Reproduce “Kinodynamic Model Predictive Control for Energy Efficient Locomotion of Legged Robots with Parallel Elasticity” 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
Yulun Zhuang, Yichen Wang, Yanran Ding
A hierarchical kinodynamic Model Predictive Control framework integrates Unidirectional Parallel Springs for enhanced energy efficiency in legged robot locomotion. The approach achieved a 38.8% reduction in Cost of Transport in simulation and a 14.8% energy consumption reduction in preliminary hardware experiments on a monoped robot.
Key insights
- Unidirectional Parallel Springs (UPS) effectively reduce energy consumption by sharing external loads and recycling kinetic energy, specifically during the stance phase without compromising swing phase.
- A hierarchical MPC structure, leveraging warm-starting from simpler models, is effective for achieving real-time performance with complex kinodynamic models.
- Explicitly incorporating UPS dynamics into the kinodynamic MPC framework optimizes motor torque distribution, shifting it towards zero and reducing peak loads on actuators.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.