Reproduce “RLMuscle: Muscle-Driven Character Animation via Reinforcement Learning” 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

Chunlei Li, Siyuan Yu, Yang Gao, Shuai Li, Peng Yu, Aimin Hao, He Wang

RLMuscle learns muscle activations in a 1D Hill-type simulation, then transfers them to a separate 3D volumetric simulation. Its 3D fiber constraint carries over the active force–length relation, while omitting velocity, tendon, and passive-force terms. In walking, an EMG prior raised in-band activation on four right-leg channels from 20% to 91%; this is not validation across movements or physiology.

Summary by alphaXiv.

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