PhD Defence Yaolei Shen | Aerial Manipulation for Deformable Objects: Modeling, Control and Motion Planning

Aerial Manipulation for Deformable Objects: Modeling, Control and Motion Planning

Yaolei Shen is a PhD student in the department Electrical Engineering. (Co)Promotors are prof.dr.ir. A. Franchi & dr. C. Gabellieri from the faculty  of Electrical Engineering, Mathematics and Computer Science (EEMCS), University of Twente.

Aerial robots have demonstrated remarkable capabilities in inspection, transportation, and infrastructure maintenance. However, most existing aerial manipulation systems are designed to interact with rigid objects, whereas many real-world applications involve deformable objects and environments. Manipulating such objects is significantly more challenging because their deformation is governed by complex continuum dynamics, while aerial robots are often underactuated and subject to limited payload capacity.

This thesis investigates the emerging field of aerial deformable object manipulation and develops a unified methodological framework for modeling, control, and motion planning of aerial robotic systems interacting with deformable objects. This work combines continuum mechanics, reduced-order modeling, differential geometry, optimization, and graph-based planning to tackle the challenges faced in aerial manipulation for deformable objects.

The first part of the thesis focuses on continuum mechanics modeling together with motion and interaction control. Reduced-order continuum models are developed to enable real-time predictive control of deformable objects, while a geometric interaction control framework is introduced to allow aerial robots to safely interact with general deformable environments under physical constraints.

The second part addresses motion planning for deformable object manipulation. Novel planning algorithms are proposed for contact-assisted object shaping and collision-free kinodynamic motion planning in cluttered environments by integrating reduced-order continuum models with trajectory optimization and graph search.

Overall, this thesis establishes a unified foundation for aerial deformable object manipulation and provides new tools for future aerial robotic systems operating in complex environments.