Current concepts of designing automatic control systems rely on dynamic behavioral modeling by using mathematical approaches like differential equations to derive corresponding functions and slowly reach limitations due to increasing system complexity. Along with the development of these concepts, an architectural evolution of automatic control systems is raised. This dissertation defines a taxonomy to illustrate this architectural evolution relying on a typical application example of adaptive cruise control (ACC). With the help of the analysis of current ACC variants with their architectures, it is indicated that the future automatic control system requires more substantial self-adaptation capability and scalability and evolves into a self-adaptive cyber-physical system, which further constitutes significant challenges for the system’s architecture design. Inspired by software engineering approaches for designing architectures of software-intensive systems, a generic architecture style is proposed. The proposed architecture style serves as a template by following the developed design principle to construct networked architectures not only for the current automatic control systems but also for self-adaptive cyber-physical systems in the future. Different triggering mechanisms and communication paradigms for designing dynamic behaviors are applicable in the networked architecture. To evaluate feasibility of the architecture style, current ACCs are retaken to derive corresponding logical architectures and examine architectural consistency compared to the previous architectures considering the control theory (e.g., in the form of block diagrams). By applying the proposed generic architecture style, an artificial cognitive cruise control (ACCC) is designed, implemented, and evaluated as a future ACC in this dissertation. The evaluation results show significant performance improvements in the ACCC compared to the human driver and current ACC variants.weiterlesen