My research focuses on structural health monitoring and intelligent operation, with an emphasis on integrating digital twins, Bayesian inference, and artificial intelligence for critical infrastructure such as bridges and offshore wind turbines. It aims to develop next-generation engineering paradigms that tightly couple data-driven methods with physical principles.

The main research areas include:

• Structural Health Monitoring (SHM)

• Digital Twin

• AI for Engineering (integration of deep learning and scientific computing)

• UAV-based intelligent inspection and computer vision

• Bayesian inference and uncertainty quantification 

1.    Digital Twin and Intelligent Structural Sensing

Targeting complex infrastructure such as bridges and wind turbines, this research develops high-fidelity digital twins by integrating physics-based models with multi-source monitoring data, establishing a unified paradigm that combines physical modeling, data-driven learning, and uncertainty quantification. Leveraging structural health monitoring and multi-modal sensing, it enables structural state perception and anomaly detection, intelligent operational modal identification under varying conditions, Bayesian-based model updating, full-field response reconstruction under sparse measurements, and real-time inversion of unknown loads (e.g., wind, traffic, seismic). The ultimate goal is to build an integrated digital twin framework with capabilities in real-time perception, diagnosis, prediction, and decision support, enabling interpretable, updatable, and predictive intelligent operation and maintenance for the safety and long-term management of critical infrastructure.

2.    UAV-enabled Intelligent Inspection and Digital Sensing

Targeting large-scale infrastructure such as bridges, buildings, and wind turbines, this research develops a UAV-based intelligent inspection framework by integrating multi-modal sensing data, including RGB images, infrared, and LiDAR. It combines autonomous path planning, deep learning-based damage detection, and high-fidelity 3D reconstruction via oblique photogrammetry and data fusion to enable accurate spatial representation and quantitative assessment of structural defects. By further integrating with BIM/BrIM through geometric and semantic mapping, the framework supports component-level damage localization, lifecycle data management, and automated report generation for maintenance decision-making. In addition, computer vision-based video measurement techniques are employed to extract structural vibration responses from UAV footage, reducing reliance on traditional contact sensors and promoting the integration of inspection and monitoring. The ultimate goal is to establish a unified UAV-enabled platform for autonomous inspection, multi-source sensing, intelligent analysis, and decision support, advancing efficient, safe, and smart operation and maintenance of infrastructure.

3.    Physics–Data Hybrid Intelligent Inference

Targeting the core needs of digital twins and intelligent operation of complex structures such as bridges and wind turbines— including full-field response prediction, unknown load inversion, and damage identification—this research develops next-generation intelligent inference methods that tightly integrate physical modeling with data-driven learning. By combining physics-based priors (e.g., finite element models, differential equations, and state-space representations) with multi-source monitoring data and deep learning, a unified framework is established for accurate sensing and inference of structural states and external actions. The approach is designed to enhance robustness, generalization, and practical applicability under sparse data, complex environments, and model uncertainties, providing reliable support for UAV-based inspection and structural health monitoring, and enabling state updating, performance prediction, and decision-making within digital twin systems. Ultimately, it aims to bridge the full pipeline from perception to modeling, prediction, and decision, advancing infrastructure management from experience-driven to physics–data fusion-driven paradigms.