Sraghvi/dual-arm-robot-dataset

Dual-Arm Robot Dataset

A LeRobot-compatible dataset containing joint position data from a dual-arm robotic system with 34 degrees of freedom.

Dataset Description

This dataset contains robot manipulation data recorded from a dual-arm robotic system. The data includes joint positions for both arms and their associated grippers, captured at 100Hz for precise motion analysis and learning.

Dataset Summary

• Total Episodes: 1
• Total Frames: 3,105
• Duration: 31.05 seconds
• Sampling Rate: 100 Hz
• Robot Type: Dual-arm robot with grippers
• Total Joints: 34 degrees of freedom

Joint Configuration

The dataset includes the following joint groups:

Arm 1 (Left Arm):

• arm1_shoulder_pan_joint
• arm1_shoulder_lift_joint
• arm1_elbow_joint
• arm1_wrist_1_joint
• arm1_wrist_2_joint
• arm1_wrist_3_joint

Arm 2 (Right Arm):

• arm2_shoulder_pan_joint
• arm2_shoulder_lift_joint
• arm2_elbow_joint
• arm2_wrist_1_joint
• arm2_wrist_2_joint
• arm2_wrist_3_joint

Gripper Joints (Both Arms):

• 3-finger gripper joints for each arm
• Palm and finger joints for precise manipulation

Dataset Structure

The dataset follows the LeRobot format with the following structure:

data.h5
├── episode_0/
│   ├── observation.joint_positions    # Joint positions (3105, 34)
│   ├── action.joint_positions        # Actions (same as observations)
│   └── timestamp                     # Timestamps for each frame
info.json                             # Dataset metadata
modality.json                         # Modality definitions

Usage

Loading the Dataset

from datasets import load_dataset

# Load the dataset
dataset = load_dataset("Sraghvi/dual-arm-robot-dataset")
print(dataset)

# Access joint data
joint_positions = dataset["train"]["observation.joint_positions"]
print(f"Shape: {joint_positions.shape}")  # (3105, 34)

Using with LeRobot

from lerobot.common.datasets.lerobot_dataset import LeRobotDataset

# Load for LeRobot training
dataset = LeRobotDataset("Sraghvi/dual-arm-robot-dataset")

Data Collection

• Source: ROS2 bag files
• Robot System: Dual-arm robotic manipulator
• Recording Method: Real-time joint state capture
• Coordinate System: Robot joint space

Applications

This dataset is suitable for:

• Robot manipulation learning
• Dual-arm coordination studies
• Joint space trajectory analysis
• Reinforcement learning for robotics
• Imitation learning from demonstrations

Technical Details

• Data Format: HDF5
• Precision: Float32
• Coordinate System: Joint angles (radians)
• Missing Data: None
• Data Quality: High-frequency, continuous recording

Citation

If you use this dataset in your research, please cite:

@dataset{dual_arm_robot_dataset_2024,
  title={Dual-Arm Robot Dataset},
  author={Sraghvi},
  year={2024},
  url={https://huggingface.co/datasets/Sraghvi/dual-arm-robot-dataset}
}

License

This dataset is released under the MIT License.

Contact

For questions or issues with this dataset, please open an issue on the Hugging Face dataset page.