Related Experiment Video
Updated: Jun 3, 2026
11:41
Evaluation of an Exclusive Spur Dike U-Turn Design with Radar-Collected Data and Simulation
Published on: February 1, 2020
Real-time lane-level abnormal traffic detection on freeways using sparse telematics data
Shixiao Liang1, Chengyuan Ma1, Pei Li2
1Department of Civil & Environmental Engineering, University of Wisconsin-Madison, Madison, WI, United States.
Accident; Analysis and Prevention
|June 1, 2026
Summary
This study introduces a novel method for real-time abnormal traffic detection on freeways using vehicle telematics data. The system provides lane-level warnings, improving safety and reducing delays compared to traditional methods.
Area of Science:
- Intelligent Transportation Systems
- Traffic Engineering
- Data Science
Background:
- Traditional abnormal traffic detection methods suffer from delays and lack precise location data.
- These limitations pose safety risks and lead to economic losses in intelligent transportation systems.
Purpose of the Study:
- To develop a real-time, lane-level abnormal traffic detection system for freeways.
- To utilize sparse telematics trajectory data for efficient and low-cost traffic monitoring.
Main Methods:
- Offline stage: Discretizing historical trajectories into spatial cells, estimating vehicle intention, and setting alert thresholds using crash reports.
- Online stage: Mapping real-time data to cells, scoring for transition anomalies, speed deviations, and lateral risks.
- Accumulating cell-specific risks to generate a risk map and issue warnings when thresholds are exceeded.
Main Results:
- Achieved a 75% identification rate with lane-level localization and 96% overall accuracy.
- Obtained an F1-score of 0.84 with a low false alarm rate of 0.6% for non-crash events.
- Detected 13% of crashes more than 3 minutes before official notification times.
Conclusions:
- The proposed telematics-based system offers a real-time, low-cost solution for lane-level abnormal traffic detection.
- This approach significantly enhances traffic safety and efficiency in intelligent transportation systems.
- Early detection capabilities can mitigate risks and economic impacts associated with traffic incidents.
