基于权衡因子的联邦滤波信息分配系数研究

Research on Information Distribution Coefficient of Federated Filter Based on Trade-off Factor

组合导航系统的融合算法普遍采用运行速度快、实时性强、计算量小的联邦滤波算法。针对该算法中当前的信息分配原则无法同时兼顾系统滤波精度及容错性的缺陷,采用了一种基于权衡因子的自适应信息分配方法。通过各子系统的误差协方差及量测噪声方差分别计算出能够提高系统滤波精度和容错性的信息分配系数,将各子系统的故障概率归一化得出该子系统的权衡因子,并在权衡因子的作用下自适应调节上述两种信息分配系数所占的比重,达到同时兼顾系统滤波精度和容错性的目的。仿真结果表明该方法能够减小系统的融合误差,保证系统的工作性能及融合精度。

The fusion algorithm of integrated navigation system generally adopted a federated filtering algorithm with fast running speed, strong real-time performance, and low computational complexity. For the current information distribution principle in the algorithm couldn't take into account the defects of system filtering accuracy and fault tolerance at the same time, an adaptive information distribution method based on the balancing factor was adopted. Through the error covariance and measurement noise variance of each subsystem, the information distribution coefficient that could improve the filtering accuracy and fault tolerance of the system was calculated. The failure probability of each subsystem was normalized to obtain the balancing factor of the subsystem. Under the balancing factor, the proportion of the above two information distribution coefficients was adjusted adaptively, so as to achieve both the accuracy and fault tolerance of the system. Simulation results showed that the proposed method successfully reduced the fusion error of the system and guaranteed the performance and fusion accuracy of the system.