基于极大后验原理的非线性系统传感器故障估计  

Sensor Fault Estimation for Class of Nonlinear Systems Using Maximum Posterior Principle

作　　者：钱默抒[1,2] 姜斌[1] 杨浩[1] 高志峰[1] 

机构地区：[1]南京航空航天大学自动化学院,南京210016 [2]南京航空航天大学无人机研究院,南京210016

出　　处：《南京航空航天大学学报》2011年第B07期100-103,共4页Journal of Nanjing University of Aeronautics & Astronautics

基　　金：国家自然科学基金(61074080)资助项目

摘　　要：针对扩展卡尔曼滤波算法(Extended Kalman filter,EKF)计算复杂,粒子滤波算法动态跟踪能力差,单一无先导扩展卡尔曼滤波算法(Unscented Kalman filter,UKF)滤波精度低等缺陷,本文根据极大后验原理(Max-imum posterior principle,MPP),针对一类非线性系统设计了一种改进型的无先导卡尔曼故障估计滤波器来估计被控系统所发生的加性传感器故障。首先根据极大后验估计原理,推导出一种最优常值故障估计器。在此基础之上,推导出次优的加性常值故障估计滤波器,并对故障估计滤波器进行了无偏性证明。最后,将得到的理论结果应用于非线性倒立摆系统,仿真验证了所提方法的有效性。Owing to the computational complexity of the extended Kalman filter （EKF）, the poor dynamic tracking ability of the particle filter algorithm, and the low accuracy of the single unscented Kalman filter （UKF）, a designed approach to an improved unscented Kalman fault estimation filter is presented for a class of nonlinear systems based on maximum posterior principle （MPP）. It can be used to estimate the sensor fault that occurs in the controlled plant. Firstly, an optimal constant fault estimator is derived according to MPP. Then, a suboptimal additive constant fault estimation filter algorithm is developed. Meanwhile, the proof of the unbiasedness for fault estimation filter is given. Finally, simulation results of an inverted pendulum example show the effectiveness of the proposed approach.

关 键 词：极大后验原理 无先导卡尔曼滤波 故障估计 

分 类 号：TP277[自动化与计算机技术—检测技术与自动化装置]