Augmenting Decision-Making of Human-in-the-Loop Operators for Resilient Cyber-Power Systems

Human operators in Industrial Cyber-Physical Systems (ICPS), such as electric power systems, are responsible for critical, real-time decision-making in control centers. They often rely on Decision Support Tools (DSTs), but multi-modal feedback can lead to cognitive and information overload, especially during extreme events. As human involvement increases, ICPS are evolving into Cyber-Physical Human Systems (CPHS) and CyberPhysical Social Systems (CPSS), where human cognition and machine intelligence are tightly integrated. This work investigates human-machine cooperation using the Cyber-Physical Transmission Resiliency Assessment Metric (CP-TRAM), a DST that integrates physical and cyber alerts to support collaborative decision-making. We developed a laboratory-based human-inthe-loop setup simulating a power grid control room, where CP-TRAM and the Cyber-Power Alarm Tool visualize real-time alarms due to physical and cyber events. An industry-grade operator training software, PowerSimulator, is used to create training scenarios on a synthetic model of Washington state's power system-the Cascadia Network. This model is connected to a hardware-in-the-loop (HIL) testbed, forming a combined hardware, software, and human-integrated environment for training and evaluation. To assess effectiveness, we simulate a cyber-induced physical event based on the MITRE ATT&CK ICS framework and compare operator performance with and without these tools. The study includes 40 participants—18 professional operators and 22 students. We use eye-tracking based metrics to compute a cognitive overload score, validating how CP-TRAM enhances detection and response to cyber events. Results show improved speed and accuracy of decision-making, confirming the DST's effectiveness under high-stress conditions.