MBZIRC2020

Sant'Anna team drone for the MBZIRC 2020 challenge

Installation

Basic dependencies

sudo apt install build-essential python-dev python-pip
sudo apt install cmake git mercurial

Install ROS

sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'
sudo apt-key adv --keyserver hkp://ha.pool.sks-keyservers.net:80 --recv-key C1CF6E31E6BADE8868B172B4F42ED6FBAB17C654
sudo apt update
sudo apt install ros-melodic-desktop-full
sudo rosdep init
rosdep update
echo "source /opt/ros/melodic/setup.bash" >> ~/.bashrc
source ~/.bashrc

Install some ROS packages

sudo apt install python-rosinstall python-rosinstall-generator python-wstool python-catkin-tools
sudo apt install ros-melodic-octomap-ros
sudo apt install ros-melodic-mavros ros-melodic-mavros-extras

Prepare files to store env variables

cd ~
touch .bash_mbzirc

Open the file .bashrc and append the following lines:

if [ -f ~/.bash_mbzirc ]; then
    . ~/.bash_mbzirc
fi

To find the file you have to go in your /home folder and click on "show hidden files".

Ardupilot

Now you have to install Ardupilot in a directory of your choice, let's say the home:

cd ~
git clone --recurse-submodules https://github.com/ArduPilot/ardupilot
cd ardupilot
chmod a+x Tools/environment_install/install-prereqs-ubuntu.sh
Tools/environment_install/install-prereqs-ubuntu.sh -y
source ~/.profile

(Optional) Test Ardupilot

If you want to verify that your Ardupilot installation was successful:

cd ArduCopter
sim_vehicle.py -j4 --map --console

And in the terminal type:

mode GUIDED
arm throttle
takeoff 100

Download Gazebo DB

The next step is to install the predefined models of Gazebo. Feel free to choose the directory where to do so (e.g. the home):

cd ~
hg clone https://bitbucket.org/osrf/gazebo_models
cd gazebo_models
hg checkout zephyr_demos
echo "export GAZEBO_MODEL_PATH=${GAZEBO_MODEL_PATH}:$PWD" >> ~/.bash_mbzirc
source ~/.bashrc

Initialize and configure catkin workspace

It's now time to prepare the actual ROS workspace for our drone.

WARNING: not all the required submodules are publicly released, so you need to ask for pemissions to the respective authors. Please contact @leoll2 for more details.

First of all, clone this repo. You can do it in any folder, we suggest the home.

mkdir drone_ws
cd drone_ws
git clone --recurse-submodules https://github.com/leoll2/MBZIRC2020.git src

Checkout to the correct branch of the vision module:

cd src/mbzirc_vision/MBZIRC2020Vision/
git checkout ros_integrated
cd -

Now build all the ROS modules: Possible add: sudo apt-get install ros-melodic-realtime_tools

catkin build
source devel/setup.bash

If you don't want to source devel/setup.bash on every terminal opening, just

echo "source $PWD/devel/setup.bash" >> ~/.bash_mbzirc

Tip: if you clone the repository by https, then for pushing/pulling you will be prompted to insert github username and pass every single time; to avoid this, you may use SSH to clone the repository (git@github.com:leoll2/MBZIRC2020.git), and add the ssh key to your local pc and github account. Obviously, you can always copy the repo via https and add the ssh remote later.|

Create links to Ardupilot and gazebo models folders (not strictly necessary, but handy), usually the path will be something like /home/your_username/ardupilot:

ln -s <path/to/ardupilot> src/ardupilot
ln -s <path/to/gazebo/models> src/gazebo_models

Add environment variables:

echo "export GAZEBO_MODEL_PATH=\${GAZEBO_MODEL_PATH}:$(rospack find mbzirc_gazebo)/gazebo_models" >> ~/.bash_mbzirc
echo "export GAZEBO_PLUGIN_PATH=\${GAZEBO_PLUGIN_PATH}:$(catkin locate)/build/mbzirc_gazebo" >> ~/.bash_mbzirc
echo "export GAZEBO_PLUGIN_PATH=\${GAZEBO_PLUGIN_PATH}:/usr/lib/x86_64-linux-gnu/gazebo-9/plugins/" >> ~/.bash_mbzirc
source ~/.bashrc

Add our vehicle parameters to Ardupilot

mkdir src/ardupilot/Tools/autotest/MBZIRC_params
cp src/mbzirc_gazebo/ardupilotStuff/MBZIRC-provaParams/gazebo-MBZIRColo.parm src/ardupilot/Tools/autotest/MBZIRC_params

Now edit src/ardupilot/Tools/autotest/pysim/vehicleinfo.py, search in the code for:

"gazebo-iris": {
  "waf_target": "bin/arducopter",
  "default_params_filename": ["default_params/copter.parm",
                              "default_params/gazebo-iris.parm"],
},

and add the following item underneat it:

# MBZIRColo
"gazebo-MBZIRColo":{
    "waf_target": "bin/arducopter",
    "default_params_filename":["default_params/copter.parm",
                                "MBZIRC_params/gazebo-MBZIRColo.parm"],
},

Test Gazebo + Ardupilot

Now go back to the ardupilot directory:

cd ArduCopter
sim_vehicle.py --console -f gazebo-MBZIRColo

Move to drone_ws\src\mbzirc_gazebo, then:

gazebo --verbose gazebo_worlds/drone_plus_env.world

START THE SIMULATION FRAMEWORK

Gazebo

First, start gazebo (+ ardupilot), having in mind the fact that gazebo is saving images on the file system. Refer to mbzirc_gazebo repo

Vision

Secondly, start the image processing part. Refer to MBZRIC2020Vision repo (inside the mbzirc_vision repo, this last one being in this repo)

leoll2 / MBZIRC2020