Reproduce “GLAMDRING: Gait Learning And Morphology co-Design via Reinforcement LearnING of CPGs” 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
Amogh Joshi, Kaushik Roy
GLAMDRING co-designs quadruped body geometry, motors, and learned gait for specified speed, payload, and power limits. It trains shared policies across sampled leg lengths, then checks joint torque–speed demands against motor envelopes. In simulation, only 3 of 32 variants trained without added payload remain actuator-feasible at 115% added payload; this result applies to the paper’s test conditions, not real-robot load testing.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.