Reproduce “MessyNav: Zero-Shot Navigation via Manipulation in Messy Environments” 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
Kenneth Llontop, Cherie Ho, Dylan Zhou, Anuva Banwasi, Priya Sundaresan, Carlota Parés-Morlans, Jeannette Bohg
MessyNav enables a mobile manipulator to reach goals through clutter by judging which objects are safe and feasible to move before planning how to clear a route. Affordance means what an object permits the robot to do in its situation. It succeeded in 10 of 15 real-robot scenes, versus 6 of 15 for a VLM planner without this stage; this 15-scene comparison has no reported significance test.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.