---
license: mit
task_categories:
- robotics
tags:
- autonomous-driving
- carla
---

# MVAdapt Dataset

This repository contains the dataset for the paper [MVAdapt: Zero-Shot Multi-Vehicle Adaptation for End-to-End Autonomous Driving](https://huggingface.co/papers/2604.11854).

MVAdapt is a physics-conditioned adaptation framework for multi-vehicle end-to-end (E2E) autonomous driving. The dataset includes driving data for 27 different vehicles in the CARLA simulator, covering a wide range of physical properties such as size, mass, and drivetrain characteristics.

[GitHub Repository](https://github.com/hae-sung-oh/MVAdapt)

## Dataset Description

The dataset provides annotations and sensor data used to train and evaluate the MVAdapt model. It includes the following fields:

- `vehicle_id`: Vehicle model ids
- `gt_waypoint`: Ground truth waypoint for vehicle model
- `bs_waypoint`: Predicted waypoint from baseline model for default vehicle model
- `gt_control`: Ground truth control for vehicle model
- `bs_control`: Predicted control from baseline model for default vehicle model
- `scene_features`: Features that extracted by backbone model (TransFuser)
- `physics_params`: Physical properties for vehicle model
- `gear_params`: Gear properties for vehicle model
- `rgb`: RGB image 
- `lidar_bev`: LiDAR BEV image
- `target_point`: Target heading point
- `ego_vel`: Speed for ego vehicle
- `command`: Command for ego vehicle

## Citation

If you find this work or dataset useful, please consider citing:

```bibtex
@misc{oh2026mvadaptzeroshotmultivehicleadaptation,
      title={MVAdapt: Zero-Shot Multi-Vehicle Adaptation for End-to-End Autonomous Driving}, 
      author={Haesung Oh and Jaeheung Park},
      year={2026},
      eprint={2604.11854},
      archivePrefix={arXiv},
      primaryClass={cs.RO},
      url={https://arxiv.org/abs/2604.11854}, 
}
```