Conferences and supporting programme
Visual Modeling of Self-Adaptive Systems
Autonomous vehicles have to be safe and reliable. While certification programs and safety standards such as ISO26262 provide guidance, safety design and development of safe and reliable functionality is time-consuming and costly. Moreover, the integration and promotion of autonomy in vehicles is an extremely challenging task, and although autonomous cars are already seen on our streets, the first severe accidents prove that they are not as secure as we had hoped them to be. We present a visual modeling solution developed to meet recommendations for using model-based approaches in safety design and the trend on automotive industry on using code generation from visual models. The presented approach is based on the KnowLang framework developed for knowledge representation and reasoning in self-adaptive systems. The modeling approach consist of a set of integrated language-specific visual models, each providing a particular view of the system, such as, its overall context, structures and their relationships, along with specific behavior. The visual modeling approach makes the system's description a relatively easy task where visual models support communication and provide feedback within the team. The visual modeling approach also relies on capabilities to manage complexity, such as partition structure from behavior by providing different user-adapted views of the system (e.g. overall and detailed) and by providing possibility to have different views to the systems (e.g. view only inheritance among structural elements). From the model-based specifications the implemented generator produces code for the KnowLang framework for further analysis and execution. This automates the routine and makes the KnowLang framework better accessible for engineers so that they can focus on safety design. The models can also be used for reporting, providing different views for different stakeholders and for documenting the system. By using visual modeling along with automatic code generation from the visual models for KnowLang, we practically reduce both development time and effort, decrease certification costs and improve productivity. We present KnowLang framework along with the developed visual modeling approach and code generator. We also show practical cases and examples of using our modeling approach when developing various self-adaptive systems such as cars and drones.
--- Date: 27.02.2018 Time: 5:00 PM - 5:30 PM Location: Conference Counter NCC Ost