Uncertainty based Fault Type Identification for Fault Knowledge Base Generation in System of Systems

Abstract

A System of Systems (SoS) is a large-scale complex system composed of Constituent Systems (CSs) that interact organically to achieve SoS level goals that cannot be achieved by individual CSs. Various uncertainties may arise in SoS in addition to the uncertainties of each CS. Uncertainty in SoS may cause it to fail in various situations. To debug these failures efficiently, a high-quality fault knowledge base is needed for the SoS. However, existing studies assume that (1) there are sufficient data when creating an initial fault knowledge base and that (2) various uncertainties related to the characteristics of the SoS cannot be considered. Therefore, this study proposes an approach to create a fault knowledge base in SoS that considers uncertainty and then determines the quality of fault data so as to build a high-quality fault knowledge base for SoS. The proposed approach categorizes faults based on the nature of uncertainty and the manifestation location, making it possible to find and add fault types that are currently not considered. Through a case study of an Advanced Driver Assistance System (ADAS), we followed the process domain experts use to create a fault knowledge base. Extracting and classifying knowledge from more than 9,000 fault data entries revealed that only 7 out of 10 fault types were observed. The proposed approach could successfully find unknown faults that were not in the generated fault knowledge base.