Fuzzy fault tree analysis based on T–S model with application to INSGPS navigation system.pdf

Soft Comput (2009) 13:31–40 DOI 10.1007/s00500-008-0290-3 ORIGINAL PAPER Fuzzy fault tree analysis based on T–S model with application to INS/GPS navigation system Hua Song · Hong-Yue Zhang · C. W. Chan Published online: 26 March 2008 ? Springer-Verlag 2008 Abstract A novel fault tree analysis (FTA) technique based on the Takagi and Sugeno (T–S) model is proposed in this paper. In the proposed technique, referred to as the TS-FTA, the events in the conventional FTA can be expressed in terms of fuzzy possibilities, and the gates that represent the relations among the top event and the primary events are replaced by T–S fuzzy gates derived from the T–S model. The magnitudes of the faults in the system are expressed in term of fuzzy vari- ables. Since the proposed TS-FTA is derived from fuzzy logic and the T–S model, it can readily handle fuzzy information and uncertainties in the relationships among events. There- fore, the TS-FTA is suitable for systems where exact infor- mation on the fault probabilities of the components and the failure mechanisms are not available. The proposed TS-FTA is applied to analyze the reliability of an Inertial Navigation System and Global Position System (INS/GPS) integrated navigation system. Keywords Fault tree analysis · Fuzzy logic · T–S model · Reliability · Navigation system 1 Introduction In modern system engineering, it is important and often nec- essary to provide a reliable analysis of the safety and reli- ability of the system. Fault tree analysis (FTA) (Lee et al. H. Song (B) · H.-Y. Zhang Department of Automatic Control, Beihang University, Beijing 100083, China e-mail: songhua7421@ C. W. Chan Department of Mechanical Engineering, The University of Hong Kong, Hong Kong, China 1985; Antonio and Nelson 1999; Vaurio 2003; Huang et al. 2004) is a logical and diagrammatic method for evaluating the probability of an accident resulting from a sequence and a combination of faults and failure events, and it is widely used to evaluate