协方差交叉融合Kalman预报器

Kalman predictor based on covariance intersection fusion

对于带未知局部预报误差互协方差的两传感器跟踪系统,通过协方差交叉融合方法,得到了协方差交叉融合稳态Kalman预报器,并用协方差椭圆的方法给出了其精度关系的几何解释。用相关方法证明了协方差交叉融合稳态Kalman预报器的精度高于每个局部稳态最优Kalman预报器,低于按矩阵加权融合稳态最优Kalman预报器。用一个Monte-Carlo仿真例子说明了协方差交叉融合稳态Kalman预报器的精度接近于稳态最优融合Kalman预报器。

By using the covariance intersection （CI） fusion predictor is presented for the two-sensor tracking system with method, the covariance intersection fusion steady-state Kalman unknown cross-covariances between the local predicting errors. The geometric interpretation of accuracy relations is given by means of the covariance ellipses. The relevant methods demon- strate that the accuracy of convariance intersection fusion Kalman predictor is higher than that of each local steady-state optimal Kalman predictor, and is lower than that of the matrix weighted fusion steady-state optimal predictor. A Monte-Carlo simulation example shows that its accuracy is close to that of the steady-state optimal fusion Kalman predictor.