonlinear measurement equations. A numerical example is simulated to validate the proposed method in a single processing unit and the distributed processing network.