This repo provides the source code and data for the following paper:
Zhou, H., Ma, K., Liang, S. et al. A unified longitudinal trajectory dataset for automated vehicle. Sci Data 11, 1123 (2024). https://doi.org/10.1038/s41597-024-03795-y
We processed a unified trajectory dataset for automated vehicles' longitudinal behavior from 14 distinct sources. The extraction and cleaning of the dataset contains the following three steps - 1. extraction of longitudinal trajectory data, 2. general data cleaning, and 3. data-specific cleaning. The datasets obtained from Step 2 and Step 3 are the longitudinal and car-following trajectory data. We also analyzed and validated the data using multiple methods. The obtained datasets are provided in Ultra-AV: A unified longitudinal trajectory dataset for automated vehicle (figshare.com). The Python code used to analyze the datasets can be found at https://github.com/CATS-Lab/Filed-Experiment-Data-ULTra-AV. We hope this dataset can benefit the study of microscopic longitudinal AV behaviors.
We have examined 13 open-source datasets, each providing distinct insights into AV behavior across various driving conditions and scenarios. These open-source datasets are from six providers:
- Vanderbilt ACC Dataset [1]. Collected in Nashville, Tennessee by Vanderbilt University research group. Available at - https://acc-dataset.github.io/datasets/.
- MircoSimACC Dataset [2]. Collected in four cities in Florida, including Delray Beach, Loxahatchee, Boca Raton, and Parkland by the Florida Atlantic University research group. Available at -https://github.com/microSIM-ACC/ICE.
- CATS Open Datasets [3]. Three datasets were gathered in Tampa, Florida, and Madison, Wisconsin by the CATS Lab. Available at - https://github.com/CATS-Lab.
- OpenACC Database [4]. Four datasets were collected across Italy, Sweden, and Hungary by the European Commission's Joint Research Centre. Available at - https://data.europa.eu/data/datasets/9702c950-c80f-4d2f-982f-44d06ea0009f?locale=en.
- Central Ohio ACC Datasets [5]. Two datasets were collated in Ohio by UCLA's Mobility Lab and Transportation Research Center. Available at -
- Waymo Open Dataset [6, 7]. Two datasets were collected in six cities including San Francisco, Mountain View, and Los Angeles in California, Phoenix in Arizona, Detroit in Michigan, and Seattle in Washington by Waymo. Available at -
- Argoverse 2 Motion Forecasting Dataset [8]. Collected from Austin in Texas, Detroit in Michigan, Miami in Florida, Pittsburgh in Pennsylvania, Palo Alto in California, and Washington, D.C. by Argo AI with researchers from Carnegie Mellon University and the Georgia Institute of Technology. Available at -
For more details on the datasets, please refer to the reference and our paper.
All the data are provided in CSV format. If you want to run our source code, please make sure you follow the prerequisites below:
- Python 3 - Ensure you have a Python 3 environment set up.
- Required Packages - Install all necessary packages listed in the
requirements.txtfile. - Original data or the processed dataset - Download the original data or our processed data from the link we provided and check the path of the data with our code if you want to process or analyze the data.
We also recommend using other software packages such as R to effectively analyze the trajectory data. These tools are well-suited for handling the dataset's format.
The code related to our data processing and validation is all stored in folder \Code. This folder contains the following files:
- main.py - The main function calls data processing and analysis functions for each dataset.
- trajectory_extraction.py - Code used in Step 1 to extract AV longitudinal trajectories.
- data_transformation.py - Code used in Step 1 to convert all datasets to a unified format.
- data_cleaning.py - Code used in Steps 2 and 3 for data cleaning.
- data_analysis.py - Code used to analyze data statistics, plot traffic performance of datasets, and plot scatter plots.
- model_calibration.py - An example tool to use the processed data to calibrate a linear car-following model.
To use this repo, run the Python script main.py. As you proceed through each Python script, always verify the paths for both the input and output files. This ensures that everything runs smoothly.
Data attributes are shown below:
| Label | Description | Notations and formulation | Unit |
|---|---|---|---|
| Trajectory_ID | ID of the longitudinal trajectory. |
|
N/A |
| Time_Index | Common time stamp in one trajectory. |
|
s |
| ID_LV | LV (lead vehicle) ID. |
|
N/A |
| Type_LV | LV is an AV or human-driving vehicle. | Label AV with 1 and human-driving vehicles with 0. | N/A |
| Pos_LV | LV position in the Frenet coordinates. | $p^{\mathrm{l}}{it}=p^{\mathrm{f}}{it}+h_{it}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m |
| Speed_LV | LV speed. | $v^{\mathrm{l}}{it}=\frac{p^{\mathrm{l}}{i(t+1)}-p^{\mathrm{l}}_{it}}{\Delta t}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m/s |
| Acc_LV | LV acceleration. | $a^{\mathrm{l}}{it}=\frac{v^{\mathrm{l}}{i(t+1)}-v^{\mathrm{l}}_{it}}{\Delta t}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m/s² |
| ID_FAV | FAV (following automated vehicle) ID. |
|
N/A |
| Pos_FAV | FAV position in the Frenet coordinates. | $p^{\mathrm{f}}{it}=p^{\mathrm{f}}{i(t-1)}+\Delta t \cdot v^{\mathrm{f}}_{it}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m |
| Speed_FAV | FAV speed. | $v^{\mathrm{f}}{it}=\frac{p^{\mathrm{f}}{i(t+1)}-p^{\mathrm{f}}_{it}}{\Delta t}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m/s |
| Acc_FAV | FAV acceleration. | $a^{\mathrm{f}}{it}=\frac{v^{\mathrm{f}}{i(t+1)}-v^{\mathrm{f}}_{it}}{\Delta t}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m/s² |
| Space_Gap | Bump-to-bump distance between two vehicles. | $g_{it}=p^{\mathrm{l}}{it}-p^{\mathrm{f}}{it} - l^{\mathrm{f}}/2 -l^{\mathrm{l}}/2, i\in \mathcal{I}, t\in \mathcal{T}_i$, where |
m |
| Space_Headway | Distance between the center of two vehicles. | $h_{it}=p^{\mathrm{l}}{it}-p^{\mathrm{f}}{it}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m |
| Speed_Diff | Speed difference of the two vehicles. | $\Delta v_{it}=v^{\mathrm{l}}{it}-v^{\mathrm{f}}{it}, i\in \mathcal{I}, t\in \mathcal{T}_i$. | m/s |
The FAV IDs are provided below:
Vanderbilt Two-vehicle ACC Dataset:
- 0 - A commercially available 2019 SUV with a full-speed range adaptive cruise control system.
