Collaborative Research: CPS: Medium: Co-Designed Control and Scheduling Adaptation for Assured Cyber-Physical System Safety and Performance

网络物理系统（CPS）的安全性和性能在很大程度上取决于控制和调度决策，这些决策通常在设计时就已确定，这极大地限制了系统能够安全运行并具有适当性能的条件。超越先前探索不同控制和调度适应单个系统设计的工作，本项目将进行更广泛和深入的研究，探讨如何共同设计网络物理系统的控制和调度，以共同、自动、动态、安全、有效地适应快速、大的、多样化的变化：(1)系统的受控行为；（二）环境；（三）其物理组成部分；(4)其平台软硬件。我们的项目将让多名研究生参与跨学科的研究，包括广泛的教育、培训和指导，涵盖计算机科学、控制理论、自然灾害工程、结构工程、机械工程和计算机工程。我们还将通过夏季REU补充和学期指导的独立学习项目让本科生参与学分，并将利用我们现有的倡议和与合作伙伴组织的关系进行K-12外展。正如我们在之前的每一次合作中所做的那样，我们的多所大学团队将从传统上在科学和技术领域代表性不足的群体中招募、指导和留住参与者，利用我们机构有效和成熟的外展计划。在这个跨学科的研究项目中，我们将开发新的正式模型、分析、系统基础设施和评估度量，以明确地表示、尊重，甚至利用控制和调度相互依赖，以确保系统的行为保持安全，同时使性能得到重大改进。我们提出的新型协同设计方法将从根本上提高网络物理系统的性能，同时尊重可能横贯网络和物理组件以及系统环境的安全约束。例如，它将实现更极端（但安全实现）的压力测试和机械系统和土木结构的适应性管理，以衡量和保持对系统及其环境条件的重大（潜在不利）变化的弹性，并制定相应的适应性缓解响应。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

The safety and performance of cyber-physical systems (CPS) depend crucially on control and scheduling decisions that often are fixed at design time, which significantly restricts the conditions under which a system can operate both safely and with suitable performance. Going beyond prior work that has explored different control and scheduling adaptations in individual system designs, this project will conduct more general and in-depth investigations, into how cyber-physical systems’ control and scheduling can be co-designed to adapt jointly, automatically, dynamically, safely, and effectively even in response to rapid, large, and diverse changes in: (1) the system’s controlled behavior; (2) its environment; (3) its physical components; and (4) its platform software and hardware. Our project will immerse multiple graduate students in cross-disciplinary research, with extensive education, training, and mentoring spanning computer science, control theory, natural hazards engineering, structural engineering, mechanical engineering, and computer engineering. We will also involve undergraduate students via summer REU supplements and in-semester mentored independent study projects for academic credit, and will leverage our existing initiatives and relationships with partner organizations for K-12 outreach. As we have done in each of our previous collaborations, our multi-university team will recruit, mentor, and retain participants from groups traditionally under-represented in science and technology fields, leveraging effective and established outreach programs at our institutions. In this cross-disciplinary research project we will develop new formal models, analyses, system infrastructure, and evaluation metrics, to explicitly represent, respect, and even exploit control and scheduling inter-dependencies, to ensure that systems’ behaviors remain safe while enabling significant improvements in performance. The novel co-design approach we propose will enable radically improved cyber-physical system performance capabilities while respecting safety constraints that may cross-cut cyber and physical components and the system’s environment. For example, it will enable more extreme (but safely realizable) stress testing and adaptive management of mechanical systems and civil structures, to gauge and maintain resilience to significant (potentially adverse) changes to conditions in a system and its environment, and to enact adaptive mitigating responses accordingly.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.