CAREER: A Unified Event-triggered Real-Time Scheduling and Control Co-Design Framework for Networked Safety-critical Control Systems

This NSF CAREER project aims to ensure the safety, reliability, and efficiency of modern interconnected autonomous systems, such as robot swarms, intelligent transportation networks, and smart manufacturing. The project will bring transformative change to how these complex systems share limited resources, including communication networks, computer processors, and shared physical space, by preventing the "traffic jams" of data or physical collision that cause catastrophic physical accidents or energy waste. This will be achieved by creating new mathematical methods that allow networked systems to smartly schedule actions and design controls that compensate for time delays in real time. The intellectual merit of the project includes establishing a unified theoretical framework that links complex timing delays with physical stability guarantees, enabling safe resource management for large-scale autonomous technologies. The broader impacts of the project include contributing to national prosperity through potential economic savings from mitigated traffic congestion and improved manufacturing efficiency. Furthermore, the project cultivates a highly skilled engineering workforce by integrating these research concepts into the Airborne Robotics Competition (ARC), an accessible, low-cost national robotic blimp competition where K-12 students gain hands-on experience with the critical importance of networked real-time robotic systems. The fundamental technical challenge in large-scale networked control systems is "correlated resource contention," where simultaneous demands for shared resources create complex non-linear timing dynamics. Traditional periodic or centralized methods fail to predict when these unpredictable scheduling delays will destabilize the physical system or determine how to scale up safely. To resolve these issues, this research develops a decentralized real-time scheduling and control co-design framework. First, novel models are formulated to accurately capture timing dynamics caused by multi-layered resource competition. Next, the project designs timing-aware event-triggered control mechanisms that only consume network resources when necessary. A key contribution of this work is establishing rigorous analytical tools to certify "schedu-stability," which is a joint guarantee of scheduling deadline feasibility and control system stability. Finally, the project creates a computationally efficient decentralized optimization framework to solve previously intractable co-design problems for large-scale systems. The optimal solutions and verified timing models resulting from this framework can be used to generate high-quality training data for methods such as imitation learning or reinforcement learning, for even larger system scales, where real-time computation of optimality is infeasible. All theoretical advancements will be integrated into an open-source software toolbox, lowering the barrier for researchers to analyze complex timing behaviors in cyber-physical systems. 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. NSF Award ID: 2539218 | Program: 01002627DB NSF RESEARCH & RELATED ACTIVIT | Principal Investigator: Ningshi Yao | Institution: George Mason University, FAIRFAX, VA | Award Amount: $628,260 View on NSF Award Search: https://www.nsf.gov/awardsearch/show-award/?AWD_ID=2539218 View on Research.gov: https://www.research.gov/awardapi-service/v1/awards/2539218.html