标量及对角阵加权的序贯估计融合算法研究

摘要

针对多传感器分布式估计融合系统,在最小化估计误差的协方差矩阵迹的准则下,采用标量加权及对角阵加权融合方法,提出了估计误差相关条件下的序贯处理式最优估计融合Kalman滤波器.该融合滤波器以两传感器估计融合算法为基础,对传感器采集信息依次进行融合计算,得到多传感器融合结果.比较两种算法与局部滤波器的估计精度,并进行了仿真.仿真结果表明了基于加权估计融合的序贯处理算法的可行性和有效性.%Under the rule of minimizing the trace of estimation error covariance matrix, the optimal estimation fusion Kalman filter weighted by scalars or diagonal matrices with sequential disposal is presented for the multiple sensors' decentralized estimation fusion system. Based on two-sensor fusion algorithm, this fusion filter computes the sensors' information successively in order to get the final multi-sensor fusion result. This paper compares the estimation precision between the fusion filter and the local filters. And the result estimated by the fusion filter is better than the local filters. The simulation shows the feasibility and validity of the estimation fusion algorithms.