Projects

Research Problem and Objectives

This project aims to develop an industrial-grade autonomous wind turbine inspection system based on drone and AI. The proposed approach integrates an image quality-aware flight control system to ensure high-resolution imagery under wind disturbances, a field-authenticated dataset labeled according to maintenance manuals and expert input, and a field-driven foundation AI model tailored to inspection tasks. Together, these components establish a closed-loop, field-to-field AI framework that enables reliable deployment beyond laboratory settings. 

Student Research Activities

Design of Autonomous Drone Flight Control System 

The student will learn to develop and test adaptive drone control methods that account for both flight performance and image quality, enabling reliable data collection in windy, real-world inspection environments. 

Development of Industrial-grade wind turbine inspection AI model 

This objective is divided into three tasks to develop a comprehensive AI-driven inspection system for wind turbines. The first task focuses on incorporating domain knowledge from maintenance manuals and field engineers into the AI framework, enabling the model to associate detected damage with maintenance urgency and actionable recommendations rather than simple visual classification. The second task aims to develop a field-driven AI inspection model that is trained and validated against real wind turbine maintenance histories, ensuring that its predictions are aligned with practical engineering decisions and operational constraints. This validation establishes a robust field-to-field closed-loop learning framework. The third task centers on real-world deployment of the AI-enabled inspection system, where drones perform autonomous inspections and the AI model provides real-time damage assessment, urgency evaluation, and maintenance suggestions to field engineers. Feedback from these deployments is iteratively incorporated to refine and improve the model’s performance and reliability in operational environments. 

Required qualifications and skills

• Undergraduate standing (at least left with one semester for graduation)

• Python/C++/Matlab programming skills