具有随机时延和丢包的分布式融合估计器

Distributed Fusion Estimator with Stochastic Delay and Packet Loss

针对分布式融合框架,考虑了具有模型不确定性、传感器失效、数据传输时延和丢包的多传感器融合估计问题.其中,利用系统矩阵中的乘性随机噪声来描述模型的不确定性,利用量测方程中相互独立的二值随机变量来描述传感器失效现象,局部最优估计传送至融合中心的过程中存在着随机时延和丢包现象,在融合中心创建符合信号存储原则的时延-丢包模型.设计了一种局部增益,使最优无偏估计器的误差协方差最小,利用线性最小方差融合估计方法,推导出递推形式下的分布式融合估计器.最后,通过算例仿真有效证明了所设计融合估计器的性能.

For system with distributed fusion framework,multi-sensor fusion estimation problem with model uncertainty,sensor failure,data transmission delay and packet loss was considered. The multiplicative stochastic noise in the system matrix was used to describe the model uncertainty,and the independent two - value random variable in the measurement equation was used to describe the sensor failure phenomenon. The local optimal estimation was transmitted to the fusion center with stochastic delay and packet loss,and a delay - loss model was created in the fusion center which accords with the signal storage principle. A local gain was designed to minimize the error covariance of the optimal unbiased estimator,and a distributed fusion estimator with recursive form was deduced by using the linear minimum variance fusion estimation method. Finally,the performance of the designed fusion estimator was proved with simulation example.