改进的冗余捷联惯组故障检测与隔离主成分析法

An Improved Principal Component Analysis Algorithm for FDI of Redundant Inertial Measurement Unit

针对冗余捷联惯组故障检测问题,提出了一种改进的主成分析(Principal Component Analysis Algorithm,PCA)算法。思路是将PCA方法和奇偶空间方法的优点结合,通过PCA方法分析影响故障的主要因素,奇偶向量运算消除载体机动影响,并对奇偶矢量进行滤波。仿真结果表明随着滤波器的作用,能够检测到的故障幅值降低,提高故障检测的灵敏度。所提出的理论及方法可行,为箭载冗余惯组故障检测与隔离提供一种理论参考。

Vehicle maneuvering lead to a significant increase in the magnitude of the fault detection for redundant inertial measurement units.In this paper,an improved principal component analysis(PCA)algorithm is proposed.The idea is to combine the advantages of PCA method and parity space method,analyze the main factors affecting the fault by PCA method,eliminate the influence of carrier maneuver by parity vector operation,and filter the parity vector.The simulation results show that the magnitude of the detected fault decreases with the function of the filter and the sensitivity of fault detection is improved.The theory and method presented in this paper are feasible,which can provide a theoretical reference for fault detection and isolation of redundant inertial measurement units.