Event
Proposes an obstacle avoidance framework for robotic arms that integrates EDF into observation space
1 reports · 1 sources
Taiga Ikeshima et al. propose a reinforcement learning framework that directly integrates the Euclidean Distance Field (EDF) into the observation space, providing continuous geometric cues for robotic arm obstacle avoidance, simplifying reward design, and achieving effective obstacle avoidance both with and without artificial potential field guidance. The method was validated on a 7-DOF redundant robotic arm, achieving zero-shot sim-to-real transfer; although training used only randomized static objects, the deployed policy still generalizes to unseen real environments and achieves online obstacle avoidance for dynamic obstacles. The paper has been accepted by the 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).
Report timeline
- Distance-Field-Based Reinforcement Learning for Reactive and Robust Obstacle Avoidance in Robotic Arms
This paper proposes a reinforcement learning framework that integrates the Euclidean Distance Field (EDF) directly into the observation space, providing the agent with continuous geometric cues to achieve stable and responsive obstacle avoidance behavior and simplifying reward design; it effectively avoids obstacles both with and without artificial potential field guidance.
arXiv cs.RO ·