Research
My research spans dependable AI, digital twins, intelligent control, and trustworthy computing infrastructures, with applications in robotics, advanced air mobility, cloud-edge systems, and cyber-physical platforms.
Neural Dynamics and Control
Data-driven dynamics, neural control, reinforcement learning, and safe autonomy for robotic and aerial systems.
Track 2Digital Twin Systems
Operational digital twins, cloud-integrated simulation platforms, and AI-enabled twin engines for UAM and AAM.
Track 3Dependability and Security
Quantitative dependability, resilience analysis, performance evaluation, and trustworthiness of complex infrastructures.
Overview
Current research domains include:
- Dynamics and control theory and systems
- AI-based digital twin systems and methods
- Computer science and software engineering with specialization in dependable, autonomous, and intelligent systems
- Dependable computing and fault-tolerance of systems and networks
- Mechatronics and aerospace robotic systems
My recent work places particular emphasis on:
- dependable and secure digital twins for advanced and urban air mobility
- reinforcement learning and data-driven control for robotic and aerial systems
- generative, neural, and physics-informed models for digital twins
- dependability, resilience, and security analysis of systems and networks
- cloud-fog-edge infrastructures for cyber-physical and IoT platforms
Neural Dynamics and Control
This track focuses on the intersection of machine learning, system dynamics, and safety-aware control. The goal is to learn useful dynamic representations from observed data while preserving the stability, robustness, and operational guarantees required by real autonomous systems.
Key questions include:
- How can we derive control-oriented dynamic models directly from observed data rather than relying only on analytical derivations?
- How can we design learning-enabled control algorithms that remain robust enough for robotics, autonomous navigation, and urban air mobility scenarios?
- How can digital twin systems become a practical tool for control design and validation?
Digital Twin Systems
This track studies digital twin architectures that couple physical vehicles with continuously updated computational models. The research spans the vehicle-level twin itself, the control and dynamics engines that power it, and the cloud infrastructure needed to operate digital twins at scale.
Representative themes include:
- neural digital twin dynamics engines (DTDE)
- neural digital twin control engines (DTCE)
- digital twin control frames (DTCF)
- cloud-backed digital twin infrastructure (DTCI)
Dependability and Security
This track develops quantitative methods for evaluating trustworthiness in complex computing infrastructures. The core concern is how systems behave under faults, attacks, resource shortages, and operational disruptions, especially when those systems support critical cyber-physical applications.
Representative systems include:
- virtualized server systems and disaster-tolerant data centers
- software-defined networks and cloud-fog-edge platforms
- Internet of Medical Things and IoT infrastructures
- unmanned aerial systems and mission-critical digital twin backends