Towards a Framework for Safe and Secure Adaptive Collaborative Systems
Real-time adaptive systems are complex systems capable to adapt their behavior to changing conditions in the environment, and/or internal state changes. Highly dynamic and possibly unpredictable environments, and uncertain operating conditions call for new paradigms of software design, and run-time adaptation mechanisms, to overcome the lack of knowledge at design time. Main application areas include vehicles or robots that need to collaborate to achieve a common task, e.g., minimize fuel consumption, moving objects at a construction site, or performing a set of operations in a factory. Moreover, these vehicles or robots need to interact and possibly collaborate with humans in a safe way, e.g., avoiding accidents or collisions, and prevent hazardous situations that may harm humans and/or machines. % This paper proposes a framework for developing safe and secure adaptive collaborative systems, with run-time guarantees. To enable this, our focus is on requirement engineering and safety assurance techniques to capture the specific safety and security properties for the collaborative system, and to provide an assurance case guaranteeing that the system is sufficiently safe. Moreover, the paper proposes an architecture and behavioral models to analyze the requirements at run-time. Finally, we design a suitable deployment platform to perform the run-time analysis and planning while guaranteeing the real-time constraints.