EFRI-ARESCI: Foundations for Reconfigurable and Autonomous Cyber-Physical Systems: Cyber-Cities and Cyber-Universities

项目负责人：Munther A. dahleh机构：麻省理工学院提案号：0735956efr - aresci：可重构和自主网络物理系统的基础：网络城市和网络大学该项目由麻省理工学院和加州理工学院的研究人员合作，重点研究自主可重构网络物理系统的新理论基础。在网络物理系统（包括关键网络基础设施、应急响应系统和分布式移动机器人）大量应用的推动下，该项目致力于对健全的数学、计算和技术基础进行基础和应用研究，以解决此类系统应对意外中断时的自主性和可重构性问题。同时，研究人员开发了网络城市和网络校园的真实测试平台，确定了研究计划的一系列挑战问题，并提供了测试该项目下开发的技术的平台。该项目考虑基于以下方法的范式转变：(i)机制设计，通过设计源自全球性能目标的本地效用函数，为分布式自治系统的互连提供框架；（ii）解决各种网络场景的通信约束下的分布式决策方法；（iii）与贝叶斯方法相比，非贝叶斯方法创造了处理意外中断所需的适当灵活性，而贝叶斯方法需要所有对手的模型；（iv）近似和松弛方法，允许构建互连的简化（抽象）模型，从中可以推导出可靠性和鲁棒性的限制。提出的方法将为应急响应和其他系统提供最优的有效响应策略。该试验台将为预测和管理大型网络化城市的活力提供另一个现实例子。该项目将为学生创造与美国和国外主要工业公司互动的机会。项目成果将被整合到麻省理工学院和加州理工学院系统领域的第一年研究生课程中。通过麻省理工学院的开放课程计划，研究人员打算让感兴趣的各方都能接触到这些发展。计划还包括在当地K-12学校开展外展活动。

PI: Munther A. DahlehInstitution: Massachusetts Institute of TechnologyProposal Number: 0735956EFRI-ARESCI: Foundations for Reconfigurable and Autonomous Cyber-Physical Systems: Cyber-Cities and Cyber-UniversitiesThis project, with investigators from the Massachusetts Institute of Technology and the California Institute of Technology, focuses on new theoretical foundations for autonomously reconfigurable cyber-physical systems. Motivated by the abundance of applications of Cyber-Physical Systems, including critical networked infrastructures, emergency response systems, and distributed mobile robotics, the project pursues fundamental and applied research towards sound mathematical, computational, and technological foundations to address issues of autonomy and reconfigurablity as such systems react to unexpected disruptions. In conjunction, the investigators develop real-life testbeds of a cyber-city and a cyber-campus, which identify a set of challenge problems for the research program and offer platforms for testing the technologies developed under this project. The project consider a paradigm shift that hinges on the following methods: (i) mechanism design which provides a framework for interconnections of distributed autonomous systems by designing local utility functions that are derived from a global performance objective; (ii) distributed decision methods under communication constraints which address various networking scenarios; (iii) non-Bayesian methods which create the appropriate flexibility necessary to deal with unexpected disruptions as contrasted to Bayesian methods that require models of all adversaries; and (iv) approximation and relaxation methods which allow constructing simplified (abstracted) models of interconnections from which limitations on reliability and robustness can be derived.The proposed approach will provide optimal efficient response strategies for emergency response and other systems. The test bed will provide another real-life example for prediction and management of liveliness of large networked cities. The project will create opportunities for students to interact with major industrial companies in the U.S. and abroad. Project outcomes will be integrated into the first-year graduate curriculum in the systems area at both MIT and Caltech. Through MIT's OpenCourseWare initiative, the investigators intend to make these developments accessible to interested parties. Plans also include outreach activities to local K-12 schools.