publications

2026

1. Video by Lennart Röstel

   GPU-Accelerated Hydroelastic Contact via Signed Distance Fields

   Lennart Röstel, Yang Liu, Jessica Yin, Miguel Angel Zamora Mora, Miles Macklin, Philipp Reist, and Tobias Widmer

   IEEE ICRA Workshop on Generalizable Contact-Rich Robotics: Control and Representation, 2026

   Joint work with the NVIDIA Newton team.

   Abs HTML CR2 Workshop Video

   We present a GPU-accelerated approach to hydroelastic contact using signed distance fields, enabling fast and accurate contact-rich simulation.

2025

1. 

   On Transient Release Dynamics in Robot Throwing: A Sliding Pivot Model

   Yang Liu, and Aude Billard

   IEEE Transactions on Robotics, 2025

   Abs HTML PDF Code

   This paper presents a sliding pivot model to model the transient release dynamics in robot throwing, providing insights into the mechanics of object release.

2. 

   Learning to Throw-Flip

   Yang Liu, Bruno Da Costa, and Aude Billard

   IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025

   Abs arXiv Video

   We present a learning-from-experience framework for robots to throw-flip.

3. 

   Learning Accurate Whole-body Throwing with High-frequency Residual Policy and Pullback Tube Acceleration

   Yuntao Ma, Yang Liu, Kaixian Qu, and Marco Hutter

   IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2025

   Abs arXiv Video

   This work presents a method for learning accurate whole-body throwing using high-frequency residual policy combined with pullback tube acceleration.

2024

1. 

   Tube acceleration: robust dexterous throwing against release uncertainty

   Yang Liu, and Aude Billard

   IEEE Transactions on Robotics, 2024

   Abs HTML PDF Video Poster Media Public Demo

   Tube Acceleration is a novel solution concept that drives the robot to stay in the family of object flying trajectories, so that the throwing outcome is robust against release uncertainty. Through insightful observations, Tube Acceleration can be found via convex optimization within 50ms.

2022

1. 

   A Solution to Adaptive Mobile Manipulator Throwing

   Yang Liu, Aradhana Nayak, and Aude Billard

   IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS), 2022

   Abs HTML PDF Code Public Demo

   Introduce Velocity Hedgehog, a dictionary of throwing configurations, carefully tailored to fit each robot’s geometry and actuator limits. It can be used for reactive and adpative throwing planning for its query efficiency.