Liang Jinling, Wang Zidong, Liu Xiaohui
Department of Mathematics, Southeast University, Nanjing 210096, China.
IEEE Trans Neural Netw. 2011 Mar;22(3):486-96. doi: 10.1109/TNN.2011.2105501. Epub 2011 Feb 22.
This paper deals with the distributed state estimation problem for a class of sensor networks described by discrete-time stochastic systems with randomly varying nonlinearities and missing measurements. In the sensor network, there is no centralized processor capable of collecting all the measurements from the sensors, and therefore each individual sensor needs to estimate the system state based not only on its own measurement but also on its neighboring sensors' measurements according to certain topology. The stochastic Brownian motions affect both the dynamical plant and the sensor measurement outputs. The randomly varying nonlinearities and missing measurements are introduced to reflect more realistic dynamical behaviors of the sensor networks that are caused by noisy environment as well as by probabilistic communication failures. Through available output measurements from each individual sensor, we aim to design distributed state estimators to approximate the states of the networked dynamic system. Sufficient conditions are presented to guarantee the convergence of the estimation error systems for all admissible stochastic disturbances, randomly varying nonlinearities, and missing measurements. Then, the explicit expressions of individual estimators are derived to facilitate the distributed computing of state estimation from each sensor. Finally, a numerical example is given to verify the theoretical results.
本文研究了一类由具有随机变化非线性和测量缺失的离散时间随机系统描述的传感器网络的分布式状态估计问题。在传感器网络中，没有能够收集来自传感器的所有测量值的集中式处理器，因此每个单独的传感器不仅需要根据其自身的测量值，还需要根据其相邻传感器的测量值，按照一定的拓扑结构来估计系统状态。随机布朗运动同时影响动态装置和传感器测量输出。引入随机变化的非线性和测量缺失是为了反映传感器网络更现实的动态行为，这些行为是由噪声环境以及概率性通信故障引起的。通过每个单独传感器的可用输出测量值，我们旨在设计分布式状态估计器来逼近网络化动态系统的状态。给出了充分条件，以保证对于所有允许的随机干扰、随机变化非线性和测量缺失，估计误差系统的收敛性。然后，推导了各个估计器的显式表达式，以便于从每个传感器进行状态估计的分布式计算。最后，给出了一个数值例子来验证理论结果。
