Skip to content
This repository has been archived by the owner on Jul 14, 2023. It is now read-only.

1. Setup Instructions

Cyril Anderson edited this page Sep 12, 2018 · 7 revisions

Table of Contents

MOVO User Guide

The MOVO User Guide is available on the kinova-movo Github repository as a zip package.

Instructions to access user guide:

  1. Download the zip package and extract the files onto your local computer
  2. Open the file index.html in your preferred web browser. (We recommend Firefox or Chrome for best results)

The User Guide, like the MOVO platform, is in BETA, and will be expanded over time. Please feel free to use "Issues" to share any comments on the docs and make any documentations requests.

MOVO Development Requirements

(OS and hardware): MOVO is officially supported only on Ubuntu 14.04 Trusty, x64 platform (core i5, core i7), 8GB RAM

Software Installations

You will need to install several development packages to develop on MOVO. Some of the packages are external to Kinova, and others are Kinova packages. If you experience any roadblocks with installation of Kinova packages, please report an issue here. For any issues specifically with external libraries, the support team for those libraries are probably the best bet for getting fixes. However, we are happy to answer any questions you have and offer help with any issues you have getting set up to develop with MOVO. This section will direct you through the installation processes.

There are two methods that can be used to install the software packages:

  • Automated installation script
  • Manual installation process

Automated installation script

To perform an automated installation, follow the steps on this page. After creating a workspace and cloning the kinova-movo repository, an interactive bash script - setup_remote_pc - is run to automate the installation.

If you experience issues with the automated installer, you can try with the manual install process. We encourage users to try the automated script first however.

Manual installation

ROS

Install ROS: ROS Ubuntu installation external link
following all steps in the Installation section. Perform the recommended install of ros-indigo-desktop-full, and make sure to follow the step to edit your .bashrc file, adding the ROS environment variables with source /opt/ros/indigo/setup.bash.

Install additional packages: sudo apt-get install ros-indigo-control-toolbox ros-indigo-moveit-core ros-indigo-costmap-2d ros-indigo-move-base ros-indigo-jsk-recognition ros-indigo-controller-manager ros-indigo-gazebo-ros* ros-indigo-hector* ros-indigo-rviz-imu-plugin ros-indigo-robot-pose-ekf ros-indigo-robot-localization ros-indigo-yocs-cmd-vel-mux ros-indigo-joint-* ros-indigo-grasp-planning-graspit-ros ros-indigo-jaco-graspit-sample ros-indigo-graspit-tools libsuitesparse-dev ros-indigo-costmap-converter ros-indigo-libg2o ros-indigo-eigen-conversions ros-indigo-eigen-stl-containers libeigen3-dev ros-indigo-trac-ik ros-indigo-moveit ros-indigo-ros-controllers ros-indigo-map-server ros-indigo-amcl ros-indigo-global-planner

Note: It is normal for the ROS installation to take a relatively long time. There are a lot of packages to install. 20-30 minutes is not unusual for this step.

MoveIt!

Install MoveIt!: MoveIt! installation external link Follow instructions for Indigo

Gazebo

Install Gazebo: Gazebo installation external link

Freenect2

Install Freenect2:

  • Create a directory Software with a sub-directory kinova.
    mkdir -p ~/Software/kinova .
  • cd ~/Software .
  • Get libfreenect2. Follow the installation instructions for Linux, downloading and installing required components. Follow the instructions for Ubuntu 14.04.
  • In terms of optional components, follow the instructions to Install OpenNI2, and, if possible, install OpenCL. You do not need CUDA. Do NOT follow the steps to install CUDA. The libfreenect2 instructions will have you download the wrong NVIDIA drivers, and will cause you to waste a day trying to figure out why your computer has locked you out.
  • Follow the instructions to build freenect. Replace the line cmake .. -DCMAKE_INSTALL_PREFIX=$HOME/freenect2 with cmake .. -DCMAKE_INSTALL_PREFIX=~/Software/freenect2 -Dfreenect2_DIR=~/Software/freenect2/lib/cmake/freenect2
  • Once built and installed, create a symlink to link the library to cmake: sudo ln -sf ~/Software/freenect2/lib/cmake/freenect2/freenect2Config.cmake /usr/share/cmake-2.8/Modules/Findfreenect2.cmake (or cmake-3.0, be sure to check which cmake you are using first, looking in /usr/share/)
  • Check to make sure your symlink worked: ls -l /usr/share/cmake-2.8/Modules/Findfreenect2.cmake
  • Edit the file .bashc file to add the line export freenect2_DIR=~/Software/freenect2/lib/cmake/freenect2/

You might encounter the following error during the make process: Could not find a package configuration file provided by "Eigen3" with any of the following names: Eigen3Config.cmake eigen3-config.cmake

Make sure the FindEigen3.cmake file is in your cmake folder. The Eigen3 installer assumes the system uses cmake-2.8: sudo cp /usr/share/cmake-2.8/Modules/FindEigen3.cmake /usr/share/<cmake-your_version>/Modules/

Kinova MOVO package Install on development computer

Install the kinova-movo package

  • cd ~/Software/kinova and clone the repository: git clone https://github.com/Kinovarobotics/kinova-movo.git
  • Create a workspace: mkdir ~/movo_ws
  • Link the MOVO files to the workspace: ln -s ~/Software/kinova/kinova-movo/ ~/movo_ws/src/ OR change the folder name: mv ~/Software/kinova/kinova-movo/ ~/movo_ws/src
  • cd ~/movo_ws/src

Edit the items that are needed

  • Open your .bashrc file and add: source ~/movo_ws/devel/setup.bash .
  • Re-source your .bashrc with source ~/.bashrc.
  • Set the values for configuration variables in movo_common/movo_config/movo_config.sh as needed. Note: this is where you can set how many arms the robot has, the gripper type, pan tilt, etc.

Create a ROS catkin workspace:

  • cd ~/movo_ws/src
  • Open your .bashrc and add: source /opt/ros/indigo/setup.bash . Make sure to re-source ~/.bashrc .
  • catkin_init_workspace - this will initialize the catkin workspace, creating a top-level CMakeLists.txt file. NOTE: If you are using the Anaconda Python distribution, there is a known conflict with ROS that will prevent catkin from working properly. To work around this, go into ~/.bashrc , find a line that looks like export PATH=/home/username/anaconda3/bin:$PATH. Comment this line out. (Remember to re-source your .bashrc file afterwards). To use Anaconda again, simply uncomment the line.
  • Build the workspace:
    • Change directory back to the root of the ROS catkin workspace, ~/movo_ws with: cd ..
    • Build the package: catkin_make

Connecting to MOVO platform

  • The development PC can be connected to MOVO in one of two ways:
    • As a wired connection using the RJ45 connector (Ethernet) on the HMI, or
    • Wirelessly, via wifi (SSID: MoVoWifi password: Welcome00)
  • Remote SSH onto movo: ssh [email protected] You will be prompted for the password; use Welcome00
  • To enable tele-operation with the joystick, connect on the development computer. Open a terminal on the dev computer and type:
    • cd ~/movo_ws
    • sws # alias sws='source ./devel/setup.bash' defined in .bashrc
    • roslaunch movo_remote_teleop movo_remote_teleop.launch
  • Enabling tele-operation with the joystick connect directly on movo. Open an ssh session on MOVO as described above, and:
    • cd ~/movo_ws
    • sws # alias sws='source ./devel/setup.bash' defined in .bashrc
    • roslaunch movo_remote_teleop movo_remote_teleop.launch