基于鲁棒正不变集的传感器故障区间估计  

Interval Estimation for Sensor Fault Based on Robust Positive Invariant Set

作　　者：张文瀚 王振华[1] 沈毅[1] ZHANG Wen-Han;WANG Zhen-Hua;SHEN Yi(School of Astronautics,Harbin Institute of Technology,Harbin 150001)

机构地区：[1]哈尔滨工业大学航天学院,哈尔滨150001

出　　处：《自动化学报》2020年第9期1986-1992,I0001,共8页Acta Automatica Sinica

基　　金：国家自然科学基金(61773145,61703296);哈尔滨工业大学深空探测着陆与返回技术国防重点学科基金(HIT.KLOF.2018.073)资助。

摘　　要：针对具有传感器故障和未知扰动与测量噪声的线性离散系统,提出了一种传感器故障区间估计方法.将传感器故障视为增广状态,原始系统转化为一个等效的广义系统.为了得到故障的点估计同时抑制扰动和噪声的影响,基于有界实引理设计了一个针对广义系统的鲁棒状态观测器.然后,通过中心对称多胞体技术实现对故障的区间估计并基于鲁棒正不变集给出了一种降低区间估计计算量的方法.最后,通过一个垂直起降(Vertical take-off and landing,VTOL)飞行器线性化模型的仿真算例验证了所提出方法的有效性与优越性.This paper proposes an interval estimation method of sensor fault for linear discrete-time system with unknown disturbances and measurement noise.A descriptor system,which is equivalent to the original system is constructed by considering the fault as an auxiliary state.Based on the descriptor system,a robust augmented state observer is designed by using bounded real lemma to obtain the point-estimation of sensor fault and attenuate the effects of disturbances and measurement noise.Then,the zonotope technique is used to realize the interval estimation of fault.Moreover,a method to reduce the computational complexity of interval estimation is proposed based on robust positive invariant set.Finally,numerical simulation of a vertical take-off and landing(VTOL)aircraft linear model is given to illustrate the effectiveness and superiority of the proposed approach.

关 键 词：故障估计 观测器 不变集 中心对称多胞体 区间估计 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]