주성분분석을 이용한 인공위성 관성센서의 고장검출기법 연구

Abstract

This paper presents the fault detection method of the inertial sensors in the satellite attitude control system using PCA (Principal Component Analysis). Component analysis is an approach to finding the fault features from the data. PCA which is one of the pattern recognition methods detects the fault feature by projecting the high-dimensional inertial sensor data onto a lower dimensional space. In this paper, we suggest how to reduce the dimension of the sensor data to two-dimension. The method enables us to detect the faulty sensor by analyzing the pattern of projected sensor data.