Exploiting Adaptive Networking for Joint Design ofCommunication and Control in Cyber Physical Systems(eNC2)

项目提案eNC2倡导在网络物理系统中联合设计通信和控制，利用自适应联网概念来支持全球(而不是本地)网络物理系统。该项目的目标是开展多学科的协作研究，以促进未来网络物理系统(CPSS)结合控制和通信(C&C)系统的基础理论，而不是通过不可知的耦合，而是探索C和C系统之间的相互依赖。目前，通信网络集成了软件定义网络(SDN)和网络功能虚拟化(NFV)等颠覆性的新技术来应对新的挑战。延迟，ii.)Resilience和iii.)高度动态的网络条件，允许控制系统在全球范围内处理网络物理系统。从控制的角度来看，我们必须解决将传感器融合-决策-学习(SDL)集成到与网络相连的单个网络物理系统的常规观测器-控制器回路中，并研究新的数学模型以优化方法来系统地推导出联合C&C结构。从通信的角度，我们将引入网络编码和压缩感知来研究NFV/SDN使能通信系统的基本极限，并处理大量的可控实体。为了演示C&C相结合的方法，我们基于OpenStack(NFV)、OpenFlow(SDN)、Kodo(网络编码)和TestMan(专有控制软件)实现了我们自己的试验台，以研究通过相互连接的协作汽车进行敏捷分层决策和学习的作用。

The project proposal eNC2 is advocating for joint design of communication and control in cyber physical systems exploiting adaptive networking concepts to enable global (and not local) cyber physical systems. The goal of the project is to carry out multi-disciplinary and collaborative research to advance the fundamental theory for future cyber physical systems (CPSs) combining control and communication (C&C) system not by agnostic coupling but exploring the inter-dependency between C&C system. Currently, communication networks integrate disruptive novel techniques such as software defined networks (SDN) and network function virtualization (NFV) to deal with the new challenges such as i.) latency, ii.) resilience, and iii.) highly dynamic network conditions, which allow the control systems to address cyber physical systems on a global scale. From the control perspective, we have to address the integration of sensor fusion - decision making - learning (SDL) into the conventional observer-controller loop for individual cyber physical systems attached to the network, and study new mathematical models for an optimization approach to systematically derive a joint C&C structure. From the communication perspective, we will introduce network coding and compressed sensing for NFV/SDN enabled communication systems to study fundamental limits and deal with massive number of controllable entities. In order to demonstrate the combined C&C approach, we implement our own testbed based on openstack (NFV), openflow (SDN), KODO (network coding), and TestMan (proprietary controlling software) to study the role of agile hierarchical decision making and learning by means of connected cooperative cars.