基于联邦滤波器的新型故障检测结构及算法

提出了一种基于联邦卡尔曼滤波器的故障检测结构,该结构利用各局部滤波器和参考滤波器共有状态之间的残差进行故障检测.并提出了2种故障检测算法:χ2检验法和Elman神经网络检验法.以组合导航系统为例进行了仿真研究,和其它算法相比该算法计算简单、可靠,不但可以快速检测出外部传感器及参考系统故障,且具有很好的容错性能,能快速检测出故障并进行隔离,使融合后系统依然保持较高精度.

New Fault Detection Structure and Algorithm Based on Federated Filter

Based on federated Kalman filter, a new fault detection structure and algorithm was presented. The structure performs fault detection with the common states of local filters and reference filter. Chi square test and Elman neural network test algorithm were presented. As an application, comparisons for these algorithms are simple and reliable, these algorithms can detect the errors for both sensors and reference system, and have excellent fault tolerance performances, fast fault identifying and isolating ability.