Research on false alarm identification method considering BIT test threshold

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Since 1970s, Built-In Test (BIT) technology was introduced to the field of avionic system. However, high False Alarm Rate (FAR) is the main element to restrain the popularization of BIT. In this paper, the method of false alarm identification considering BIT test threshold is proposed. System’s operational state is divided into three states: normal, false alarm and fault. In order to find the character­istics of each state division, Hidden Markov Model (HMM) and Self-Organizing Map (SOM) algorithm is used for training and processing. Then, the thresholds are obtained as the limitation of normal, false alarm and fault according to SOM clustering result. Finally, the result of diagnosis on BIT of current equipment is based on the test threshold largely. Finally, based on the test threshold, the result of diagno­sis on BIT of current equipment is given. Different thresholds have big influence on recognition of false alarm. The effectiveness of false alarm recognition of proposed method is validated by an application of an equipment.

Original languageEnglish
Title of host publicationSafety and Reliability – Theory and Applications - Proceedings of the 27th European Safety and Reliability Conference, ESREL 2017
EditorsMarko Cepin, Radim Briš
PublisherCRC Press/Balkema
Pages805-811
Number of pages7
ISBN (Print)9781138629370
DOIs
StatePublished - 2017
Event27th European Safety and Reliability Conference, ESREL 2017 - Portorož, Slovenia
Duration: 18 Jun 201722 Jun 2017

Publication series

NameSafety and Reliability - Theory and Applications - Proceedings of the 27th European Safety and Reliability Conference, ESREL 2017

Conference

Conference27th European Safety and Reliability Conference, ESREL 2017
Country/TerritorySlovenia
CityPortorož
Period18/06/1722/06/17

Fingerprint

Dive into the research topics of 'Research on false alarm identification method considering BIT test threshold'. Together they form a unique fingerprint.

Cite this