TruckV2X: A Truck-Centered Perception Dataset

Paper

Project Page

TruckV2X is the first large-scale truck-centered cooperative perception dataset, designed to address unique perception challenges in autonomous trucking (e.g., extensive blind spots, occlusions from large vehicle sizes, and dynamic trailer movements). It features multi-modal sensing (LiDAR and cameras) and supports multi-agent cooperation, including interactions between tractors, trailers, connected and automated vehicles (CAVs), and road-side units (RSUs). This dataset establishes performance benchmarks for heavy-duty vehicle scenarios and accelerates research on multi-agent autonomous trucking systems.

Usage Instructions

Option 1: Load via Hugging Face Hub

You can directly load the dataset using the datasets library for quick experimentation or exploration:

from datasets import load_dataset

dataset = load_dataset("XieTenghu1/TruckV2X", trust_remote_code=True)
# For security reasons, 🤗 Datasets do not allow running dataset loading scripts by default, 
# and you have to pass trust_remote_code=True to load datasets that require running a dataset script.

Note: This method may be more user-friendly and convenient for quick access, but it does not expose the raw file structure, which is required by downstream frameworks like OpenCOOD for direct usage.

Option 2: Clone + Local Extraction (Recommended)

For full flexibility and seamless integration with cooperative perception frameworks (e.g., OpenCOOD), we strongly recommend downloading and using the dataset locally:

# Clone the dataset repository
git lfs install
git clone https://huggingface.co/datasets/XieTenghu1/TruckV2X
cd TruckV2X

# Unzip all splits (each Town is separately zipped)
# Unzip all zip files under train/
for zipfile in train/*.zip; do
  dirname="${zipfile%.zip}"
  unzip "$zipfile" -d "$dirname"
done

# Unzip all zip files under val/
for zipfile in val/*.zip; do
  dirname="${zipfile%.zip}"
  unzip "$zipfile" -d "$dirname"
done

# Unzip all zip files under test/
for zipfile in test/*.zip; do
  dirname="${zipfile%.zip}"
  unzip "$zipfile" -d "$dirname"
done

Dataset Structure

The dataset is organized into training, validation and testing subsets with the following structure:

TruckV2X/
├── train/
│   ├── Town1_1/
│   │   ├── cav/
│   │   │   ├── 000000_camera0.jpg
│   │   │   ├── 000000_camera1.jpg
│   │   │   ├── 000000_camera2.jpg
│   │   │   ├── 000000_camera3.jpg
│   │   │   ├── 000000_lidar0.pcd
│   │   │   ├── 000000.yaml
│   │   │   ├── 000001_xxx
│   │   │   ├── ...
│   │   ├── roadside/
│   │   │   ├── 000000_camera0.jpg
│   │   │   ├── 000000_camera1.jpg
│   │   │   ├── 000000.yaml
│   │   │   ├── 000001_xxx
│   │   │   ├── ...
│   │   ├── tractor/
│   │   │   ├── 000000_camera0.jpg
│   │   │   ├── 000000_camera1.jpg
│   │   │   ├── 000000_camera2.jpg
│   │   │   ├── 000000_camera3.jpg
│   │   │   ├── 000000_camera4.jpg
│   │   │   ├── 000000_lidar0.pcd
│   │   │   ├── 000000_lidar1.pcd
│   │   │   ├── 000000.yaml
│   │   │   ├── 000001_xxx
│   │   │   ├── ...
│   │   ├── trailer/
│   │   │   ├── 000000_camera0.jpg
│   │   │   ├── 000000_camera1.jpg
│   │   │   ├── 000000_camera2.jpg
│   │   │   ├── 000000_camera3.jpg
│   │   │   ├── 000000_camera4.jpg
│   │   │   ├── 000000_lidar0.pcd
│   │   │   ├── 000000_lidar1.pcd
│   │   │   ├── 000000.yaml
│   │   │   ├── 000001_xxx
│   │   │   ├── ...
│   ├── Town1_4/
│   ├── ...
├── val/
│   ├── Town3_0/
│   ├── ...
├── test/
│   ├── Town1_0/
│   ├── ...

License

This dataset is licensed under the MIT License.

Citation

If you use TruckV2X or find our work inspiring in your research, please cite our paper:

@ARTICLE{11096563,
  author={Xie, Tenghui and Song, Zhiying and Wen, Fuxi and Li, Jun and Liu, Guangzhao and Zhao, Zijian},
  journal={IEEE Robotics and Automation Letters}, 
  title={TruckV2X: A Truck-Centered Perception Dataset}, 
  year={2025},
  volume={10},
  number={9},
  pages={9312-9319},
  keywords={Agricultural machinery;Laser radar;Cameras;Robot sensing systems;Roads;Vehicle dynamics;Sensors;Safety;Vehicle-to-infrastructure;Training;Vehicle-to-everything;cooperative perception;autonomous trucking;dataset},
  doi={10.1109/LRA.2025.3592884}}