{"title":"Space Telemetry Anomaly Detection Based on Statistical PCA Algorithm","authors":"B. Nassar, W. Hussein, M. Mokhtar","volume":102,"journal":"International Journal of Electronics and Communication Engineering","pagesStart":637,"pagesEnd":646,"ISSN":"1307-6892","URL":"https:\/\/publications.waset.org\/pdf\/10002768","abstract":"The critical concern of satellite operations is to ensure\r\nthe health and safety of satellites. The worst case in this perspective\r\nis probably the loss of a mission, but the more common interruption\r\nof satellite functionality can result in compromised mission\r\nobjectives. All the data acquiring from the spacecraft are known as\r\nTelemetry (TM), which contains the wealth information related to the\r\nhealth of all its subsystems. Each single item of information is\r\ncontained in a telemetry parameter, which represents a time-variant\r\nproperty (i.e. a status or a measurement) to be checked. As a\r\nconsequence, there is a continuous improvement of TM monitoring\r\nsystems to reduce the time required to respond to changes in a\r\nsatellite's state of health. A fast conception of the current state of the\r\nsatellite is thus very important to respond to occurring failures.\r\nStatistical multivariate latent techniques are one of the vital learning\r\ntools that are used to tackle the problem above coherently.\r\nInformation extraction from such rich data sources using advanced\r\nstatistical methodologies is a challenging task due to the massive\r\nvolume of data. To solve this problem, in this paper, we present a\r\nproposed unsupervised learning algorithm based on Principle\r\nComponent Analysis (PCA) technique. The algorithm is particularly\r\napplied on an actual remote sensing spacecraft. Data from the\r\nAttitude Determination and Control System (ADCS) was acquired\r\nunder two operation conditions: normal and faulty states. The models\r\nwere built and tested under these conditions, and the results show that\r\nthe algorithm could successfully differentiate between these\r\noperations conditions. Furthermore, the algorithm provides\r\ncompetent information in prediction as well as adding more insight\r\nand physical interpretation to the ADCS operation.","references":"[1] D.L. Iverson, R. Martin, M. Schwabacher, L. Spirkovska, W. Taylor,\r\nR. Mackey, and J.P. Castle, \u201cGeneral Purpose Data-Driven System\r\nMonitoring for Space Operations,\u201d in Proc. of AIAA Infotech @\r\nAerospace Conference, Seattle, WA, October 2010.\r\n[2] D.L. Iverson, \u201cData Mining Applications for Space Mission Operations\r\nSystem Health Monitoring,\u201d NASA Ames Research Center, Moffett\r\nField, California, 94035 Space Operations Conference, 2008.\r\n[3] T. Yairi, M. Inui, A. Yoshiki, Y. Kawahara, and N. Takata, \u201cSpacecraft\r\nTelemetry Data Monitoring by Dimensionality Reduction Techniques,\u201d\r\nin Proc. SICE Annual Conference, Japan, 2010.\r\n[4] I. Verzola, A.E. Lagny, and J. Biswas, \u201cA Predictive Approach to\r\nFailure Estimation and Identification for Space Systems Operations,\u201d in\r\nProc. 13th international conference on space operations, Pasadena,\r\nCalifornia, USA, May 2014.\r\n[5] J. MacGregora, A. Cinarc, \u201cMonitoring, fault diagnosis, fault-tolerant\r\ncontrol, and optimization: Data-driven methods,\u201d Journal of Computers\r\nand Chemical Engineering, vol. 47, June 2012.\r\n[6] J. Peng, L. Fan, W. Xiao, and J. Tang, \u201cAnomaly Monitoring Method\r\nfor Key Components of Satellite,\u201d Scientific World Journal, vol. 2014,\r\nArticle ID 104052, January 2014.\r\n[7] S. Lindsay, and D. Woodbridge, \u201cSpacecraft State- of-health (SOH)\r\nAnalysis via Data Mining,\u201d in Proc. 13th international conference on\r\nspace operations, Pasadena, California, USA, May 2014.\r\n[8] S. Wold, K. Esbensen, and P. Geladi, \u201cPrincipal component analysis\u201d,\r\nChemometrics and intelligent laboratory systems, vol. 2: 37-52, 1987.\r\n[9] J.E. Jackson, A User\u2019s Guide to Principal Components. Wiley, New\r\nYork, 1991.\r\n[10] S. Wold, M. Sjostrom, and L. Eriksson, \u201cPLS-regression: a basic tool of\r\nChemometrics\u201d, Chemometrics and intelligent laboratory systems, vol.\r\n58:109-130, 2001.\r\n[11] L. Simar, and W. Hardle, Applied Multivariate Statistical Analysis. Tech\r\nmethod and data technologies. Springer-Verlag, Berlin and Louvain-la-\r\nNeuve, 2003.\r\n[12] P.R. Goulding, B. Lennox, D.J. Sandoz, K.J. Smith, and Marjanovic,\r\n\u201cFault detection in continuous processes using multivariate statistical\r\nmethods,\u201d International journal of systems science, vol. 31(11), pp.\r\n1459-1471, 2000.\r\n[13] T.K. Ralston, G. DePuy, and J.H. Graham, \u201cComputer-based monitoring\r\nand fault diagnosis a chemical process case study,\u201d Instrument Society of\r\nAmerica Transactions ISA, vol. 40, pp. 85-98. E, USA, 2003.\r\n[14] Y. Zhan, and V.Makis, \u201cA robust diagnostic model for gearboxes\r\nsubject to vibration monitoring,\u201d Journal of Sound and Vibration, vol.\r\n290, pp. 928\u2013955, 2006.","publisher":"World Academy of Science, Engineering and Technology","index":"Open Science Index 102, 2015"}