Reproduce “Shape control of simulated multi-segment continuum robots via Koopman operators with per-segment projection” 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
Eron Ristich, Jiahe Wang, Lei Zhang, Sultan Haidar Ali, Wanxin Jin, Yi Ren, Jiefeng Sun
A data-driven Koopman operator approach enables real-time, closed-loop shape control for simulated multi-segment tendon-driven soft continuum robots. This method incorporates a per-segment projection scheme that improves model accuracy by an order of magnitude and facilitates reaching complex target shapes with multiple segments.
Key insights
- The per-segment projection scheme significantly improves Koopman model accuracy by an order of magnitude, effectively mitigating the challenges posed by highly nonlinear dynamics in multi-segment soft robots.
- Using the change in control input (∆u) for the Koopman operator, in conjunction with the projection scheme, yields models with the highest accuracy compared to using absolute control inputs.
- While MPC with Koopman models enables computationally efficient shape control, inherent steady-state errors in the identified Koopman models prevent the controller from maintaining the absolute lowest possible error, highlighting an area for future improvement.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.