/docker-xenial-ros-vrep-custom-shell-env

customized Docker image with ROS Kinetic, OpenGL support, shared X11 socket and NVIDIA hardware acceleration. Used for testing V-REP integration with ROS.

Primary LanguageShell

Docker ROS Kinetic & V-REP

Use the Dockerfile and the build & run scripts to create and run a Docker container consisting of ROS Kinetic (full version) for Ubuntu Xenial with NVIDIA hardware acceleration, OpenGL support and shared X11 socket.

This image is used for testing the V-REP simulator with ROS Kinetic from within a Docker container.

This image combines the build steps of the following two images

Please refer to those repositories for more details about the respective images and the build process.

Contents

  1. Requirements
  2. Building the image
  3. Running the container
  4. V-REP - ROS interface

1. Requirements

This docker container has been build and tested on a machine running Ubuntu 16.04 with the following packages installed

Dependencies

This Docker image uses as base the custom docker_ros_kinetic_cuda9 image available from this repository.

2. Building the image

  1. Make sure you have built the docker_ros_kinetic_cuda9 image on your local machine, as mentioned in the Dependencies section above.
  2. In a terminal run ./build.sh to build a docker image with the name provided as the first argument and with a specified non-root user (default = docker) for the docker container.
    • If you have given a different name to the base image from step 1, make sure to change this line in the Dockerfile 
       FROM ros_kinetic_full_cuda9
      with the corresponding name.

3. Running the container

In a terminal enter ./run.sh with the image name from the previous step. This will run and remove the docker image upon exit (i.e., it is ran with the --rm flag).

./run.sh GIVEN_IMAGE_NAME

V-REP usage

Note that V-REP simulator is not installed nor copied in the Docker container. Rather, the host V-REP installation folder is volume mounted in the Docker container at runtime by adding the following line to the nvidia-docker run command (see the run.sh contents),

  -v $HOME/Projects/devs/v-rep-edu:/home/docker/v-rep

where you should replace the source path $HOME/Projects/devs/v-rep-edu with your own V-REP installation location.

A similar mounting can be done for the CATKIN Workspace directory, e.g., if you want to save the workspace data on the host side.

Testing functionality

Test V-REP

When the Docker container is launched, a terminator window opens. Navigate to $HOME/v-rep and type

./vrep.sh

The V-REP simulator GUI should launch.

Test ROS

While inside the container call roscore. The ros master should be launched.

4. V-REP - ROS Interface

There are two ways you can build the V-REP - ROS Interface

  1. Follow the online V-REP tutorial at
    • http://www.coppeliarobotics.com/helpFiles/en/rosTutorialIndigo.htm
    • the tutorial works fine also with ROS Kinetic and you should be able to build your own ROS Interface (libv_repExtRosInterface.so) and follow all the suggested examples.
    • note that in the example based on scene rosInterfaceTopicPublisherAndSubscriber.ttt the camera data is published on the topic /image and not visionSensorData, hence the correct command to visualize the image data published on the rostopic is
    $ rosrun image_view image_view image:=/image
  2. Follow the instructions in the README file provided with the V-REP installation.
    • For V-REP PRO EDU V3.4.0 rev1, this file is titled ros_vrep_rosinterface_install_guide.txt and can be found in your own V-REP's installation folder, at <v_rep_edu_folder>/programming/ros_packages/
    • Note that the ros_vrep_rosinterface_install_guide.txt is not up to date and you may encounter the following issues when trying to build the ROS interface yourself:
      • both mentioned git repositories have been moved to the CoppeliaRobotics github account
      • xsltproc package is required as mentioned in the dependency of the v_repStubsGen repository. The Dockerfile has been updated to perform this step for you.
      • no need to perform Step 5 - install v_repStubsGen (https://github.com/fferri/v_repStubsGen.git) if you clone the v_repExtRosInterface repository recursively by doing, 
        git clone --recursive https://github.com/CoppeliaRobotics/v_repExtRosInterface.git vrep_ros_interface
  3. Update the binaries in the mounted V-REP folder with the ones from the built catkin workspace
  4. Launch V-REP and monitor the terminal, you should see
./vrep.sh
...
Plugin 'RosInterface': loading...
Plugin 'RosInterface': load succeeded.
...
  1. Upon successful RosInterface load, checking the available nodes gives this:
$ rosnode list
/rosout
/vrep_ros_interface
  1. Test with the examples provided in the online tutorial