Abstract:
Uncrewed Aerial Vehicles (UAVs) are increasingly used for high-altitude tasks, many of which require not only perception but also active interaction with the environment. This has led to growing interest in aerial manipulation—combining aerial mobility with manipulation capabilities. In this talk, we explore how to move toward general aerial manipulation: enabling a single system to perform a variety of aerial manipulation tasks across diverse scenarios.

To address this, we propose a unified aerial manipulation framework with three key components: (i) hybrid motion-force control for dynamic and compliant physical interaction, (ii) contact-aware trajectory planning to ensure safe and feasible motion under contact constraints, and (iii) an end-effector-centric interface to support intuitive teleoperation and high-level policy learning.
We demonstrate the system’s versatility on tasks ranging from surface inspection to object manipulation, achieving improved precision, robustness, and task generality. We believe this work takes a meaningful step toward enabling autonomous aerial robots to operate as capable physical agents in the real world.

Committee:
Prof. Guanya Shi (advisor)
Prof. Sebastian Scherer (advisor)
Prof. Wennie Tabib
Xiaofeng Guo