带未知有色观测噪声的自校正融合Kalman滤波器

对于带未知有色观测噪声的多传感器线性离散定常随机系统, 未知模型参数和噪声方差的一致的融合估值器用递推增广最小二乘法(RELS)和求解相关函数方程得到. 将这些估值器代入到最优解耦融合Kalman滤波器中, 得出了自校正解耦融合Kalman滤波器, 并用动态方差误差系统分析(DVESA)和动态误差分析(DESA)方法证明了它收敛于最优解耦融合Kalman滤波器, 因而具有渐近最优性. 一个带3传感器跟踪系统的仿真例子说明了其有效 性.

Self-tuning fusion Kalman filter with unknown colored observation noises

For the multisensor linear discrete time-invariant stochastic system with unknown colored observation noises, the consistent fused estimators of unknown model parameters and noise variances are obtained by using the recursive extended-least-squares (RELS) method and solving the correlation function equations. Substituting them into the optimal decoupled fused Kalman filter, we obtain a self-tuning decoupled fused Kalman filter. By means of the dynamic variance error system analysis (DVESA) method and the dynamic error system analysis (DESA) method, this filter is proved to be convergent to the optimal decoupled fusion Kalman filter with asymptotic optimality. A simulation example for a targettracking system with 3 sensors shows its effectiveness.