This work proposes a learning-based approach to dynamic non-prehensile object reorientation, enabling fast reorientation of large, grasp-infeasible objects within a constrained task space using uni-manual manipulation. Our policy is trained in simulation via reinforcement learning, utilizing a carefully designed observation space, action space, and reward function to reorient randomly sized cuboids with varied physical properties. Given an object model and a target rotation direction, the policy plans offline trajectories suitable for both simulation and real-world deployment. Although the policy is sensitive to modeling uncertainties, accurate modeling enables successful sim-to-real transfer across different objects and rotation directions.
@misc{DNP_Reorientation_SII2026,
author = "{Abdullah Mustafa and Ryo Hanai and Ixchel Ramirez and Floris Erich and Ryoichi Nakajo and Yukiyasu Domae and Tetsuya Ogata}",
title = "Learning Dynamic Non-Prehensile Object Reorientation via Reinforcement Learning",
year = "2026",
}