VLA-Model
construction-multimodel-dataset
Construction Traversability Dataset
A construction-site RGB semantic segmentation dataset developed for research on
terrain understanding, traversability estimation, and multimodal RGB–LiDAR
perception for mobile robots.
Overview
This dataset contains RGB images and pixel-wise semantic segmentation masks
collected in construction-site environments. The dataset is intended to support
research on construction-site scene understanding and traversability-aware
robot… See the full description on the dataset page: https://huggingface.co/datasets/vla-model/construction-multimodel-dataset.construction-syn-data
Construction-Site Multimodal Dataset (CMAD)
Real-world RGB + LiDAR + pose data collected on a live construction site with a custom ground-robot sensor rig. Three sequences, 16,401 synchronized frames (~78 minutes) of camera, LiDAR, and 6-DoF odometry, laid out in a directory/loader convention compatible with the public RELLIS-3D benchmark, so existing off-road traversability-estimation tooling built around that convention can consume it with only a root-path change.
This dataset… See the full description on the dataset page: https://huggingface.co/datasets/vla-model/construction-syn-data.construction-rosbags
Construction-Site Multimodal Robotics ROS 2 Recordings
The recordings preserve the multimodal sensor streams and their temporal relationships during real robot operation. All four recording sessions are closed-loop robot trajectories, providing continuous sensor data while the robot operated and revisited regions of the construction sites.
Dataset Structure
construction-traversability-rosbags/
├── site1/
│ ├── site1_session01/
│ │ ├── metadata.yaml
│ │… See the full description on the dataset page: https://huggingface.co/datasets/vla-model/construction-rosbags.vision_intelligence_vla_model_data