CRII: CPS SaTC: Securing Smart Cyberphysical Systems against Man-in-the-Middle Attacks

网络物理系统越来越多地成为世界各地黑客的头号目标。我们还看到，连接互联网的关键基础设施激增，可实现轻松监控、可视化和控制。2013年2月，美国总统签署了一项名为“改善关键基础设施网络安全”的行政命令，强调了保护此类关键基础设施免受恶意攻击的迫切需要。因此，该项目正在开发数学工具和方法来设计与智能控制算法相结合的安全通信协议，以确保网络物理系统的安全。PI正在利用网络安全和优化方面的研究来设计算法，这些算法优化系统性能并提供对广泛类别的网络攻击的免疫力。PI和他的学生们还在创建一个试验台，在上面测试设计的算法。该项目更广泛的影响是引入新的数学方法来设计安全系统，并形成跨不同学科的新合作。该项目还在培训一名研究生和两名本科生，他们正在参与该项目。该项目的一些部分将作为案例研究包括在未来的本科课程中。此外，该项目旨在提高下一代工程师和科学家对关键基础设施的网络安全的认识。这是首批广泛研究加密-解密和最佳控制策略联合设计的项目之一。分析中的主要困难是：(I)控制系统的动态；(Ii)外部因素造成的不确定性；以及(Iii)由于加解密算法和实时操作系统造成的信息传输延迟。PI利用分散最优控制、密码学、网络安全和博弈论等基础理论，为大型基础设施设计加密-解密-控制策略。由于控制系统固有的动态特性，用于加密数据的密钥被动态改变，以减少信息泄露给对手的可能性。此外，如果控制器检测到数据被篡改，则它会以一种保证系统稳定性和性能的方式运行。这些算法考虑了(I)由于攻击而在信道中引入的错误，(Ii)由于加密-解密协议造成的延迟，以及(Iii)由实时操作系统施加的实时通信和计算限制。这样设计的算法预计对攻击具有弹性，并能够适应不断变化的系统参数。

Cyber-physical systems have increasingly become top targets for hackers around the world. We are also seeing proliferation of internet-connected critical infrastructures that allow for easy monitoring, visualization, and control. In February 2013, US president signed an executive order "Improving Critical Infrastructure Cybersecurity" that underscores the urgent need for securing such critical infrastructure against malicious attacks. Accordingly, this project is developing mathematical tools and approaches to design secure communication protocol in tandem with smart control algorithms to secure cyber-physical systems. The PI is leveraging research in network security and optimization to devise the algorithms, which optimize system performance and provides immunity against a wide class of cyber-attacks. The PI and his students are also creating a testbed on which the designed algorithms will be tested. The broader impact of the project is to introduce novel mathematical methods for designing secure systems and form new collaboration cutting across various disciplines. The project is also training a graduate student and two undergraduate students, who are working on the project. Some part of the project will be included as case studies in future undergraduate classes. Further, the project aims at increasing the awareness about cybersecurity of critical infrastructures among next generation of engineers and scientists.This is among the first projects to extensively investigate the joint design of encryption-decryption and optimal control strategies. The key difficulties in the analyses are: (i) the dynamics of the control system; (ii) uncertainties due to exogenous factors; and (iii) the delays in the information transfer due to encryption-decryption algorithms and the real-time operating systems. Using the foundational theory of decentralized optimal control, cryptography, network security, and game theory, the PI is designing encryption-decryption-control strategies for large-scale infrastructures. Due to inherently dynamic nature of the control system, the cryptographic key used for encrypting the data is changed dynamically to reduce the possibility of leakage of information to the adversary. Further, if a controller senses tampering of data, then it acts in a manner that guarantees system stability and performance. These algorithms take into account (i) errors introduced in the channel due to attacks, (ii) delays due to encryption-decryption protocol, and (iii) real-time communication and computational constraints imposed by real-time operating systems. The algorithms thus designed are expected to be resilient to attacks and will be able to adapt itself to changing system parameters.