CSR-EHCS(CPS),TM: Architectures, Abstractions and Algorithms for Cyber-Physical Networks

网络物理网络(CPN)是计算和物理组件的紧密集成，具有比传统监测和控制系统中发生的更多样和更复杂的相互作用和相互依赖。在软件体系结构研究的基础上，该项目将为一种新的体系结构描述语言开发概念证明，该语言带有注释和视图，将为CPN的设计提供一个统一的框架。还将开发和演示利用基于组件的CPN体系结构描述的结构特征的初始算法。以建筑物的环境监测和控制为目标应用领域，将扩展现有的形式和方法，以纳入更广泛的一套旨在支持成分分析和核查的网络物理特征。原型设计算法将在新的性能度量的背景下开发，这些度量捕捉关键特征，并在CPN的计算、通信和控制的目标和约束之间进行设计权衡。还将开发从CPN架构的多个视角衍生的核查和设计问题网络的概念，使之有可能评估和利用跨越工程和计算机科学中不同的数学形式主义和不同方法造成的当前网络-物理边界的权衡。作为概念的证明，将开发方法来在相互依赖的抽象和问题的有限类网络上合成和迭代验证和设计结果。这些初步结果将在节能建筑设计中加以应用和评估。这项研究将为开发一套通用的体系结构工具和算法，用于CPN的组成分析、设计和验证奠定基础。

Cyber-physical networks (CPNs) are tight integrations of computational and physical components with interactions and interdependencies that are much more varied and complex than what occurs in traditional monitoring and control systems. Building on research in software architectures, this project will develop a proof of concept for a novel architecture description language with annotations and views that will provide a unified framework for the design of CPNs. Initial algorithms that leverage the structural features of component-based architectural descriptions of CPNs will also be developed and demonstrated. Using environmental monitoring and control of buildings as the target application domain, existing formalisms and methods will be extended to incorporate a wider set of cyber-physical features aimed at supporting compositional analysis and verification. Prototype design algorithms will be developed in the context of new performance metrics that capture critical features and design tradeoffs among the objectives and constraints for computation, communication and control in CPNs. The concept of networks of verification and design problems derived from multiple views of CPN architectures will also be developed, making it possible to assess and leverage tradeoffs that cut across current cyber-physical boundaries created by disparate mathematical formalisms and dissimilar methodologies in engineering and computer science. As a proof of concept, methods will be developed to compose and iterate verification and design results over a restricted class of networks of interdependent abstractions and problems. These preliminary results will be applied and evaluated in the context of the design of energy efficient buildings. This research will lay the foundations for developing a general set of architectural tools and algorithms for the compositional analysis, design, and verification of CPNs.