Visual-Inertial-Wheel Based Velocity Estimation for a Vertical Wall Climbing Mobile Robot

Abstract

Wall-climbing mobile robots are emerging as pivotal tools in various industries, offering unique capabilities to navigate complex and hazardous environments and perform tasks traditionally undertaken by humans. In such environments, it is often difficult to navigate and localize due to adverse conditions, including poor lighting quality, the absence of features that external sensors can leverage, and the poor condition of the navigation surface, which causes wheel slips. To effectively perform odometry by means of robot velocity estimation, we propose a visual–inertial–wheel odometry estimation using a modular deep learning–based model that leverages insightful information collected by a down-facing camera, an inertial measurement unit mounted on the robot, and the wheel encoders. The proposed model enables modular plugin modules that focus on wheel slip estimation and perception, ensuring reliability when these models are unusable due to environmental conditions. We aim to correctly identify the undesired behavior of the robot, such as wheel slippage, to characterize the navigation surface and estimate the mobile robot's velocities in real time. The presented architecture is successfully evaluated on an internally designed vacuum-based climbing mobile robot, and the results are compared to state-of-the-art methods.

Received: 19 July 2025 | Revised: 29 September 2025 | Accepted: 26 February 2026

Conflicts of Interest

The authors declare that they have no conflicts of interest to this work.

Data Availability Statement

Data sharing is not applicable to this article as no new data were created or analyzed in this study.

Author Contribution Statement

Stefano Mutti: Conceptualization, Methodology, Software, Validation, Formal analysis, Investigation, Resources, Data curation, Writing – original draft, Writing – review & editing, Visualization, Supervision, Project administration. Simone Sabbadini: Methodology, Validation, Investigation, Resources, Data curation, Visualization. Lorenzo Latini: Methodology, Validation, Investigation, Resources, Data curation, Visualization. Diego Gitardi: Conceptualization, Software, Validation, Formal analysis, Data curation, Writing – original draft, Writing – review & editing, Supervision. Anna Valente: Writing – original draft, Writing – review & editing, Supervision, Project administration.