Navigation Framework for a Hybrid Steel Bridge Inspection Robot

TL;DR

This paper presents a comprehensive navigation framework with four algorithms enabling autonomous inspection of steel bridges by a robot, improving efficiency and reducing human intervention.

Contribution

The paper introduces four novel algorithms for depth data processing, structure segmentation, graph construction, and path planning tailored for steel bridge inspection robots.

Findings

Algorithms effectively process depth data and generate inspection paths.

Experimental results demonstrate successful application on steel bridge structures.

Source code is publicly available for further research and development.

Abstract

Autonomous navigation is essential for steel bridge inspection robot to monitor and maintain the working condition of steel bridges. Majority of existing robotic solutions requires human support to navigate the robot doing the inspection. In this paper, a navigation framework is proposed for ARA robot [1], [2] to run on mobile mode. In this mode, the robot needs to cross and inspect all the available steel bars. The most significant contributions of this research are four algorithms, which can process the depth data, segment it into clusters, estimate the boundaries, construct a graph to represent the structure, generate a shortest inspection path with any starting and ending points, and determine available robot configuration for path planning. Experiments on steel bridge structures setup highlight the effective performance of the algorithms, and the potential to apply to the ARA robot to run on real bridge structures. We released our source code in Github for the research community to use.