Reproduce “Cutting the Cord: System Architecture for Low-Cost, GPU-Accelerated Bimanual Mobile Manipulation” 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
Artemis Shaw, Chen Liu, Justin Costa, Rane Gray, Alina Skowronek, Kevin Diaz, Nam Bui, Nikolaus Correll
The authors evolve XLeRobot into an untethered, low-cost two-armed mobile robot by reinforcing printed links, separating motor and compute power paths, applying firmware current limits, and adding onboard Jetson compute. High-Shell links match High-Infill’s 2 mm deflection at 15.8% less mass. In five one-minute stress-test trials, the revised system held 12.0 V; this result is limited to the tested loads.
Key insights
- The High-Shell 3D-printing strategy significantly improved structural rigidity (67.1% increase in yield force) while reducing mass, providing a robust base for manipulators.
- The Tri-Bus power topology effectively eliminated voltage collapses and brownouts observed in shared-bus designs during high-duty cycle operations, maintaining stable voltage.
- Onboard GPU acceleration with the Jetson Orin Nano allows for real-time inference of complex manipulation policies like ACT (36ms latency) and supports advanced VLA models without thermal throttling, enabling untethered autonomy.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.