No abstract

Models play an essential role in the design process of cyberphysical systems. They form the basis for simulation and analysis and help in identifying design problems as early as possible. However, the construction of models that comprise physical and digital behavior is challenging. Therefore, there is considerable interest in learning such hybrid behavior by means of machine learning which requires sufficient and representative training data covering the behavior of the physical system adequately. In this work, we exploit a combination of automata learning and model-based testing to generate sufficient training data fully automatically. Experimental results on a platooning scenario show that recurrent neural networks learned with this data achieved significantly better results compared to models learned from randomly generated data. In particular, the classification error for crash detection is reduced by a factor of five and a similar F1-score is obtained with up to three orders of magnitude fewer training samples.

No abstract

The Internet of Things (IoT) extends the Internet to include also wireless embedded computers that are often equipped with sensors and actuators to monitor and control their physical environment. The IoT is increasingly used for safety-critical applications such as smart factories or networked cars, where a failure of the IoT may lead to catastrophic consequences. The IoT is therefore in urgent need of dependability, where reliability, availability, and security properties can be guaranteed even in harsh environments (e.g., radio interference) and under deliberate attacks (e.g., exploiting side channels). In this paper we give an overview of recent research activities in the LEAD project “Dependable Internet of Things in Adverse Environments” towards a dependable IoT, specifically dependable wireless communication and localization using Ultra-Wide-Band technology, secure execution of real-time software, protocol testing and verification, and dependable networked control. We also present the TruckLab testbed, where our research results can be integrated and validated in a platooning use case. In this testbed, model trucks are automatically controlled to follow a lead truck.

No abstract