Reproduce “MABEL: An Open Anthropomorphic Mobile Bimanual Platform for Teleoperation and Learning” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Qilong Cheng, Rooholla Khorrambakht, Mohamed Yassine Kabouri, Quang Nam Nguyen, Luca Macesanu, Haoran Zhou, Yipeng Wang, Ludovic Righetti
MABEL combines mobile, adjustable-height hardware and dexterous hands with headset-based whole-body teleoperation. Its shared controller carries commands from demonstrations and simulation to learned policies, while modeled fingertip reach spans floor level to 2.23 m. The authors demonstrate an ACT grasp-and-place policy succeeding in 8 of 10 trials, but the base and lift stayed nearly fixed, so mobile learning remains untested.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.