带多随机滞后和丢包网络化多传感器系统分布式递推融合估计

Distributed Recursive Fusion Estimation for Networked Multi-Sensor Systems with Multiple Random Delays and Packet Dropouts

文章研究了带多随机滞后和丢包网络化多传感器系统的分布式递推融合估计问题.利用伯努利分布随机变量描述传感器到估值器的随机滞后和丢包现象.通过定义新的变量,将具有随机滞后和丢包的系统等价转化为随机参数化系统.给出了线性最小方差意义下的局部最优线性滤波器,推导了局部估计误差之间的互协方差阵,以及先验融合估计与局部估计之间的互协方差阵.进而,提出了线性无偏最小方差意义下的分布式递推融合滤波器.分析了算法的稳定性和稳态特性.与现有的分布式加权融合算法相比,文章提出的算法可以明显提高估计精度.仿真例子验证了算法的有效性.

This paper is concerned with the distributed recursive fusion estimation problem for the networked systems with multiple random delays and packet dropouts.Bernoulli distributed random variables are used to model the phenomena of random delays and packet dropouts from sensors to estimators.By defining new variables,the system with random delays and packet dropouts is equivalently transformed into a stochastic parameterized system.The local optimal linear filters in the linear minimum variance sense are given.Cross-covariance matrices between local filters and between the prior fusion estimator and local filters are derived.Further,a distributed recursive fusion filter is presented in the linear unbiased minimum variance sense.Stability and the steady-state property of the proposed algorithm are analyzed.Compared with the existing distributed weighted fusion algorithms,the proposed algorithms can obviously improve the estimation accuracy.A simulation example verifies the effectiveness of the algorithms.