ROS Integration

This section covers the ROS integration capabilities of DexRobot MuJoCo, which provides seamless control and monitoring through ROS topics and services.

Contents

Overview

The ROS node (dexrobot_mujoco_ros.py) provides:

  • Joint and pose control interfaces

  • State feedback and monitoring

  • Data recording (CSV, ROS bags, MP4)

  • VR visualization capabilities

  • Automatic ROS1/ROS2 compatibility via ros_compat

Core Features

Joint Control

  • Direct joint position control via joint_commands

  • Hand pose control with 6-DoF floating base

  • Position scaling and velocity damping

  • Real-time position tracking

State Feedback

  • Joint state publishing

  • Body pose tracking

  • Touch sensor data

  • Custom sensor outputs

Data Recording

  • CSV file recording

  • ROS bag recording

  • MP4 video recording

  • Synchronized multi-modal data

VR Visualization

  • Stereo image rendering

  • Web-based visualization

  • Real-time streaming

  • Adjustable viewpoints

Quick Start

Basic Usage

  1. Launch the node:

    python nodes/dexrobot_mujoco_ros.py model.xml

  2. Send joint commands:

    # ROS2
ros2 topic pub /joint_commands sensor_msgs/msg/JointState \
    "{name: ['r_f_joint1_1'], position: [0.5]}"

# ROS1
rostopic pub /joint_commands sensor_msgs/JointState \
    "{name: ['r_f_joint1_1'], position: [0.5]}"

  3. Monitor states:

    # ROS2
ros2 topic echo /joint_states
ros2 topic echo /body_poses
ros2 topic echo /touch_sensors

# ROS1
rostopic echo /joint_states
rostopic echo /body_poses
rostopic echo /touch_sensors

Advanced Usage

Enable VR and recording:

python nodes/dexrobot_mujoco_ros.py model.xml \
    --enable-vr \
    --output-formats ros csv mp4 \
    --output-csv-path data.csv \
    --output-mp4-path video.mp4 \
    --output-bag-path recording.bag

Next Steps

Continue reading: