arXiv：cs.LG· Zexi Zhang, Zecheng Zhu, Zidong Chen, Zulkhuu Tuya, Stephen James·· 4 小时前AI 评分49
BiGym 2.0：面向 Unitree G1 人形家务操作的策略基准测试
BiGym 2.0: Benchmarking Learned and Agent-Developed Policies for Humanoid Household Manipulation
AI 导读
BiGym 2.0 将 BiGym 适配到 Unitree G1，用统一全身控制器在 20 项家务任务上提供每任务 60 条 VR 演示，并基准测试 VLA 微调、模仿学习、示范驱动强化学习与冷启动编码智能体。
正文
Abstract:Humanoid household manipulation requires the arms to act while the body balances, steps and changes posture. We present BiGym 2.0, an adaptation of BiGym for the Unitree G1 across 20 household tasks using a unified whole-body controller for demonstration and evaluation. The suite provides 60 native human virtual-reality demonstrations per task with synchronised multi-camera views and full-body execution records. We benchmark vision-language-action fine-tuning, imitation learning, demo-driven reinforcement learning, and cold-start coding agents given the interaction budget of online reinforcement learning. With the same onboard views, proprioception and whole-body controller for every method, vision-language-action fine-tuning has the highest nine-task mean, and agent-developed programs outperform every demo-driven reinforcement learning baseline on this mean and lead on bimanual reaching. Cross-workspace stacking remains open, $\pi_{0.5}$ stays low on pick-box, and multi-object transport is hard for imitation learning, demo-driven reinforcement learning and coding agents. All environments, human demonstrations, and evaluation traces are open-sourced at this https URL.
| Subjects: | Robotics (cs.RO); Machine Learning (cs.LG) |
| Cite as: | arXiv:2610.07594 [cs.RO] |
| (or arXiv:2610.07594v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2610.07594 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Zexi Zhang [view email]
[v1]
Tue, 6 Oct 2026 01:33:35 UTC (8,315 KB)
来源：arXiv：cs.LG · arxiv.org