按三种不同加权准则的信息融合Kalman滤波器的性能比较

Performance comparision of information fusion Kalman filters weighted by three different ways

用Kalman滤波方法,在三种不同的线性最小方差最优融合准则下分别提出两传感器按矩阵加权,按对角阵加权和按标量加权的信息融合稳态Kalman滤波器.它们可以处理带相关的输入和观测噪声和带相关的观测噪声系统.一个目标跟踪系统的仿真例子说明了它们的有效性.仿真结果表明,同按矩阵加权和按对角阵加权融合滤波器相比,按标量加权融合滤波器的精度没有明显损失,但却显著地减小了计算负担,构成一种快速信息融合估计算法,适合实际应用.

By the Kalman filtering method, under three different linear minimum variance optimal information fusion criterions, the two - sensor information fusion steady - state Kalman filters weighted by matrices, diagonal matrices, and scalars are presented, respectively. They can be used to handle systems with correlated input and measurement noises,and with correlated measurements noises. A simulation result shows that compared with the fused filters weighted by matrices and diagonal matrices, the loss of the accuracy of the fused filter weighted by scalars is not obvious, but its computational burden is obviously reduced, so that it is a fast fused filtering algorithm, and is suitable for real applications.