- Home
- Training data
- Robotics & Sensor
Robotics and sensor data.
Demonstrations and multi-sensor logs for imitation learning, perception, and vision-language-action models, with observations, actions, and calibration in sync.
What we can supply
- Robot teleoperation trajectories
- Manipulation and grasping demonstrations
- Autonomous driving multi-sensor logs
- LiDAR and 3D point clouds
- IMU and wearable motion data
- Aerial and drone navigation
- Sim-to-real paired data
- Egocentric video for embodied AI
Labels and enrichment
- Synced observations and actions
- Proprioception
- Grasp pose and force
- 3D boxes and tracks
- Per-point semantics
- Calibration
- Task success
Formats and delivery
ROS bag, MCAP, or HDF5 with JSON metadata, or converted to your layout. Delivered to AWS S3, Google Cloud Storage, Azure, or your own storage, with a datasheet and provenance records.
Common uses
- Imitation learning
- Vision-language-action models
- Perception
- Sensor fusion
- Mapping and SLAM
- Activity recognition
Questions
Can you collect on our hardware setup?
Tell us the robot, gripper, and sensor configuration, and we collect on matching or comparable hardware.
Is calibration included?
Every multi-sensor delivery includes calibration and time-sync data.
What formats do you deliver?
The format your stack uses, including ROS bag, MCAP, or HDF5, with metadata as JSON.
Own data like this? License it through ModelCore