English | 日本語

sciurus17_examples

This package includes examples to control Sciurus17 using sciurus17_ros.

How to launch Sciurus17 base packages

1. Connect cables of a head camera, a chest camera and a control board to a PC.
2. Power on the Sciurus17 and the camera device names are shown in the /dev directory.
3. Open terminal and launch sciurus17_bringup.launch of sciurus17_bringup package.

This launch file has arguments:

• use_rviz (default: true)
  
• use_head_camera (default: true)
  
• use_chest_camera (default: true)
  

Using virtual Sciurus17

To launch Sciurus17 base packages without Sciurus17 hardware, unplug the control board's cable from the PC, then launch nodes with the following command:

roslaunch sciurus17_bringup sciurus17_bringup.launch

Using real Sciurus17

Launch the base packages with the following command:

roslaunch sciurus17_bringup sciurus17_bringup.launch

Using without cameras

Launch the base packages with arguments:

roslaunch sciurus17_bringup sciurus17_bringup.launch use_head_camera:=false use_chest_camera:=false

Using without RViz

To reduce the CPU load of the PC, launch the base packages with arguments:

roslaunch sciurus17_bringup sciurus17_bringup.launch use_rviz:=false

Using Gazebo simulator

Launch the packages with the following command:

roslaunch sciurus17_gazebo sciurus17_with_table.launch

# without RViz
roslaunch sciurus17_gazebo sciurus17_with_table.launch use_rviz:=false

Run Examples

Following examples will be executable after launch Sciurus17 base packages.

• gripper_action_example
• neck_joint_trajectory_example
• waist_joint_trajectory_example
• pick_and_place_demo
• hand_position_publisher
• head_camera_tracking
• chest_camera_tracking
• depth_camera_tracking
• preset_pid_gain_example
• box_stacking_example

gripper_action_example

This is an example to open/close the grippers of the two arms.

Run a node with the following command:

rosrun sciurus17_examples gripper_action_example.py

Videos

back to example list

neck_joint_trajectory_example

This is an example to change angles of the neck.

Run a node with the following command:

rosrun sciurus17_examples neck_joint_trajectory_example.py

Videos

back to example list

waist_joint_trajectory_example

This is an example to change angles of the waist.

Run a node with the following command:

rosrun sciurus17_examples waist_joint_trajectory_example.py

Videos

back to example list

pick_and_place_demo

This is an example to pick and place a small object with right hand while turning the waist.

Run a node with the following command:

rosrun sciurus17_examples pick_and_place_right_arm_demo.py

This is an example to pick and place a small object with left hand.

Run a node with the following command:

rosrun sciurus17_examples pick_and_place_left_arm_demo.py

This is an example to pick and place a small object with both hands.

Run a node with the following command:

rosrun sciurus17_examples pick_and_place_two_arm_demo.py

Videos

back to example list

hand_position_publisher

This is an example to receive link positions from tf server.

This example receives transformed positions l_link7 and r_link7 based on base_link from tf server, then publishes these positions as topics named /sciurus17/hand_pos/left and /sciurus17/hand_pos/right.

Run a node with the following command:

rosrun sciurus17_examples hand_position_publisher_example.py

back to example list

head_camera_tracking

This is an example to use the head camera images and OpenCV library for ball tracking and face tracking.

Firstly, install the OpenCV library of Python with the following command:

pip2 install opencv-python

Then, run a node with the following command:

rosrun sciurus17_examples head_camera_tracking.py

For ball tracking

Edit ./scripts/head_camera_tracking.py as follows:

def _image_callback(self, ros_image):
    # ...

        # Detect an object (specific color or face)
        output_image = self._detect_orange_object(input_image)
        # output_image = self._detect_blue_object(input_image)
        # output_image = self._detect_face(input_image)

For face tracking

Edit ./scripts/head_camera_tracking.py as follows:

This example uses Cascade Classifier for face tracking.

Please edit the directories of cascade files in the script file. USER_NAME depends on user environments.

class ObjectTracker:
    def __init__(self):
        # ...

        # Load cascade files
        # Example:
        # self._face_cascade = cv2.CascadeClassifier("/home/USER_NAME/.local/lib/python2.7/site-packages/cv2/data/haarcascade_frontalface_alt2.xml")
        # self._eyes_cascade = cv2.CascadeClassifier("/home/USER_NAME/.local/lib/python2.7/site-packages/cv2/data/haarcascade_eye.xml")
        self._face_cascade = cv2.CascadeClassifier("/home/USER_NAME/.local/lib/python2.7/site-packages/cv2/data/haarcascade_frontalface_alt2.xml")
        self._eyes_cascade = cv2.CascadeClassifier("/home/USER_NAME/.local/lib/python2.7/site-packages/cv2/data/haarcascade_eye.xml")

def _image_callback(self, ros_image):
    # ...

        # Detect an object (specific color or face)
        # output_image = self._detect_orange_object(input_image)
        # output_image = self._detect_blue_object(input_image)
        output_image = self._detect_face(input_image)

Videos

This orange ball can be purchased from this page in RT ROBOT SHOP.

back to example list

chest_camera_tracking

This is an example to use the chest camera images and OpenCV library for ball tracking.

Firstly, install the OpenCV library of Python with the following command:

pip2 install opencv-python

Then, run a node with the following command:

rosrun sciurus17_examples chest_camera_tracking.py

Execute face tracking and ball tracking simultaneously

Launch nodes with the following commands for face tracking with the head camera and for ball tracking with the chest camera.

rosrun sciurus17_examples head_camera_tracking.py

# Open another terminal
rosrun sciurus17_examples chest_camera_tracking.py

Videos

back to example list

depth_camera_tracking

This is an example to use the depth camera on the head for object tracking.

Firstly, install the OpenCV library of Python with the following command:

pip2 install opencv-python

Then, run a node with the following command:

rosrun sciurus17_examples depth_camera_tracking.py

The default detection range is separated into four stages.

To change the detection range, edit ./scripts/depth_camera_tracking.py as the followings:

    def _detect_object(self, input_depth_image):
        # Limitation of object size
        MIN_OBJECT_SIZE = 10000 # px * px
        MAX_OBJECT_SIZE = 80000 # px * px

        # The detection range is separated into four stages.
        # Unit: mm
        DETECTION_DEPTH = [
                (500, 700),
                (600, 800),
                (700, 900),
                (800, 1000)]

back to example list

preset_pid_gain_example

This is an example to change PID gains of the servo motors in bulk using preset_reconfigure of sciurus17_control.

Lists of PID gain preset values can be edited in sciurus17_control/scripts/preset_reconfigure.py.

Launch nodes preset_reconfigure.py and preset_pid_gain_example.py with the following command:

roslaunch sciurus17_examples preset_pid_gain_example.launch

back to example list

box_stacking_example

This is an example to detect boxes via PointCloudLibrary and stack the boxes.

Launch nodes with the following command:

roslaunch sciurus17_examples box_stacking_example.launch

To visualize the result of box detection, please add /sciurus17/example/markers of visualization_msgs/MarkerArray in Rviz.

Videos

back to example list