Research map / Vision-language and multimodal / Robot action and navigation
RoboTwin 2.0: A Scalable Data Generator and Benchmark with Strong Domain Randomization for Robust Bimanual Robotic Manipulation
Tianxing Chen, Zanxin Chen, Baijun Chen, Zijian Cai, Yibin Liu, Zixuan Li, Qiwei Liang, Xianliang Lin, Yiheng Ge, Zhenyu Gu, Weiliang Deng, Yubin Guo and 14 more
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add multimodal VLM feedback to code generation (ablation) | RoboTwin 2.0 code agent · 10 code-gen tasks | ASR: 62.1% -> 71.3% | Improved |
| add multimodal VLM feedback to RoboTwin 1.0 code generation (ablation) | RoboTwin 1.0 code agent · 10 code-gen tasks | ASR: 47.4% -> 63.9% | Improved |
| fine-tune VLA on clean non-randomized synthetic data (ablation) | RDT and Pi0 · 8 sim tasks, domain-randomized eval | success rate: RDT 18.8% -> 14.6%; Pi0 22.5% -> 24.9% | No clear effect |
| pretrain on domain-randomized RoboTwin 2.0 data (ablation) | RDT · 8 sim tasks, domain-randomized eval | success rate: 18.8% -> 24.8% | Improved |
| pretrain on domain-randomized RoboTwin 2.0 data (ablation) | Pi0 · 8 sim tasks, domain-randomized eval | success rate: 22.5% -> 29.1% | Improved |
| add 1k domain-randomized synthetic trajectories to 10 real demos (ablation) | RDT · real-world, unseen background, cluttered | success rate: 9.0% -> 42.0% | Improved |
| train on 1k synthetic trajectories only (zero-shot) (ablation) | RDT · real-world, unseen backgrounds | success rate: 15.5% -> 36.5% (clean); 9.0% -> 29.5% (cluttered) | Improved |
| add 1k synthetic data to real demos on Click Bell, seen clean (ablation) | RDT · real-world Click Bell, seen, uncluttered | success rate: 36.0% -> 24.0% | Hurt |
| add embodiment-aware grasp adaptation (RoboTwin 2.0 vs 1.0) (ablation) | data generation pipeline · 50 tasks, 5 embodiments | data collection success rate: +8.3% average; Piper +22.7% | Improved |
| add embodiment-aware grasp adaptation on high-DoF arms (ablation) | data generation pipeline · Franka, UR5 | data collection success rate | No clear effect |
| evaluate on domain-randomized (Hard) instead of clean (Easy) conditions | RDT and Pi0 · RoboTwin 2.0 benchmark | success rate: -20.8% RDT; -30.1% Pi0 | Hurt |
| use VLA pretraining vs non-pretrained policies under Hard conditions | RDT, Pi0 · RoboTwin 2.0 benchmark Hard | average success rate: Pi0 16.3, RDT 13.7 vs ACT 1.7, DP 0.6 | Improved |
Closest papers by meaning
- RoboTwin: Dual-Arm Robot Benchmark with Generative Digital Twins (CVPR 2025 · similarity 0.92)
- RDT-1B: a Diffusion Foundation Model for Bimanual Manipulation (ICLR 2025 · similarity 0.90)
- HumanoidGen: Data Generation for Bimanual Dexterous Manipulation via LLM Reasoning (NeurIPS 2025 · similarity 0.89)
- ManipTrans: Efficient Dexterous Bimanual Manipulation Transfer via Residual Learning (CVPR 2025 · similarity 0.89)
- PIVOT-R: Primitive-Driven Waypoint-Aware World Model for Robotic Manipulation (NeurIPS 2024 · similarity 0.89)
- RobotSmith: Generative Robotic Tool Design for Acquisition of Complex Manipulation Skills (NeurIPS 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
