基于稳态Kalman滤波的相关观测融合方法及其功能等价性

对于带相关观测噪声和带不同观测阵的多传感器线性离散定常随机系统,用加权最小二乘（WLS）法提出了两种加权观测融合稳态Kalman滤波方法,可处理状态、白噪声和信号融合估计。基于稳态信息滤波器证明了它们功能等价于集中式融合稳态Kalman滤波方法,因而具有渐近全局最优性,且可显著减少计算负担。两个跟踪系统数值仿真例子验证了它们的功能等价性。

Correlated Measurement Fusion Methods Based on Steady-state Kalman Filtering and Their Functional Equivalence

For the multisensor linear discrete time-invariant stochastic control systems with correlated measurement noises and with different measurement matrices, two weighted measurement fusion steady-state Kalman filtering methods are presented by using the weighted least squares(WLS)method. Based on the steady-state information filter, it is proved that they are functionally equivalent to the centralized fusion steady-state Kalman filtering method, so that they have their asymptotic global optimality, and can reduced the computational burden. Two numerical simulation examples for tracking systems verify the functional equivalence.