Fabio Amadio
Papers
6
Total Citations
72
H-Index
5
About
Fabio Amadio is a roboticist whose research sits at the intersection of learning from demonstration, reinforcement learning, and human-robot interaction. His work focuses on enabling robots to acquire complex manipulation skills—particularly bimanual tasks and the handling of non-rigid objects like cloth—by exploiting structural symmetries and intuitive interfaces. Amadio’s most cited paper (26 citations) introduces a method for leveraging symmetries in reinforcement learning to learn bimanual robotic tasks using probabilistic movement primitives (ProMPs). He has also made significant contributions to telemanipulation, developing a target-guided architecture for assisted grasping (20 citations) and a shared-autonomy framework for learning bimanual tasks from demonstration. His work on controlled Gaussian process dynamical models for robotic cloth manipulation (9 citations) addresses the challenging problem of handling deformable objects. Additionally, Amadio has contributed to human-robot interaction with the FollowMe framework for person following (8 citations) and to model-based reinforcement learning with VF-MC-PILCO, an algorithm designed for systems with only raw position measurements (4 citations). His research advances the frontier of intuitive, data-efficient robot learning for real-world manipulation.
Research Focus
Key Achievements
Top Papers
- 1Exploiting Symmetries in Reinforcement Learning of Bimanual Robotic Tasks26 citations · 2019
- 2A Target-Guided Telemanipulation Architecture for Assisted Grasping20 citations · 2022
- 3
- 4
- 5
- 6Learning Control from Raw Position Measurements4 citations · 2023