Papers
5
Total Citations
32
H-Index
4
About
Dominik Winkelbauer is a robotics researcher whose work sits at the intersection of robotic manipulation, deep learning, and 3D perception. His research focuses primarily on enabling robots to grasp and manipulate objects intelligently, even under conditions of partial observability and complex hand kinematics. Among his most notable contributions is a two-stage learning architecture for planning stable grasps with an 18-DOF four-fingered robotic hand — a landmark challenge given the enormous dimensionality of the search space — which has garnered 11 citations since its 2022 publication. Building on this foundation, Winkelbauer has pioneered methods that combine 3D shape completion with grasp prediction, allowing robots to reason about objects from incomplete sensor data and act with greater versatility in assistive robotics contexts. His 2023 work on uncertainty-aware shape completion further advances the field by flagging geometrically unreliable regions, improving downstream planning safety. Winkelbauer has also addressed the practical challenge of humanoid robot calibration, proposing a self-contained method using only a head-mounted RGB camera. His most recent work bridges dextrous grasping and in-hand manipulation through reinforcement learning, pointing toward more seamlessly integrated real-world robotic systems.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3Shape Completion with Prediction of Uncertain Regions6 citations · 2023
- 4
- 5