Reproduce “Contact-Anchored Proprioceptive Odometry for Legged and Wheel-Legged 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
Minxing Sun, Yao Mao
The Contact-Anchored Proprioceptive Odometry (CAPO) framework provides robust, purely proprioceptive state estimation for legged and wheel-legged robots by using contact anchoring and novel drift suppression mechanisms. It achieved a horizontal-loop error of 0.1638 m for a point-foot robot on a 200m trajectory and maintained vertical stability with less than 0.1 m error over repeated step-up/step-down cycles.
Key insights
- Recording and continuously re-anchoring to world-frame contact positions during stance phases is effective at suppressing long-term horizontal and vertical position drift without external sensors.
- A dynamic height-memory mechanism, which clusters and corrects touchdown heights to dominant support planes, significantly improves elevation consistency over varied and extended traversals.
- Filtering noisy raw encoder velocities through an Inverse-Kinematics Cubature Kalman Filter (IKVel-CKF) greatly enhances the robustness of contact-anchored updates, particularly in dynamic and impact-prone locomotion.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.