Semantic Support for Log Analysis of Safety-Critical Embedded Systems

TL;DR

This paper presents semantic techniques and full text search methods to assist human experts in analyzing and classifying test logs from safety-critical embedded systems, aiming to improve diagnosis speed and accuracy.

Contribution

It introduces a semantic support framework for log analysis in safety-critical systems, enhancing fault diagnosis and requirements tracing capabilities.

Findings

Improved accuracy in fault classification.

Faster identification of relevant test logs and requirements.

Enhanced support for human experts in diagnosis process.

Abstract

Testing is a relevant activity for the development life-cycle of Safety Critical Embedded systems. In particular, much effort is spent for analysis and classification of test logs from SCADA subsystems, especially when failures occur. The human expertise is needful to understand the reasons of failures, for tracing back the errors, as well as to understand which requirements are affected by errors and which ones will be affected by eventual changes in the system design. Semantic techniques and full text search are used to support human experts for the analysis and classification of test logs, in order to speedup and improve the diagnosis phase. Moreover, retrieval of tests and requirements, which can be related to the current failure, is supported in order to allow the discovery of available alternatives and solutions for a better and faster investigation of the problem.