Advanced Manipulation Learning (AML)

Live Docs for AML

• Live documentation for AML can be found here
• Also see the docs folder in the AML root directory for the doxygen documentation.

Setting up AML - The Very Simple Way

• A set of scripts for installing AML natively in your host machine, or for getting it setup in a docker container can be found here

Setting AML manually

Dependencies

• ROS (Indigo)
• BaxterSDK
• Baxter Simulator
• Saywer Robot

Python Libraries

• numpy
• numpy-quaternion
• pygame
• PySide
• pybox2d
• Pillow
• Scipy
• Pandas
• six
• decorator
• matplotlib
• ipython
• cv2

This document lists various setup instructions after a fresh installation of Ubuntu 14.04 on your machine. The end part of the document also contains some of the possible errors during installation and their solutions.

1. Installing ROS - Indigo: Follow instructions on this page. Important Note: install “desktop-full” version

2. Installing virtual environment

   • Install packages

   sudo apt-get install python-setuptools
   sudo easy_install pip
   sudo apt-get install python-pip
   sudo pip install virtualenv
   sudo pip install virtualenvwrapper
   

   • Add the following two lines in your ~/.bashrc script:

   export WORKON_HOME=~/.venvs  
   source /usr/share/virtualenvwrapper/virtualenvwrapper.sh  
   export PIP_VIRTUALENV_BASE=~/.venvs
   

   • Close the bashrc file and source them:

   source ~/.bashrc
   mkvirtualenv robotics
   workon robotics
   

3. Install CUDA 8.0 and Cuda-NN 5.1

   You can get the compiled binary files from following websites:

   a. CUDA

   b. CUDANN

   • You will have to create an account in Nvdia to download the libraries

   • Goto "Download cuDNN v5.1 (Jan 20, 2017), for CUDA 8.0", Dowload cuDNN v5.1 Library for Linux

   • From the terminal goto the download directory

      tar -zxf cudnn-8.0-linux-x64-v5.1.tgz
      cd cuda/
      sudo cp include/* /usr/local/cuda/include/
      sudo cp lib64/* /usr/local/cuda/lib64/
     

   • Add the following bit to ~/.bashrc file

      export CUDA_HOME=/usr/local/cuda-8.0 
      export LD_LIBRARY_PATH={CUDA_HOME}/lib64 
      PATH={CUDA_HOME}/bin:{PATH} 
      export PATH
     

4. Install tensorflow in virtual environment

   workon robotics
   pip install --upgrade tensorflow-gpu
   

5. Create a new workspace and clone aml repository in it

   mkdir ~/catkin_ws
   cd catkin_ws
   mkdir baxter_ws
   cd baxter_ws
   mkdir src
   cd src
   git clone https://github.com/RobotsLab/AML.git
   

6. Baxter simulator setup

   • Dependencies

   sudo apt-get install gazebo2 ros-indigo-qt-build ros-indigo-driver-common ros-indigo-gazebo-ros-control ros-indigo-gazebo-ros-pkgs ros-indigo-ros-control ros-indigo-control-toolbox ros-indigo-realtime-tools ros-indigo-ros-controllers ros-indigo-xacro python-wstool ros-indigo-tf-conversions ros-indigo-kdl-parser
   

   • Simulator (specific versions)

   cd ~/catkin_workspaces/baxter_ws/src
   wstool init .
   wstool merge aml/3rdparty/baxter/rethink_packages.rosinstall
   wstool update
   wstool merge sawyer_robot/sawyer_robot.rosinstall
   wstool update
   

   • Check if any other unmet dependencies, run this line from ws folder

   rosdep install --from-path . --ignore-src --rosdistro indigo -y -r
   

   • Final step

   cd ../
   catkin_make
   

7. Few other dependencies

   pip install numpy numpy-quaternion pygame decorator ipython jupyter matplotlib Pillow scipy six PySide pandas
   pip install pybullet
   pip install git+git://github.com/pybox2d/pybox2d
   

8. Installing opencv for python

   Note: this compilation could take a while! And install this only after removing opencv-python (this is unofficial version) if installed previously.

   • Go to a folder of your choice

   git clone https://github.com/Itseez/opencv.git
   cd opencv
   git checkout 3.2.0
   mkdir build
   cd build
   cmake ..                             
   sudo make                   
   sudo make install         
   sudo /bin/bash -c 'echo "/usr/local/lib" > /etc/ld.so.conf.d/opencv.conf'  
   sudo ldconfig
   

   • Set symlink to virtual environment (on assumtion that your venv name is "robotics")

   cd ~/.venvs/robotics/lib/python2.7/site-packages/
   ln -s /usr/local/lib/python2.7/site-packages/cv2.so cv2.so
   

   • Check installation

   workon robotics
   python
   >> import cv2
   

   • To compile samples

   cd <opencv folder path>/opencv/samples
   cmake .               
   sudo make 
   

Possible Errors:

1. Could not find any downloads that satisfy the requirement tensorflow

   Solution: pip install --upgrade pip

2. No module named catkin_pkg.package

   Solution: pip install catkin_pkg

3. No module named rospkg

   Solution: pip install -U rospkg

4. ImportError: No module named 'em'

   Solution: pip install empty

5. Could not stop controller 'left_joint_velocity_controller' since it is not running

   Solution: goto ~/catkin_ws/baxter_ws/src/ baxter_gazebo/src/baxter_gazebo_ros_control_plugin.cpp

   Edit lines:::SwitchController::Request::STRICT to ::SwitchController::Request::BEST_EFFORT (This happens in two places)

   Note: You have to rebuild the catkin_make from baxter_ws