Dependable Embedded Systems

With recent advances in Internet connectivity, network enabled devices have become a substantial part of our everyday lives, extending into areas of increasing importance and criticality.  Consequently, those systems must meet high standards with respect to dependability: Systems must be secure, safe, correct, reliable, maintainable, adaptable, and robust. However, these requirements for dependability are challenged by the growing complexity of modern embedded systems. On one hand, embedded systems are integrated into systems of systems that pool their resources and capabilities in order to offer more functionality and performance than simply the sum of its constituents. On the other hand, we face a paradigm shift towards more intelligent and highly adaptive systems where quality guarantees are hard to achieve. The opportunities seem unlimited, and yet these characteristics appear to be overwhelming in their complexity that any assurances of dependability seem out of reach.

Our answer to this problem is twofold: (1) dependable embedded systems require rigor: the complexity has to be mastered by precise abstraction techniques that enable compositional specification and verification; (2) smart embedded systems demand smart development methods: the complexity due to “intelligent” software has to be mastered by new “intelligent” and rigorous software and methods.

The Dependable Embedded Systems (DES) Lab’s research addresses this growing demand of dependability along these lines. Specifically, the DES Lab will conduct interdisciplinary research across the fields of formal methods for verification and testing, machine learning and embedded compositional software, with a special focus on safety and security.