系统干扰分布阵未知的GPS/SINS故障诊断算法

Fault diagnosis algorithm for GPS/SINS with unknown perturbation distribution matrix

针对系统模型不确定性、未知输入扰动,为对干扰解耦以及不依赖系统未知输入扰动分布阵先验信息,提出了系统干扰分布阵未知的GPS/SINS(global positioning system/strapdown interial navigation system)故障诊断算法。设计了MEP-UIO(model error prediction-unknow input observer)故障诊断观测器,改进了传统未知输入故障诊断观测器(UIO)假设系统未知扰动分布阵已知的不足;利用凸二次规划最优化原理,构造了关于未知扰动分布阵的目标函数,提出了满足目标函数最小的未知输入扰动分布阵的最优估计算法以及状态估计误差方差最小的故障诊断系统增益阵设计方法。仿真结果表明,提出的MEP-UIO故障诊断观测器设计算法相比传统Kalman滤波精度更高,验证了该故障诊断算法的有效性。

Aiming at the system model uncertainty and unknown input disturbances,in order to decouple the unknown input disturbances and not depend on the priori information of the system unknown input disturbance distribution matrix,a fault diagnosis algorithm for GPS/SINS （global positioning system/strapdown interial navigation system）with unknown perturbation distribution matrix is presented. The MEP-UIO（ model eror prediction-unknow input observer） fault diagnosis observer was designed. We improved the deficiency of conventional unknown input observer （UIO） that is assuming the system unknown disturbance distribution matrix is known a priori. We constructed the objective function of the unknown disturbance distribution matrix adopting the convex quadratic programming optimal principle. We also proposed the optimal estimation algorithm of the unknown input disturbance distribution matrix that meets the requirement of minimum objective function and the design method of the fault diagnosis system gain matrix that makes the variance of the state estimation error minimum. The simulation results show that the proposed design algorithm of MEP-UIO fault diagnosis observer possess better accuracy compared with traditional Kalman filter, which verifies the efficiency of the fault diagnosis algorithm.