Secure and Resilient Cyber-Physical Systems

网络物理系统(CP)是包括物理和计算组件的工程交互网络的智能系统。这些集成系统通过支持智能电网、石油和天然气炼油厂以及智能交通系统等关键基础设施(CI)的技术进步，提供了提高生活质量的新功能。然而，这些进步也带来了新的严峻挑战，首要的挑战是安全。安全性是配置项体系结构的一个必要功能，以确保其功能不会受到恶意活动的影响。拟议的研究将开发人工智能(AI)辅助工具，用于智能状况监测、网络威胁追踪和检测、威胁预测和威胁缓解，以保护加拿大的独联体。重点是：a)开发用于故障和异常检测的智能状态监测策略；b)开发最先进的多视图和多层元学习使能模型，用于准确的威胁搜索和检测；c)创建基于随机博弈论的多智能体强化学习的攻击预测模型，以防止在CI中的攻击传播和级联失败；d)开发深度模糊学习系统来收集和共享网络威胁情报，以适当地缓解和应对网络攻击。这项研究与加拿大的网络安全战略和CI国家战略和行动计划相一致，以建设一个更安全、更有保障和更具弹性的加拿大。

Cyber-physical systems (CPS) are smart systems that include engineered interacting networks of physical and computational components. These integrated systems provide new functionalities to improve quality of life by enabling technological advances in our Critical Infrastructure (CI) such as smart power grids, oil and gas refineries, and smart transportation systems. However, the advancements also lead to new and serious challenges, the primary one being security. Security is a necessary feature of the CIs architecture to ensure that their capabilities are not compromised by malicious activities. The proposed research will develop Artificial Intelligence (AI)-aided tools for intelligent condition monitoring, cyber threat hunting and detection, threat prediction, and threat mitigation to protect Canada's CIs. The focus is on; a) developing intelligent condition monitoring strategies for fault and anomaly detection; b) developing a state-of-the-art multi-view and multi-layer meta-learning enabled model for accurate threat hunting and detection; c) creating a novel stochastic game-theoretic multi-agent reinforcement learning-based attack prediction model to prevent attack propagation and cascading failure in CI; d) developing deep fuzzy learning systems to gather and share cyber threat intelligence for proper attack mitigation and response to cyber-attacks. The research aligns with Canada's Cyber Security Strategy and the National Strategy and Action Plan for CI to build a safer, more secure and more resilient Canada.