Project Overview
Autonomous cyber-physical-social systems are increasingly deployed in mission-critical environments where safety, security, and reliability are essential. These systems must operate under uncertainty, adapt to dynamic conditions, and withstand intelligent adversaries while maintaining acceptable mission performance. Traditional security and control approaches, which often treat cyber, physical, and human components in isolation, are insufficient to address the tightly coupled and interdependent nature of modern autonomous systems. This project focuses on developing principled, AI-driven foundations for designing autonomous systems that remain secure, resilient, and trustworthy in contested and uncertain environments.
The research investigates uncertainty-aware decision making, adaptive defense, and mission impact assessment for autonomous systems operating under adversarial dynamics. Core technical contributions include probabilistic and game-theoretic modeling of attack–defense interactions, deep reinforcement learning-based intrusion response mechanisms, and cyber deception strategies that proactively reduce adversary effectiveness. These techniques are integrated with mission-level reasoning frameworks that capture interdependencies across cyber, physical, and social components, enabling the system to reason about cascading effects and trade-offs between security, performance, and mission objectives.
A central emphasis of this work is human-machine teaming, where autonomous agents and human operators collaborate through shared situational awareness and explainable decision support. By incorporating human feedback, uncertainty visualization, and adaptive autonomy, the project aims to improve operator trust and decision quality during time-critical missions. The proposed approaches are evaluated using realistic simulation environments and cyber-physical testbeds, with applications spanning autonomous vehicles, tactical networks, and mission-critical infrastructure. Overall, this research advances the foundations of trustworthy autonomy by unifying AI, cybersecurity, and human-centered system design.
