Continuous engineering of embedded systems

We investigate the late phases of the embedded systems' life cycles, in particular the treatment of change requests, the integration of legacy components, and the problem of emerging platforms. We propose to tackle these issues in a model-driven design paradigm, on the behavioral models, and to employ techniques from automata theory, model checking and automata learning. The main practical impact of our approach is its support of the systematic completion and update of user/customer requirements, which by their nature are quite partial and concentrate on the most prominent scenarios. Our technique generalizes these typical requirement skeletons by extrapolation and it indicates via automatically generated traces where the requirement specification is too loose and additional information is required. This works in the initial phases of system development, but also in case of change requests, where our technique hints at possible problems with their realization (feature interactions), and helps to keep the requirement model in synchrony along the chain of new releases.