Lead: Professor Kerstin Eder
Verification is the process used to gain confidence in the correctness of a system with respect to its specification. While specifications of traditional systems are crisp, evolving functionality requires flexible and adaptive specifications. We are researching verification techniques for adaptive systems, considering formal methods, simulation-based testing and runtime monitoring and verification.
Conformance and compliance testing are particular challenges for systems with evolving functionality. Beyond core functionality, our investigation also includes non-functional properties of autonomous systems such as performance, energy consumption and security. These are closely interlinked, giving rise to the need for multi-objective optimization both at design time and at runtime.
Core to our research is the development, validation and maintenance of verifiable models that can be used as reference points for verification at design time as well as for informed decision-making at runtime. Design-for-verification is a key concept that we are developing and promoting in close collaboration with the technology developers, exploring the design features that make evolving autonomous systems (more) verifiable.
For more information on verification see the Trustworthy Systems Laboratory.