Reproduce “LLM-guided Planar Loco-Manipulation for Quadruped Robots” 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

Georges Jetti, Michael Khayyat, Jean-Pierre Sleiman, Stefano Arrigoni, Francesco Braghin

The authors present an event-driven planner for quadruped interactive navigation that selects navigation or pushing skills, objects, and geometric targets from task instructions and structured scene poses. It executes one skill at a time using learned low-level policies, then replans. On held-out scenes it completed 109/120 trajectories; hardware tasks succeeded in 34/45 trials, despite 44/45 correct skill sequences.

Summary by alphaXiv.

Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.