Visualization of Test Failure Data to Support Fault Localisation in Distributed Embedded Systems within the Automotive Industry

摘要

In this thesis we present the design, developmentand evaluation of a software tool with the purpose of assisting inthe localisation of root causes of test case failures in distributedembedded systems, specifically vehicle systems controlled by anetwork of electronic control units (ECUs). Fault localisation isespecially hard in such systems due to its distributed nature, andoften organisations rely on the knowledge of in-house specialistsfor detecting and rectifying the underlying root cause of testcase failures. The study took place in-situ at the Research andDevelopment division of Volvo Car Corporation, a large automotivemanufacturer. Researchers had access to a vast number oftest execution logs from large-scale software integration testingunder a continuous integration process. The main objectives ofthe research were to develop and evaluate a data visualisationtool to support root-cause identification of failures in orderto foster a continuous feedback loop in the fault localisationprocess. Our contributions encourage the improvement of testingquality and supporting the development and adoption of test casewriting guidelines and test failure debugging procedures. Theresearch concludes that the use of data visualisation techniquescan considerably boost the failure debugging procedures bypresenting data in a clear and concise manner and making useof test harnesses to directly assist in reducing possible causes offailures. Additionally it encourages a systematic and continuousanalysis of the current state of testing by aggregating, categorisingand displaying large amounts of historical data in a concisemanner that allows stakeholders to identify patterns and trendsin test results.