Related Experiment Video
Updated: Oct 22, 2025
07:10
Fluorescent Paper Strips for the Detection of Diesel Adulteration with Smartphone Read-out
Published on: November 9, 2018
9.6K
Smartphone Sensing of Road Surface Condition and Defect Detection
Dapeng Dong1, Zili Li2,3
1Department of Computer Science, Maynooth University, W23 F2H6 Maynooth, Ireland.
Sensors (Basel, Switzerland)
|August 28, 2021
Summary
Smartphones can detect road defects using accelerometers and machine learning, offering frequent, low-cost pavement monitoring. This crowdsourced data enables timely maintenance, improving road safety and efficiency.
Area of Science:
- Civil Engineering
- Transportation Engineering
- Data Science
Background:
- Traditional road monitoring is expensive, slow, and infrequent.
- Smartphones offer a potential low-cost, high-frequency data collection alternative.
- Advances in mobile tech, AI, and big data enable new monitoring approaches.
Purpose of the Study:
- To investigate the feasibility of using smartphone accelerometers for road surface defect detection.
- To develop and evaluate a machine learning approach for analyzing crowdsourced road data.
- To demonstrate the potential for quasi-real-time road condition monitoring.
Main Methods:
- Collected accelerometer data from smartphones via an Android application during test drives.
- Processed data using power spectral density analysis.
- Applied a k-means unsupervised machine learning algorithm to identify road defects.
Main Results:
- Achieved an average accuracy of 84% in detecting road surface defects.
- Demonstrated successful processing of low-rate accelerometer data for defect identification.
- Validated the potential of crowdsourced smartphone data for road monitoring.
Conclusions:
- Smartphone-based data collection is a viable method for frequent road surface monitoring.
- Crowdsourced data analysis can significantly reduce inspection and maintenance costs.
- This approach supports timely road maintenance, enhancing serviceability and safety.
