广义系统多传感器信息融合Kalman状态预报器

Multi-sensor information fusion Kalman state predictor for descriptor systems

研究了广义离散随机线性系统的多传感器信息融合状态估计问题。依据Kalman滤波理论,在线性最小方差信息融合准则下,推导出广义系统分别按矩阵、对角阵和标量加权的三种多传感器信息融合Kalman预报器,并给出了两个广义子系统之间的预报误差互协方差阵的计算公式。仿真说明,融合后的Kalman预报器的精确度高于每一个子系统;按标量加权信息融合Kalman预报器与按矩阵加权和对角阵加权信息融合Kalman预报器相比,虽然精确度有所降低,但损失不明显,并可减少计算负担,便于实时应用。

Multi-sensor information fusion state estimation problem for descriptor discrete-time stochastic linear systems is studied. Based on Kalman filtering theory and linear least variance information fusion criterion, three kinds of multi-sensor information fusion Kalman state predictors weighted by matrices, diagonal matrices and scalars are given. The computation formula for the prediction error cross-covariance matrix is given between any two descriptor subsystems. A simulation example shows that the accuracy of fused Kalman predictors is improved. And the accuracy of information fusion Kalman predictor weighted by scalars is less than those weighted by matrices and diagonal matrices, but the computation burden is reduced greatly and it is suitable for real- time application.