Reproduce “MarMot: Metamorphic Runtime Monitoring of Autonomous Driving Systems” 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
Jon Ayerdi, Asier Iriarte, Pablo Valle, Ibai Roman, Miren Illarramendi, Aitor Arrieta
MarMot monitors lane-keeping by checking whether steering commands remain consistent across related camera images. It transforms an image and reruns the controller, comparing outputs against an expected relationship such as reversed steering after a horizontal flip. Deviations trigger alarms, not prove failures. On LeoRover, it detected all internal-anomaly recordings on both circuits at 10%, limited to tested mutations and conditions.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.