MicroSimACC Dataset:
- 0 - Toyota Corolla LE 2020
CATS ACC Dataset:
- 0 - Lincoln MKZs 2016 (Black)
- 1 - Lincoln MKZs 2017 (Red)
CATS Platoon Dataset:
- 0 - Lincoln MKZs 2016 (Black)
- 1 - Lincoln MKZs 2017 (Red)
CATS UWM Dataset:
- 0 - Lincoln MKZs 2017 (Red)
OpenACC Casale Dataset:
- 0 - Rexton
- 1 - Hyundai
OpenACC Vicolungo Dataset:
- 0 - Ford(S-Max)
- 1 - KIA(Niro)
- 2 - Mini(Cooper)
- 3 - Mitsubishi(OutlanderPHEV)
- 4 - Mitsubishi(SpaceStar)
- 5 - Peugeot(3008GTLine)
- 6 - VW(GolfE)
OpenACC Asta Dataset:
- 0 - Audi(A6)
- 1 - Audi(A8)
- 2 - BMW(X5)
- 3 - Mercedes(AClass)
- 4 - Tesla(Model3)
OpenACC ZalaZone Dataset:
- 0 - AUDI_A4
- 1 - AUDI_E_TRON
- 2 - BMW_I3
- 3 - JAGUAR_I_PACE
- 4 - MAZDA_3
- 5 - MERCEDES_GLE450
- 6 - SMART_TARGET
- 7 - SKODA_TARGET
- 8 - TESLA_MODEL3
- 9 - TESLA_MODELS
- 10 - TESLA_MODELX
- 11 - TOYOTA_RAV4
Ohio Single-vehicle Dataset:
- 0 - retrofitted Tesla Sedan
Ohio Two-vehicle Dataset:
- 0 - retrofitted Tesla Sedan
- 1 - retrofitted Ford Fusion Sedan
Waymo Perception Dataset:
- 0 - Waymo ADS-equipped vehicle
Waymo Motion Dataset:
- 0 - Waymo ADS-equipped vehicle
Argoverse 2 Motion Forecasting Dataset:
- 0 - Argo AI self-driving Ford
For more details on the labels and the vehicle types, please refer to our paper.
Developer - Hang Zhou (hzhou364@wisc.edu).
Code reviewer - Ke Ma (kma62@wisc.edu).
If you have any questions, please feel free to contact CATS Lab in UW-Madison. We're here to help!
[1] Wang, Yanbing, George Gunter, Matthew Nice, and Daniel B. Work. "Estimating adaptive cruise control model parameters from on-board radar units." arXiv preprint arXiv:1911.06454 (2019).
[2] Yang, Mingyuan, Pablo Chon-Kan Munoz, Servet Lapardhaja, Yaobang Gong, Md Ashraful Imran, Md Tausif Murshed, Kemal Yagantekin, Md Mahede Hasan Khan, Xingan Kan, and Choungryeol Lee. "MicroSimACC: an open database for field experiments on the potential capacity impact of commercial Adaptive Cruise Control (ACC)." Transportmetrica A: Transport Science (2024): 1-30.
[3] Shi, Xiaowei, and Xiaopeng Li. "Empirical study on car-following characteristics of commercial automated vehicles with different headway settings." Transportation research part C: emerging technologies 128 (2021): 103134.
[4] Makridis, Michail, Konstantinos Mattas, Aikaterini Anesiadou, and Biagio Ciuffo. "OpenACC. An open database of car-following experiments to study the properties of commercial ACC systems." Transportation research part C: emerging technologies 125 (2021): 103047.
[5] Xia, Xin, Zonglin Meng, Xu Han, Hanzhao Li, Takahiro Tsukiji, Runsheng Xu, Zhaoliang Zheng, and Jiaqi Ma. "An automated driving systems data acquisition and analytics platform." Transportation research part C: emerging technologies 151 (2023): 104120.
[6] Hu, Xiangwang, Zuduo Zheng, Danjue Chen, Xi Zhang, and Jian Sun. "Processing, assessing, and enhancing the Waymo autonomous vehicle open dataset for driving behavior research." Transportation Research Part C: Emerging Technologies 134 (2022): 103490.
[7] Ettinger, Scott, Shuyang Cheng, Benjamin Caine, Chenxi Liu, Hang Zhao, Sabeek Pradhan, Yuning Chai et al. "Large scale interactive motion forecasting for autonomous driving: The waymo open motion dataset." In Proceedings of the IEEE/CVF International Conference on Computer Vision, pp. 9710-9719. 2021.
[8] Wilson, Benjamin, William Qi, Tanmay Agarwal, John Lambert, Jagjeet Singh, Siddhesh Khandelwal, Bowen Pan et al. "Argoverse 2: Next generation datasets for self-driving perception and forecasting." arXiv preprint arXiv:2301.00493 (2023).