Model-driven V&V processes for computer based control systems: a unifying perspective

A recent trend in software engineering is to support the development process by providing flexible tool chains allowing for eff ective model driven approaches. These solutions are very appealing in industrial settings since they enable the creation of developement and veri cation processes, enhancing abstraction and reuse, and hence improving productivity. This paper addresses advantages and challenges in extending model-driven approaches to system engineering and speci cally to verification and validation (V&V) of critical computer-based systems. Specifically, the paper highlights the needs for real-world industrial contexts and proposes the defi nition of a unifying model-driven process for V&V of functional and non-functional system properties. Some enabling techniques which aim at improving the reuse of model-driven artifacts are addressed to deal with process scalability and e effectiveness. Two sample applications are described for ERTMS/ETCS signalling system in order to show the advantages of the approach: formal modeling for performance evaluation of message delivery between train and track controllers and test case generation for the veri cation of functional requirements of trains outdistancing.