Flexible & Intelligent Complex Systems (FICS)

The Flexible & Intelligent Complex Systems (FICS) research group works on complex systems, autonomy, artificial intelligence, machine learning and their applications on cyber-physical systems, which operate in dynamic, uncertain and safety-critical environments.

FICS researchers develop flexible and intelligent decision-making frameworks for high-dimensional scenarios, bridging foundational research in AI/ML with real-world deployments in aerospace, transportation, robotics, communications and cybersecurity.

The goals of this lab include:

• Engineer safe, secure and adaptive autonomous systems capable of operating in real-world conditions with uncertainty, constraints and adversarial risks.
• Design, validate, and explain AI/ML models, particularly neural networks with emergent modularity and interpretable behaviors.
• Develop decision-making algorithms that integrate learning, control and system modeling under uncertainty.
• Advance domain-specific autonomy across air, space, maritime and ground systems spanning diverse applications, including navigation, wireless communication, robotics and embedded security.

Equipment

• High-end computing platforms for machine learning research and simulation.
• Edge-AI systems (e.g., Nvidia Jetson) for research in onboard autonomy, with aerospace and embedded applications.
• 1/10th-scale autonomous vehicles for validating learning-based control, perception and planning in AV scenarios.
• Multi-drone systems for research in swarming, coordination and decentralized policy learning.
• Robotic arms for studying imitation learning, manipulation under uncertainty and perception–action loops.

Capacity

• AI/ML-controlled CPS Development-Simulation-Test-Improvement Closed Loop, Multi-Agent Systems and Wireless Networks Research, Modeling and Simulation.
• Verification and Validation (V&V) Frameworks for autonomous agents, including black-box validation, edge-case exploration and explainability analysis using tools like SHAP and transfer entropy.
• Formalized safety and security methodologies, addressing operational envelopes, runtime monitoring and adversarial robustness in AI-enabled CPS.
• Closed-loop autonomy pipelines: Iterative design-simulation-deployment-validation cycles using real-time data and targeted scenario injection.
• Simulation-to-reality transfer studies, quantifying and mitigating domain shift between virtual environments and physical platforms.
• Core research themes: deep reinforcement learning, neuroevolution, modular architectures, multi-agent systems, scenario-based testing and intelligent exploration of decision boundaries.

Student Engagement and Mentorship

• Research Experiences for Undergraduates (REU) projects, including AI for robotics, Cybersecurity, CPS safety and autonomous decision-making.
• Capstone design project mentorship, offering real-world problems involving simulation, testing and integration of machine learning with embedded systems.
• Multi-semester project development, guiding students through agile workflows (e.g., GitHub, backlogs, sprints) while emphasizing coding, hands-on development, testing and deliverables.
• Opportunities to work hands-on with drones, autonomous vehicles, edge devices,and simulation environments.