The Future of Robot Rescue Simulation - Software day 5

Installation requirements

The workshop has several tutorials and hands-on assignments, which require simulation and programming environments. This page describes the software-packages which should be installed on your computer for the fourth day of the workshop.

What is needed to be prepared for the workshop, is Linux laptop with Ubuntu 14.04.4 LTS (Trusty 64bits), with ROS Indigo and Gazebo5. On top of that you have installed the GazeboUSARSim plugin and the Matlab Robotics System Toolbox. On this page we assume that you installed those packages following the previous instructions.

Software for day 5

To have multiple robots in their own namespace, you could follow the instructionof the pioneer3at_demo provided by Stefan Kohlbrecher and Nate Koenig. Yildiz demonstrated that this worked with their example code.

What is needed for to control a real robot is a working version of the simulation of a Nao robot.

Follow the steps of the ros Nao installation
mkdir ~/naoqi
Download pynaoqi-python2.7-2.1.4.13-linux64.tar.gz
cp ~/Downloads/pynaoqi-python2.7-2.1.4.13-linux64.tar.gz ~/naoqi
cd ~/naoqi
tar xzf pynaoqi-python2.7-2.1.4.13-linux64.tar.gz
echo 'export PYTHONPATH=~/naoqi/pynaoqi-python2.7-2.1.4.13-linux64:$PYTHONPATH' >> ~/.bashrc
python
inside python shell
>>> from naoqi import ALProxy
>>> quit()

If this works, the NaoQi Python bindings are correctly installed. Continue with the ROS packages for the Nao robot.

sudo apt-get install ros-indigo-nao-robot
sudo apt-get install ros-indigo-nao-bringup
sudo apt-get install ros-indigo-naoqi-bridge
sudo apt-get install ros-indigo-nao-extras

Correct working of the naoqi_driver can be verified by:

ROS_IP= roscore
rosrun naoqi_driver naoqi_driver_node --qi-url=tcp://:9559 --roscore_ip --network_interface
rosnode info /nao_robot
kill driver

Correct working of the full nao_bringup can be verified by:

Download world.stl
roslaunch nao_bringup nao_full_py.launch nao_ip:= roscore_ip:=
rosrun rviz rviz
add the MoveIt as MotionPlanner. add world.stl as Scene Object:
You should be able to see the camera image inside rviz.

Hands-on

If you like to have fun, try to plan to grab the door with the Nao robot or to recognize a tomato.

Alternatively, try rocon gateway, multimaster fkie or topic proxy.

If this communication works, it could be included into a new version of the Rules.

Alternative versions

The packages used during this workshop have complex dependencies. It is well possible that you don't have to downgrade from ROS Gravy or upgrade from Gazebo5 and still be succesfull in the workshop. Actually, such alternative installation could give usefull information on the sensitivities and dependencies of the packages, so please record your experience. Yet, our experience is that the least time is wasted when you upgrade/downgrade to the versions which are recommended here.

Workshop Program

Tutorial Preparation

Masaru Shimizu, Chukyo University, Aichi, Japan
Lauren Tabolinsky & Arie Weeren - MathWorks
Nate Koenig - Open Source Robotics Foundation
Satoshi Kochiyama - New Energy and Industrial Technology Development Organization, Japan
Stefan Kohlbrecher - TU Darmstad, Germany
Ricardo Tellez - The Construct Sim
Arnoud Visser, Universiteit van Amsterdam, The Netherlands