A Toolkit for the Analysis of Models of Real-Time Embedded Software

点击翻译按钮获取中文摘要

With the number, size and complexity of real-time embedded systems increasing, the development costs for the software they run have become a concern in many domains including automotive, telecom, avionics, and medical. Growing pressure by governments and regulatory bodies to adopt more rigorous validation, verification, and certification techniques further exacerbates the problem.  With their emphasis on abstraction and automation, model-based development techniques and tools, based on, e.g., the Unified Modeling Language (UML), have the proven potential to allow for the efficient, effective development of high-quality complex embedded software and are increasingly adopted in industry.  ******However, the current state-of-the-art provides insufficient support for the analysis of models, i.e., for automated checks allowing developers to detect flaws in the model early in the design. For instance, for UML-based models as used in parts of the telecom industry, analysis capabilities offered by existing tools are incomplete and unable to adapt to the needs of specific domains, companies, groups, or users; model-level analyses investigated in academia often also fall short in this respect.******Inspired by the success of modern analysis toolkits for source code such as GrammaTech CodeSonar and Coverity Code Advisor the long-term goal of the proposed research is the development of an open, customizable analysis infrastructure and toolkit for industrial UML-based models of real-time embedded systems; more concretely, the toolkit would consist of an Application Programmer Interface (API) containing the data structures, routines, guidelines, and examples necessary for the implementation of a wide range of industrially relevant analyses of these models.******To achieve this goal, we will leverage (1) the background and knowledge collected during our past work on the formal semantics and verification of UML, (2) our industry connections to industrial users of model-based development techniques, and (3) the increasing interest in and availability of open source tools for model-based development in general, and the creation of the PolarSys Eclipse Industry Working Group in particular. More precisely, we will first conduct two case studies on model-level analysis; these case studies will be carried out in collaboration with industry (e.g., Ericsson Canada) to ensure relevance. Then, the insights gained on the required features and requirements for the analysis toolkit will be used to design and implement the toolkit. In the final step, the toolkit will be evaluated and refined. ******The proposed research will advance the state-of-the-art in the foundations and industrial use of model-based software development. Moreover, it will give several students (5 PhD, 3MSc, and 5 BSc students) the knowledge and skills they need to become sought-after graduates and make valuable contributions to Canadian industry.**