CAREER: Efficient, Secure and Robust Control of Cyber Physical Systems

随着计算机、通信带宽变得越来越快、越来越便宜，计算和通信能力将嵌入到放置在复杂物理环境中的互联设备中。我们将连接计算和通信的网络世界与物理世界的系统称为网络物理系统（CPS）。计算和通信的融合提出了重大的工程挑战，因为规模，需要桥接物理和软件领域，需要高效，安全和可靠地操作。我们建议为三个关键的子问题开发创新的方法和工具：-有效的资源分配，即有效解析潜在的大量数据以提取与故障/攻击检测、估计、控制和决策相关的信息的能力。我们建议研究传感器调度的新方法，以解决有关估计和检测应用的能源效率和性能保证。-安全性，即检测恶意攻击的能力，以优雅地减少功能来保证操作的连续性，并最终通过重新配置和恢复全部功能来挫败攻击。我们建议通过定义威胁类别和设计攻击弹性控制系统，将安全性集成到系统设计过程中。CPS必须设计成能够处理变化。我们建议建模和分析感知、通信和计算基础设施之间的相互作用，以实现决策机制的鲁棒/可重构设计。更广泛的影响拟议工作的技术影响植根于稳健，安全的CPS设计工具的开发，这将减少设计周期时间，提高运营效率，并为广泛使用CPS提供令人信服的经济案例。效率意味着更少的能源消耗，从而减少排放。可靠性和安全性将被证明是在许多工业应用中采用这种技术的关键。该项目的教育影响源于通过尖端课程将研究直接转移到教室，并通过建立合适的试验台为下一代工程专业学生提供全面的学习体验而转移到实验室。

As computers, communication bandwidth become increasingly faster and cheaper, computing and communication capabilities will be embedded in interconnected devices placed in complex physical environments. We refer to systems that bridge the cyber-world of computing and communications with the physical one as cyber-physical systems (CPS).Intellectual MeritThe convergence of computing and communications raises significant engineering challenges because of the scale, the need to bridge physical and software domains, and the need to operate efficiently, securely and reliably. We propose to develop innovative methods and tools for three critical sub-problems:-Efficient Resource Allocation, i.e. the capability to efficiently parse the potentially enormous amount of data to distill the information relevant to failure/attack detection, estimation, control, and decision-making. We propose to study novel approaches for sensor scheduling to address energy efficiency and performance guarantees with respect to estimation and detection applications.-Security, i.e. the capability to detect malicious attacks, to guarantee continuity of operations with gracefully decreased functionalities, and to ultimately thwart attacks by reconfiguring and restoring full functionality. We propose to integrate security in the system design process, by defining classes of threats and designing attack resilient control systems.-Robustness/Reconfigurability. CPS must be engineered to handle change. We propose to model and analyze the interaction between sensing, communication and computing infrastructure to enable robust/ reconfigurable design of decision-making mechanisms.Broader ImpactThe technological impact of the proposed work is rooted in the development of design tools for robust, secure CPS that will reduce design cycle time, increase operational efficiency and make a compelling economic case for widespread use of CPS. Efficiency implies less energy consumption and therefore lower emissions. Reliability and security will prove crucial to the adoption of such technology in many industrial applications. The educational impact of this project stems from the transition of research directly to classrooms through cutting-edge courses and to laboratories by building suitable test-beds that will offer a comprehensive learning experience for the next generation of engineering students.