一类多速率传感器动态系统的建模与估计

研究了一类时变线性动态系统的多速率多传感器数据融合状态估计问题.首先,在不同传感器以不同采样率对同一目标进行观测的情况下,提出了一种多速率建模方法,该方法可将多采样率的融合估计问题转化为同采样率的状态估计问题.随后,利用Kalman滤波对目标状态进行了在线估计,并利用有反馈分布式融合结构对上述估计进行了有机融合,从而获得了目标状态的最优融合估计值.该方法不需要对状态或观测进行扩维,计算量适当,保证了算法的实时性.以景象匹配辅助GPS/INS组合导航为例,在两种采样关系下,分别做了仿真,仿真结果验证了算法的有效性.

Modeling and estimation of a class of multi-rate sensor dynamic systems

The state estimation problem of a kind of time-vary linear dynamic system with the multi-rate multi-sensor data fusion was studied. Firstly, a multi-rate system modeling algorithm was presented in times of multiple sensors observing the same target with different sampling rates. So the multiple sampling rate system could be transformed to the same sampling rate system, and the fusion of multi-rate multi-sensor state estimation problem was simplified. Online estimation of the target state was done with the Kalman filter. By use of the distributed fusion structure with feedback, the estimations were fused and the optimal state estimation was generated. The augmentation of state or measurement dimensions were avoided in this presented algorithm, and the real-time property of the algorithm was guaranteed. A kind of integrated navigation systems including GPS, INS and scene matching was simulated with the different two sampling rates separately. The results show that the presented algorithm is effective.