Allan Munck: Systems with embedded hardware and software are typically developed by using a document-based approach where the intended systems are specified in a series of documents but increased complexity makes it desirable to use a model-based approach, where the intended systems are modeled in an appropriate modeling language and subsequently simulated.
The transition from using documents to using models can be overwhelming due to the tremendous number of available processes, methodologies, modeling languages and tools. So the project aims at devising a method to simply identify and implement an appropriate modeling setup that supports both the immediate and the future modeling needs. The project therefore starts with a comprehensive investigation of the various modeling processes, methods, languages and tools that may be relevant for model-based systems engineering.
Increasing complexity of embedded systems will lead to increasing model complexity. Therefore there is also an increasing risk of having undetected errors in the models. Subjecting the models to automated testing can reduce such risks. From the software engineering discipline, we know that the so-called that test-driven development method, where test case code is written before production code, can provide better product. We hypothesize that using a similar test-driven approach to modeling can provide better models, which in the end will lead to better products. The project therefore aims at developing a test-driven, model-based systems engineering framework and accompanying development tool.
All proposed processes, methods and tools will be evaluated on ongoing industrial cases from the company GN ReSound.
Effective start/end date 01/08/2013 → 16/08/2017
Published as PhD report: Test-Driven, Model-Based Systems Engineering.
Supervisors: Jan Madsen, Paul Pop
Section for Embedded Systems Engineering