多传感器信息融合的时间不确定性

分析了测量融合方案的测量信息时间不确定性，将时间不确定性转化为系统不确定性．以跟踪问题为例，讨论了卡尔曼滤波器、确定性系统的鲁棒滤波器和参数不确定性系统的鲁棒滤波器3种滤波器的应用，给出了3种滤波器的求解结果，并对其表现进行了比较．仿真结果表明，在时间不确定性条件下，3种滤波器的性能都有所下降，2种鲁棒滤波器的滤波误差的方差始终比卡尔曼滤波器的方差小。

Uncertainty in the Time-Tagging of Multisensor Data Fusion

The problem of time-tagging error of multisensor data fusion strategies was investigated. It is usually encountered with the data collected by sensors having different sample rates, which is caused by processor delays and sample-to-sample timing variations. Kalman filter is a common method of dealing with multisensor data fusion problems. But Kalman filter is sensitive to uncertainty in the system parameters. The filtering problems for uncertainty in the time-tagging is converted into the filtering problems for parametric uncertainty. The design and implementation of three kinds of fusion filters, Kalman filter, robust filter and robust filter with parametric uncertainty, were described. A velocity-acceleration tracking filter is used to illustrate the success of robust filtering in the fusion of data sampled with time-tagging uncertainty. The analysis of the fused state estimate covariances of the three measurement fusion methods shows that the robust filtering scheme is insensitive to uncertainty in the system parameters, while the Kalman filter is degraded.