Reproduce “Robotic Inspection and Characterization of Subsurface Defects on Concrete Structures Using Impact Sounding” 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
Ejup Hoxha, Jinglun Feng, Diar Sanakov, Ardian Gjinofci, Jizhong Xiao
Researchers from CCNY Robotics Lab developed the "Impact-Rover," a mobile robot that automates impact-sounding data collection on concrete structures, including vertical surfaces. The system integrates vision-based localization with unsupervised learning on frequency-domain features to accurately detect and map subsurface defects like voids and delaminations, providing interpretable visualizations for concrete condition assessment.
Key insights
- Sound signal differences between healthy and defective concrete are significantly more discernible in the frequency domain, especially regarding energy levels.
- Feature engineering, including the extraction of spectral moments and their subsequent enhancement, is crucial for improving the clarity and visibility of defects in the resulting maps.
- Unsupervised clustering methods, like K-Means (with feature enhancement) and Spectral Clustering, can effectively automate defect identification and mapping without prior labeled data, even differentiating surrounding weakened areas.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.