Reproduce “Height Control and Optimal Torque Planning for Jumping With Wheeled-Bipedal Robots” 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, Yuan Xu, Binxin Huang, Mandan Chao, Guowei Shi, Xin Yang, Kuangen Zhang, Chenglong Fu
Researchers at Southern University of Science and Technology developed a W-JBD model and a Bayesian optimization approach (BOTP) to enable precise height control and energy-efficient jumping for wheeled-bipedal robots. The BOTP method achieved jumping height errors within 0.2% of the target while generating continuous torque trajectories, significantly improving motor protection and reducing energy consumption compared to model-based methods.
Key insights
- Accurate jumping height control requires detailed dynamic modeling, including factors like leg contraction during the aerial phase, which simpler models often overlook.
- Generating smooth, continuous torque trajectories is essential for motor protection and energy efficiency, a goal better achieved through simulation-based Bayesian optimization than purely model-based calculations.
- A synergistic approach, using a dynamic model to initialize and constrain the search space for data-driven optimization, drastically improves the efficiency and practicality of finding optimal control parameters.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.