OmniClimb: A Passively Adaptive Rail-Climbing Robot for Multi-Curvature Mobility
Xiang Yue, Zhiyu Chen, Junqing Hu, Xuejie Qiao, Yan Xu
Abstract
Stable mobility of rail-climbing robots in complex aerial worksites remains a significant challenge for industrial inspection and maintenance. Existing designs often struggle with reliable adhesion, efficient locomotion, and adaptability to omnidirectional movement on rails with varying curvature and orientation. This study proposes OmniClimb, a novel omnidirectional rail-climbing robot integrating a gravity-actuated locking mechanism and a variable-posture passive adaptation mechanism. The front locomotion unit employs gravity-actuated locking mechanism to ensure consistent contact stress and friction while self-adjusting to rail inclination. A variable-posture passive adaptation mechanism dynamically adjusts the pitch between front and rear frames, enhancing the gravity-actuated locking effect on the rear unit during curved-rail climbing. A multistage compound transmission system combined with an five-wheel-drive configuration improve power transmission efficiency and disturbance rejection in complex rail environments. Dynamic models were established, and structural parameters were optimized using ADAMS multi-body dynamics simulations. Experiments on a 3D-printed physical prototype demonstrated stable mobility and locking capability on multi-curvature rails under loaded conditions, achieving a maximum speed of 0.053 m/s with controlled motor torque fluctuation (mean 1.25 N·m). OmniClimb's validated adaptability, stability, and efficient climbing performance provide a promising novel solution for deploying rail robots in complex scenarios such as transmission towers and utility tunnels.
BibTeX
@inproceedings{ral2026_omniclimbapassiv,
title = {OmniClimb: A Passively Adaptive Rail-Climbing Robot for Multi-Curvature Mobility},
author = {Xiang Yue and Zhiyu Chen and Junqing Hu and Xuejie Qiao and Yan Xu},
booktitle = {RA-L 2026},
year = {2026}
}