FT-AED Dataset

Institute for Software Integrated Systems Logo

Freeway Traffic Anomalous Event Detection (FT-AED) Dataset

Early and accurate detection of anomalous events on the freeway, such as accidents, can improve emergency response and clearance. However, existing delays and mistakes from manual crash reporting records make it a difficult problem to solve. Current large-scale freeway traffic datasets are not designed for anomaly detection and ignore these challenges. Therefore, we introduce the Freeway Traffic Anomalous Event Detection Dataset, the first large-scale lane-level freeway traffic dataset for anomaly detection. Our dataset consists of a month of weekday radar detection sensor data collected in 4 lanes along an 18-mile stretch of Interstate 24 heading toward Nashville, TN, comprising over 3.7 million sensor measurements. We also collect official crash reports from the Nashville Traffic Management Center and manually label all other potential anomalies in the dataset. This dataset has the potential to be used for developing and benchmarking anomaly detection methods to enchance automatic incident detection methods and reduce reporting delays.

📊 Download the Dataset 📄 Read the Paper 🖥️ Run Anomaly Detection Code

Interactive Data Visualization

Dataset Details

The dataset consists of the following features.
  • time_unix: the time the measurement was taken.
  • milemarker: location of sensor on the interstate. This comes from the Tennessee Department of Transportation's coordination system and is consistent with what is labeled on the road.
  • lane: the lane number. There are four lanes, where lane 1 is the left-most lane.
  • speed: 30-second average speed of the vehicles passing through the area.
  • occupancy: 30-second average percentage of time that the detection zone of the radar sensor is occupied by a vehicle.
  • volume: 30-second average number of vehicles passing through the area.
  • crash_record: whether a crash was officially reported at this timestamp.
  • human_label: human anomaly label for this timestamp.

There are 196 nodes, across the 4 lanes and 49 milemarkers. Data was recorded using a Radar Detection System (RDS) sensor every 30 seconds weekday mornings (4:00 AM - 12:00 PM) in October, 2023. There are 3,763,200 data points.

Citation

If you find the dataset or code useful for your research, please use the following citation.

@misc{coursey2024ftaed,
  title={FT-AED: Benchmark Dataset for Early Freeway Traffic Anomalous Event Detection},
  author={Austin Coursey and Junyi Ji and Marcos Quinones-Grueiro and William Barbour and Yuhang Zhang and Tyler Derr and Gautam Biswas and Daniel B. Work},
  year={2024},
  eprint={2406.15283},
  archivePrefix={arXiv},
}

Team Members

Austin Coursey [1], Junyi Ji [2], Marcos Quinones-Grueiro [1,2], Yuhang Zhang [2], Tyler Derr [3], Gautam Biswas [1], Daniel Work [2]

Please contact austin[dot]c[dot]coursey[at]vanderbilt[dot]edu with questions!

The Modeling and Analysis of Complex Systems (MACS) Lab [1], the Work Research Group [2], and the Network and Data Science (NDS) Lab [3] at Vanderbilt University.

Acknowledgements

This work is supported by a grant from the U.S. Department of Transportation Grant Number 693JJ22140000Z44ATNREG3202 and the Tennessee Department of Transportation under the grant RES2023-20. This material is based upon work supported by the National Science Foundation under Grant No. CNS-2135579 (Work). The U.S. Government assumes no liability for the contents or use thereof. We would also like to acknowledge the Tennessee Department of Transportation (TDOT) for providing the data used in this research.