容错导航系统的一种联合滤波器结构及其算法

A Federated Filter Structure and Algorithm for a Fault-Tolerant Navigation System

本文提出了一种采用联合卡尔曼滤波器结构的容错导航系统的滤波算法。这种在多处理机环境下并行处理的分布式滤波算法有效地解决了具有多子系统的容错导航系统的实时要求。主滤波器的状态联合算法保证了获得系统最优公共状态估值。基于联合滤波器结构,水文提出了容错导航系统的故障检测与隔离(FDI)算法,将故障检测隔离决策归结为各子系统的状态估值与基于系统特性及先验噪声模型的状态估值之间的均值一致性检验问题。FDI算法决策并隔离发生故障的子系统对整个系统性能的影响。基于公共状态量的概念,简化了FDI算法的计算。

A decentralized algorithm for the federated Kalman filter has been presented. The algorithm, which can be parallelly implemented in the environment of multiprocessors, satisfies efficiently realtime requirement of the navigation system with multi-subsystems. Common state estimates can be optimally obtained by the state fusion algorithm of the master filter. A failure detection and isolation (FDD algorithm, based on the federated structure, which turns FDI decision into testing the hypothesis on the consistency of the mean values between the estimates of the subsystem states and the propagating state of the a priori model, has also been presented. The FDI decision can be made to isolate the influence of the fault subsystem on the entire system. The FDI algorithm has been simplified by means of the common state concept.