RoboTwin

RoboTwin is a simulation benchmark that facilitates data-driven learning for bimanual manipulation tasks, focusing on the evaluation of coordination in robotic arm movements.

Papers using RoboTwin (11)

$\pi_0$-EqM: Equilibrium Matching for Closed-Loop Vision-Language-Action Control2026BiCoord: A Bimanual Manipulation Benchmark towards Long-Horizon Spatial-Temporal Coordination2026EquiBim: Learning Symmetry-Equivariant Policy for Bimanual Manipulation2026Ada3Drift: Adaptive Training-Time Drifting for One-Step 3D Visuomotor Robotic Manipulation2026HALO: A Unified Vision-Language-Action Model for Embodied Multimodal Chain-of-Thought Reasoning2026Action-Geometry Prediction with 3D Geometric Prior for Bimanual Manipulation2026Robotic VLA Benefits from Joint Learning with Motion Image Diffusion2025PointNet4D: A Lightweight 4D Point Cloud Video Backbone for Online and Offline Perception in Robotic Applications2025Compose Your Policies! Improving Diffusion-based or Flow-based Robot Policies via Test-time Distribution-level Composition2025AVR: Active Vision-Driven Precise Robot Manipulation with Viewpoint and Focal Length Optimization2025Rethinking Bimanual Robotic Manipulation: Learning with Decoupled Interaction Framework2025