Official Documentation

• Checkout the docs

Dependencies

• ROS

  Go to this website and follow all the instructions there to install your ros environment.

  Then install catkin. I prefer catkin build to catkin make since it offers its own separation of build space advantages.

• Catkin build

  The best way to install catkin build is via PyPI. Do this in terminal:

    sudo pip install -U catkin_tools
  

• Build your workspace

  export ROS_DISTRO=`rosversion -d`              # Set ROS distribution
  mkdir -p ~/catkin_ws/src                       # Create workspace
  cd ~/catkin_ws/src                             # Navigate to source space
  # rosinstall_generator --deps ros_tutorials > .rosinstall  # Get list of packages
  # wstool update                                            # Checkout all packages
  cd ~/catkin_ws                                 # Navigate to ros workspace root
  catkin init                                              # Initialize workspace

Sphero ROS Build

cd to the src folder of your catkin_ws and then clone the master branch of this repo:

  cd ~/catkin_ws/src
  git clone https://github.com/service-lab/sphero-ros.git -b master

Now cd to the root of your catkin_ws folder and then build:

  cd ~/catkin_ws
  catkin build

Adding Paired Bluetooth Addresses to Launch file

Pair the bluetooth addresses for other sphero devices that you have using the bluetooth settings tab on your ubuntu distro, e.g. >> Dash + <search for BlueTooth + Click on the "+" sign to add other bluetooth devices>.

Now note the names and addresses of these devices that you have paired with and copy them into the the appropriate tabs in the multisphero launcher file.

Bringup each robot into roscore

  roslaunch sphero_node multi_sphero.launch

This should bring up each robot to the ros graph with each robot ready to receive commands on any of the topics

/sphero_xxx/cmd_vel
/sphero_xxx/collision
/sphero_xxx/disable_stabilization
/sphero_xxx/imu
/sphero_xxx/odom
/sphero_xxx/set_angular_velocity
/sphero_xxx/set_back_led
/sphero_xxx/set_color
/sphero_xxx/set_heading

Example Command Velocity Code

  rosrun sphero_run sphero_move.py

Vision based detection and tracking of the Sphero robots

Detection and tracking of the Sphero robots is done in the balls_detector package. The package is designed to support a Kinect or a generic Webcam. However, at this stage, the Kinect camera has not been tested nor fully supported. We used the Logitech C920 webcam to test the package, to reproduce our results, please use a webcam. If you decide to use Kinect, install Kinect libraries and dependencies by following the instructions on openkinect.

To use the balls_detector package with a webcam, run the nodes manually in terminal by follow the steps below:

1. Connect to the spheros and set the color of the LEDs to white at maximum brightness (255,255,255). You may configure the launch file in ~/catkin_ws/src/sphero_ros/sphero_ros/sphero_node/launch/multi_sphero.launch to connect to the desired robots.
2. Install v4l if not installed. This will allow us to modify the camera parameters. To install v4l, in a terminal window, run the following command:

sudo apt-install v4l-utils

3. In a new terminal window, change directory to your catkin workspace, then roslaunch the camera.launch file in balls_detector as such:

cd ~/catkin_ws/
roslaunch balls_detector camera.launch

4. When the launch file starts, a pop-up window should become visible. That window should be dark and only the Sphero robots should appear in it. The detection algorithm should display yellow circles on the Spheros once they are detected. If that does not happen, then the detection parameters need to be modified. To modify camera parameters, change them manually in /balls_detector/config/detect_param.yaml. Another window should pop-up. In that window, the Sphero robots should be displayed as they blink to tag them correctly. The robots will blink in the order they are specified in the file /balls_detector/config/tag_param.yaml. Make sure that the robot names are correct, that no extra robots exist, and that the correct number of robots is specified in that file. Tagging the robots takes some time during which the robots must remain in the frame and should not be moved. To view results use rostopic echo <topic_name> to display the data published to the topics (/locs/ordered, /locs/tagged, or /locs/detected)/.

Notes:

• The detection parameters are dependent on the height of the camera and the orientation of the robots. Modify the parameters based on camera position if necessary. When we tested it, the camera was placed on a tripod, and placed at a hight of 1.65 meters with the lens pointing downwards.
• In the first iteration, do not have any other objects in the frame, have only the Sphero robots present and colored white.

TODO:

Kinect support need to be extended.

3D reconstruction

To get a 3D reconstruct of the location of the robots, camera calibration must be performed at first. To do so, install the image_pipeline package by cloning the package from github using this link. Once installed, follow this tutorial to calibrate your camera. Save the data obtained and go to the location in whch it was stored. Extract the files and copy the generate YAML file as is to the directory ~/catkin_ws/src/sphero_ros/reconstruction/config. Name the file findNRT_config.yml. Then, run the roconstruction package though terminal by using roslaunch to launch reconstruction.launch

To get 3D reconstructed data, follow the following procedure:

1. Stop any previously started nodes.
2. Type in a terminal: roslaunch sphero_node multi_sphero.launch.
3. Type in a new terminal: roslaunch reconstruction reconstruction.launch
4. In a new terminal, type: roslaunch balls_detector camera.launch. This order should be followed. Failure to do so may prevent some nodes from starting correctly.