Software engineering : testing real-time embedded systems using timed automata based approaches

摘要

Real-time Embedded Systems (RTESs) have an increasing role in controlling society infrastructures that we use on a day-to-day basis. RTES behaviour is not based solely on the interactions it might have with its surrounding environment, but also on the timing requirements it induces. As a result, ensuring that an RTES behaves correctly is non-trivial, especially after adding time as a new dimension to the complexity of the testing process. This research addresses the problem of testing RTESs from Timed Automata (TA) specification by the following. First, a new Priority-based Approach (PA) for testing RTES modelled formally as UPPAAL timed automata (TA variant) is introduced. Test cases generated according to a proposed timed adequacy criterion (clock region coverage) are divided into three sets of priorities, namely boundary, out-boundary and in-boundary. The selection of which set is most appropriate for a System Under Test (SUT) can be decided by the tester according to the system type, time specified for the testing process and its budget. Second, PA is validated in comparison with four well-known timed testing approaches based on TA using Specification Mutation Analysis (SMA). To enable the validation, a set of timed and functional mutation operators based on TA is introduced. Three case studies are used to run SMA. The effectiveness of timed testing approaches are determined and contrasted according to the mutation score which shows that our PA achieves high mutation adequacy score compared with others. Third, to enhance the applicability of PA, a new testing tool (GeTeX) that deploys PA is introduced. In its current version, GeTeX supports Control Area Network (CAN) applications. GeTeX is validated by developing a prototype for that purpose. Using GeTeX, PA is also empirically validated in comparison with some TA testing approaches using a complete industrial-strength test bed. The assessment is based on fault coverage, structural coverage, the length of generated test cases and a proposed assessment factor. The assessment is based on fault coverage, structural coverage, the length of generated test cases and a proposed assessment factor. The assessment results confirmed the superiority of PA over the other test approaches. The overall assessment factor showed that structural and fault coverage scores of PA with respect to the length of its tests were better than the others proving the applicability of PA. Finally, an Analytical Hierarchy Process (AHP) decision-making framework for our PA is developed. The framework can provide testers with a systematic approach by which they can prioritise the available PA test sets that best fulfils their testing requirements. The AHP framework developed is based on the data collected heuristically from the test bed and data collected by interviewing testing experts. The framework is then validated using two testing scenarios. The decision outcomes of the AHP framework were significantly correlated to those of testing experts which demonstrated the soundness and validity of the framework.