A Generic Architecture Style for Self-Adaptive Cyber-Physical Systems

Más información sobre el libro

Current concepts for designing automatic control systems depend on dynamic behavioral modeling through mathematical methods, such as differential equations, which face limitations as system complexity increases. This dissertation outlines an architectural evolution of automatic control systems, using adaptive cruise control (ACC) as a key example. Analyzing existing ACC variants reveals that future control systems must enhance self-adaptation and scalability, evolving into self-adaptive cyber-physical systems, which pose significant architectural design challenges. Drawing inspiration from software engineering, a generic architecture style is proposed as a template for constructing networked architectures for both current and future systems. This architecture accommodates various triggering mechanisms and communication paradigms to design dynamic behaviors. To assess the feasibility of this architecture style, existing ACCs are revisited to derive logical architectures and evaluate their consistency with previous designs, particularly through control theory frameworks like block diagrams. Utilizing the proposed architecture, an artificial cognitive cruise control (ACCC) is developed, implemented, and assessed as a next-generation ACC. Evaluation results indicate notable performance enhancements of the ACCC compared to human drivers and existing ACC variants, demonstrating the effectiveness of the new architectural approach.