【摘要】 提出了一种模糊基函数网络(FBFN)辅助扩展卡尔曼滤波(EKF)的PL(伪卫星)/INS(惯性导航)紧组合导航系统多故障诊断及识别方法.分析了PL/INS紧组合导航系统观测方程的线性化误差,在此基础上利用FBFN对观测方程泰勒展开式Hessian矩阵进行学习,实际应用中,将FBEN的输出作为EKF观测方程的输入项进行滤波。最后,利用大数定律对滤波残差进行平稳性检验,实现了故障诊断以及故障参数识别。仿真结果表明,针对导航系统中多个PL信号同时发生故障的情况,此方法能有效检测故障,并能准确识别出故障参数。更多还原

【Abstract】 A method for multiple faults identification of PL(Pseudo-Satellite)/INS(Inertial Navigation System) tightly coupled navigation system based on extended Kalman filter by using fuzzy basic function network(FBFN) is presented.Firstly,the linearization error of PL/INS tightly coupled navigation system measurement function is analyzed.And then,FBFN is applied to take advantage of the Hessian matrix of the measurement function.After that,the output of the FBFN can be used as the input of the EKF.The fault detection and parameter identification are fulfilled through stationary test,which is based on the law of large numbers.The simulation results show that,in case of multiple signal faults in PL simultaneously,this method can effectively detect faults and precisely identify the parameters of failure.更多还原