Prognostic Analysis and Failure Alarming of Software Reliability for GPS Navigation System

• Main Authors: Kang-Yu Liu, 劉康宇
• Other Authors: Ying-Che Chien
• Format: Others
• Language: zh-TW
• Published: 2012
• Online Access: http://ndltd.ncl.edu.tw/handle/12747347724604884774

碩士 === 中原大學 === 工業與系統工程研究所 === 100 === This study applies prognostic analysis method on software reliability and establishes software failure alarm mechanism for prevention of impact caused by software failure. As software system innovation grows in very fast manner, software with complex functions and high reliability are considered as top design objectives. The GPS navigation system is selected as a case study for this research. The study defines software failure modes, applies Fault Tree Analysis method to identify software failure causes, and establishes software operational profiles. The GPS navigation systems were tested based on different operational profiles to identify significant software performance indexes which may cause impact on product features (e.g. GPS positioning time). The selected software performance indexes are defined as software failure precursors. Cumulative degradation levels associated with the failure precursors can be defined based on software test results. This study suggested to use Sequential Probability Ratio Test (SPRT) method for building software failure alarm mechanism. The objective is to let software user has opportunity to do system maintenance prior to software failure so as to avoided impact and loss caused by the failure. In addition, a Markov Reliability Model based on system states of GPS navigation system was built for software system availability analysis corresponding on different operational profile. The analysis results can be applied for software availability assessment before the software release for market.