多星座组合导航自适应信息融合滤波算法

The Adaptive Information Fusion Filtering Algorithm for Multi-constellation Integrated Navigation System

针对多星座卫星组合导航,提出了一种双重自适应联合卡尔曼滤波算法,采用描述机动载体运动的"当前"统计模型,首先建立一种基于载体加速度方差自适应的动态定位卡尔曼滤波模型,并分别对GPS,GLONASS和GALILEO系统设计相应的自适应子滤波器,然后采用有重置的联合自适应滤波器对各个子滤波器进行数据融合处理,各子滤波器的信息分配系数根据各卫星导航系统输出的几何精度因子(GDOP)进行自适应调节。通过对GPS/GLONASS/GALILEO多星座组合导航系统的仿真,分析对比了加权平均滤波、常规联合滤波和本文提出的双重自适应滤波。结果表明:该双重自适应算法有效提高了组合导航系统的精度和可靠性,能更好地适应于量测噪声不断变化的卫星组合导航系统。

Regarding the multi-constellation integrated navigation,a double-adaptive federated filtering algorithm is proposed in this paper.By Assuming a current statistical model for maneuvering target,a dynamic positioning Kalman filtering model is established,which is based on adaptive acceleration and variance of maneuvering target,and adaptive sub-filters for GPS,GLONASS and GALILEO system are designed respectively.Then data fusion and processing for sub-filters are implemented by means of adaptive federated filtering algorithm.The information distribution coefficients of sub-filters are adaptively adjusted according to the geometry dilution of precision（GDOP）,which are produced in real time by each satellite-navigation system.The double-adaptive filtering algorithm is applied on GPS/GLONASS/GALILEO multi-constellation integrated navigation system.The simulation results compared with weighted average filter and general federated filter show that the double-adaptive filter algorithm effectively improves the precision and reliability of integrated navigation.The algorithm proposed is more applicable for the satellite integrated navigation system in which measurement noise is time-varying.