Reproduce “The CRAM Cognitive Architecture for Robot Manipulation in Everyday Activities” 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
Michael Beetz, Gayane Kazhoyan, David Vernon
CRAM enables robots to perform underspecified table-setting and cleanup by completing reusable plans with context-specific objects and movement parameters. It uses knowledge, reasoning, perception, and simulation to select motions without requiring every movement. In a fridge task, the gm_epl baseline repositioned the base an average of 3.6 times while grasping a milk box, a task-specific measure.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.