Dynamic Road Anomaly Detection: Harnessing Smartphone Accelerometer Data with Incremental Concept Drift Detection and Classification
2024

Real-Time Road Anomaly Detection Using Smartphone Data

publication Evidence: moderate

Author Information

Author(s): Imen Ferjani, Suleiman Ali Alsaif

Primary Institution: Imam Abdulrahman Bin Faisal University

Hypothesis

How can we design a machine learning model that dynamically adapts to continuous changes in data patterns for effective road anomaly detection?

Conclusion

The study demonstrates a novel approach for real-time road anomaly detection that effectively adapts to changing data patterns using smartphone accelerometer data.

Supporting Evidence

  • The proposed method achieved a 96% success rate in detecting road anomalies.
  • The system adapts in real-time to changes in road conditions without needing retraining.
  • Incremental learning enhances model responsiveness and efficiency.

Takeaway

This study shows how smartphones can help detect road problems like potholes by using their sensors to learn and adapt to changes in the road conditions.

Methodology

The study employs a hybrid anomaly detection method combining unsupervised and supervised learning to manage concept drift in real-time.

Limitations

The method's performance on alternative data collection systems needs further validation.

Digital Object Identifier (DOI)

10.3390/s24248112

Want to read the original?

Access the complete publication on the publisher's website

View Original Publication