aiqbench analysis
DuoBench offers a reproducible benchmark for dual-arm robot policies, with eleven tasks, stage-based failure analysis, and both simulation and real-hardware results. The authors report that current imitation-learning and vision-language-action policies struggle with bimanual coordination, and that simulation reproduces the same failure modes seen on hardware. For AI readers, the practical value is a shared evaluation testbed for learned robot policies, though the claims rest on the benchmark's own policy results and have not yet been validated by independent replication.
DuoBench: A Reproducible Benchmark for Bimanual Manipulation in Simulation and the Real World
arXiv:2606.11901v2 Announce Type: replace-cross Abstract: Bimanual robot systems substantially expand manipulation capabilities, but coordinating two arms introduces additional control complexity and failure modes that are not well captured by existing benchmarks. We introduce DuoBench, an extensible benchmarking framework for bimanual manipulation policies on the FR3 Duo platform. DuoBench comprises eleven tasks spanning four coordination categories, implemented in simulation and partially replicated in the real world through reproducible task recipes with 3D-printable assets. In addition, we propose a stage-based evaluation scheme that supports fine-grained semantic failure analysis beyond binary success. It includes stage-conditional pass rates with confidence intervals as well as human-teleoperated datasets for all benchmark tasks. We benchmark several imitation-learning and vision-language-action policies in simulation and on real hardware. Our results show that current policies remain challenged by bimanual manipulation, particularly at the beginning of interaction, during parallel arm execution, and in sim-to-real transfer settings. We further find that simulation reveals the same failure modes as real-world experiments. DuoBench provides a reproducible testbed for analyzing them and studying future methods for dual-arm policy learning. Code, datasets, and videos are available at https://duobench.github.io
Abstract:Bimanual robot systems substantially expand manipulation capabilities, but coordinating two arms introduces additional control complexity and failure modes that are not well captured by existing benchmarks. We introduce DuoBench, an extensible benchmarking framework for bimanual manipulation policies on the FR3 Duo platform. DuoBench comprises eleven tasks spanning four coordination categories, implemented in simulation and partially replicated in the real world through reproducible task recipes with 3D-printable assets. In addition, we propose a stage-based evaluation scheme that supports fine-grained semantic failure analysis beyond binary success. It includes stage-conditional pass rates with confidence intervals as well as human-teleoperated datasets for all benchmark tasks. We benchmark several imitation-learning and vision-language-action policies in simulation and on real hardware. Our results show that current policies remain challenged by bimanual manipulation, particularly at the beginning of interaction, during parallel arm execution, and in sim-to-real transfer settings. We further find that simulation reveals the same failure modes as real-world experiments. DuoBench provides a reproducible testbed for analyzing them and studying future methods for dual-arm policy learning. Code, datasets, and videos are available at this https URL
Robotics (cs.RO); Artificial Intelligence (cs.AI)
arXiv:2606.11901 [cs.RO]
(or arXiv:2606.11901v2 [cs.RO] for this version)
https://doi.org/10.48550/arXiv.2606.11901 Focus to learn more arXiv-issued DOI via DataCite
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.
Full text under the publisher's open licence · 273 words
Related coverage
HiPi: Reproducible High-Fidelity Piezoresistive Sensors for Robotic Manipulation
arXiv · Jun 11, 2026
ContactBench 1.0: A Benchmark for Everyday Contact-Rich Robotic Manipulation
arXiv · Oct 20, 2026
The Missing Touch: Spatially Distributed Tactile Feedback Brings Teleoperation Closer to Human Dexterity
arXiv · Aug 21, 2026
Learning Distance-Conditioned Object Transport for Humanoid Loco-Manipulation from a Single Motion Clip
arXiv · Sep 21, 2026