respeaker_ros

A ROS Package for Respeaker Mic Array

Supported Devices

• Respeaker Mic Array v2.0

  

Preparation

1. Build this package

   Assumed that ROS is properly installed.

   mkdir -p ~/catkin_ws/src && ~/catkin_ws/src
   git clone https://github.com/furushchev/respeaker_ros.git
   cd ~/catkin_ws
   source /opt/ros/kinetic/setup.bash
   rosdep install --from-paths src -i -r -n -y
   catkin config --init
   catkin build
   source ~/catkin_ws/devel/setup.bash

2. Register respeaker udev rules

   Normally, we cannot access USB device without permission from user space. Using udev, we can give the right permission on only respeaker device automatically.

   Please run the command as followings to install setting file:

   roscd respeaker_ros
   sudo cp -f $(rospack find respeaker_ros)/config/60-respeaker.rules /etc/udev/rules.d/60-respeaker.rules
   sudo systemctl restart udev

   And then re-connect the device.

3. Install python requirements

   roscd respeaker_ros
   sudo pip install -r requirements.txt

4. Update firmware

   git clone https://github.com/respeaker/usb_4_mic_array.git
   cd usb_4_mic_array
   sudo python dfu.py --download 6_channels_firmware.bin  # The 6 channels version 

5. Run executables

   roslaunch respeaker_ros respeaker.launch
   rostopic echo /sound_direction     # Result of DoA
   rostopic echo /sound_localization  # Result of DoA as Pose
   rostopic echo /is_speeching        # Result of VAD
   rostopic echo /audio               # Raw audio
   rostopic echo /speech_audio        # Audio data while speeching

   You can also set various parameters via dynamic_reconfigure.

   sudo apt install ros-kinetic-rqt-reconfigure  # Install if not
   rosrun rqt_reconfigure rqt_reconfigure

   To set LED color, publish desired color:

   rostopic pub /status_led std_msgs/ColorRGBA "r: 0.0
   g: 0.0
   b: 1.0
   a: 0.3"

Use cases

Voice Recognition

• ros_speech_recognition
• julius_ros

Author

Yuki Furuta <[email protected]>

License

Apache License