联邦滤波器在小型飞行器距离定位融合估计中的应用

为了进一步修正小型飞行器距离定位误差，给出了一种利用联邦滤波方法进行数据融合的小型飞行器定位误差修正方法，解决了由于差分GPS长时间静默影响定位精度的问题。仿真和数据回放结果表明：采用卡尔曼滤波方法可提高定位精度，采用联邦滤波方法效果更好，使用中方法可以得到飞行器更准确的距离定位，可满足多种类型的飞行器距离定位系统的精度要求。

A New Method for Correcting Location Error of Unmanned Aerial Vehicle (UAV) Using Federated Filter Data Fusion

Section 2 discusses distance location for UAV by using federated filter, which consists of radio sensor, altitude sensor and difference GPS. We considered four possible arrangements of this federated filter and finally selected the arrangement shown in Fig.1(b) because its two subfilters involving altitude sensor and difference GPS respectively are fully isolated from each other and thus make the federated filter more fault tolerant and better in real time performance. Section 3 discusses data fusion based on this federated filter. When electromagnetic environment is not bad, as is usually the case in time of peace, the difference GPS filter is working normally and simulation results in Fig.2(a) show that values estimated by data fusion based on federated filter are very close to real values and the variance in Fig.2(b) of federated filter converges to 15m. When the electromagnetic environment is very bad, as is usually the case in war zone, the difference GPS is silent for a long time, simulation results in Fig.3(a) still show that values estimated by data fusion based on federated filter are quite close to real values but the variance in Fig.3(b) of federated filter converges to 17m in Fig.3(b),bigger than the 15m in Fig.2(b). In a previous paper [5] , we used data fusion based on Kalman filter; the variance in Fig.4 of Kalman filter converges to 20m,higher than the 17m in Fig.3(b). Federated filter can give much more accurate estimation than Kalman filter.