Reproduce “Route-Guided CBF-QP Repair for Safe Vision-Language-Action Models” 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
Chan Park, Sunwoo Hong, Kirak Kim, Yoonjeong Park, Alexander W. Olson, Ali Pesaranghader
Route-Guided CBF-QP Repair guides a vision-language-action robot around blocked paths and back toward its task. It detects obstruction in the VLA’s short-horizon actions, selects a bypass route, and adds soft route guidance while retaining a hard CBF safety constraint. In SafeLIBERO’s four-suite average, SafeSuccess rises from 47.19% with AEGIS to 49.69%, though Goal-suite rates decline and collisions remain.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.