Online fault detection of HRG based on an improved support vector machine

Author

Zi-Yang Qi ; Qing-Hua Li ; Guo-xing Yi ; Yang-Guang Xie ; Hong-Tao Dang

Author_Institution

Space control & inertial Technol. Res. center, Harbin Inst. of Technol., Harbin, China

Volume

01

fYear

2013

fDate

14-17 July 2013

Firstpage

316

Lastpage

319

Abstract

An improved support vector machine (SVM) model is proposed to perform online fault detection of the navigation system with hemispherical resonator gyro (HRG). The proposed model is based on sliding window SVM prediction and least square (LS) method, which can satisfy the prediction demand of the HRG output characteristic of nonlinearity, non-determinism and randomness. The proposed model can overcome the explosion of calculation of traditional SVM method, and it also improves the prediction accuracy compared to the GM(1,1) model and BP neural network. Finally, simulations of HRG fault patterns are used to verify the correctness and effectiveness of the online fault detection model.

Keywords

fault diagnosis; gyroscopes; mechanical engineering computing; navigation; random processes; support vector machines; HRG fault pattern simulation; HRG nondeterminism; HRG nonlinearity; HRG output characteristic; HRG randomness; LS method; hemispherical resonator gyro; improved support vector machine model; least square method; navigation system; online fault detection model; sliding window SVM prediction; HRG; Least square method; Moving window; Prediction model; SVM;

fLanguage

English

Publisher

ieee

Conference_Titel

Machine Learning and Cybernetics (ICMLC), 2013 International Conference on

Conference_Location

Tianjin

Type

conf

DOI

10.1109/ICMLC.2013.6890487

Filename

6890487