HPeC project

The HPeC project aims at demonstrating the relevancy of self-adaptive hardware architectures to respond to the growing demands of high performance computing, in an increasing class of embedded systems that also have demanding footprint and energy efficiency constraints.

This github repository aims at improving the teamwork in the project. It helps by providing the same files at the whole team in order to avoid version and compliance issues.

Getting Started

These instructions will get you a copy of the project up and running on your local machine for development and testing purposes.

Please contact me if you encounter an unexpected error at erwan.moreac@univ-ubs.fr.

1. Prerequisites

• Ubuntu 16.04 (Xenial) or equivalent (Debian Jessie)

There is no guarantee that the following project runs in another distro than Ubuntu 16.04.

You need to install the following applications:

• Git

• ROS kinetic

  • Gazebo 7, provided with kinetic full install
  • MAVLINK, provided with kinetic full install

• SITL ardupilot

• MAVROS

• Additional packages:

1. We need octomap packages to compile the project

$ sudo apt-get install ros-kinetic-octomap ros-kinetic-octomap-mapping ros-kinetic-octomap-msgs ros-kinetic-octomap-ros ros-kinetic-octomap-rviz-plugins ros-kinetic-octomap-server

2. In order to run gazebo and ros this package is required:

$ sudo apt-get install ros-kinetic-gazebo-ros-pkgs

3. Geographic lib is mandatory to execute MAVROS, to install the dataset, run the script at the following location:

/opt/ros/kinetic/lib/mavros/install_geographiclib_datasets.sh

4. We need cmake modules to use the catkin_make compilation:

$ sudo apt-get install ros-kinetic-cmake-modules

As soon as all programs have been installed, we configure the terminal environment by editing the .bashrc. First, we source ROS to get all commands:

$ echo 'source /opt/ros/kinetic/setup.bash' >> ~/.bashrc

Second, we source gazebo setup with:

$ echo 'source /usr/share/gazebo/setup.sh' >> ~/.bashrc

Now, reload your .bashrc in your terminal:

$ source ~/.bashrc

2. Upgrading to Gazebo 8

This section is mandatory since the project includes some Gazebo plugins which depend of Gazebo 8 features. The first step is to remove the current gazebo 7 installed with ROS kinetic full install:

$ sudo apt-get purge gazebo7

We install gazebo 8 thereafter:

$ sudo apt-get install gazebo8

A next step is to get gazebo-ros packages for gazebo 8:

$ sudo apt-get install ros-kinetic-gazebo8-ros-pkgs

As well as for the package ros-control:

$ sudo apt-get install ros-kinetic-gazebo8-ros-control

Once it's done, you can check gazebo version by typing $ gazebo --version. To complete the upgrade, we MUST install the Plugin ardupilot_gazebo (or reinstall if you used gazebo 7).

3. Ardupilot-Gazebo plugin setup

Follow instructions specified in Plugin ardupilot_gazebo without forget to adapt commands to your Gazebo version.

4. Installing the project

This section aims at creating a catkin workspace and include sources from this repository. The environment configuration is introduced. Then, compiling and execution of the project are explained.

First step, create a catkin workspace directory and go inside it.

$ mkdir -p  ./catkin_ws
$ cd ./catkin_ws/

Clone the git repository in catkin_ws/ and rename it /src.

$ git clone https://github.com/Kamiwan/HPeC-sources.git ./src

Then, check if the repo is properly in place.

$ cd ./src/
$ git status

You should get something like "Your branch is up-to-date with 'origin/master'. Nothing to validate, the working copy is clean."

Now we configure the project to compile it.

Replace the path of image_transport lib by yours in the CMakeLists.txt of following packages:

• detection_tracking
• emergency_landing
• generic_task_wrapper
• harris_detector
• mission_manager
• rosnode_model
• stabilisation_imu

in the include directories

## Specify additional locations of header files
include_directories(
  ${catkin_INCLUDE_DIRS} 
  /Your/Path/image_transport
  ${OpenCV_INCLUDE_DIRS}
)

and to link executable library. Image_transport include directory and binary file are supposed to come from your ROS kinetic install path with something like /opt/ros/kinetic/[...].

## Specify libraries to link a library or executable target against
 target_link_libraries(the_actual_node
   ${catkin_LIBRARIES}
    /Your/Path/libimage_transport.so
   ${OpenCV_LIBRARIES}
 )

Then, replace the PROJECT_PATH in gazebo plugin by your in gazebo_plugin/gazebo_world_plugin.cpp without slash at the end

  #define PROJECT_PATH "/Your/Path"

Once every paths are updated you can compile the whole project with the following command at the root of your catkin workspace:

$ cd ../
$ catkin_make -j 1

The -j 1 option is to only use one core to compile since catkin can have some errors due to compile dependencies separately.

The command should end with:

[100%] Built target px4_comm_node_copter

Now your project directory is ready to run current installed packages and also to create new ones. :)

Running the tests

1. Configuration

Before to start, we configure paths in order to make Gazebo find HPeC models and worlds. Please adapt the path to your own setup. The first environment variable we need to update is GAZEBO_MODEL_PATH,

$ echo 'export GAZEBO_MODEL_PATH=/Your/Path/catkin_ws/src/mybot_gazebo/models:${GAZEBO_MODEL_PATH}' >> ~/.bashrc

and the other one is GAZEBO_RESOURCE_PATH,

$ echo 'export GAZEBO_RESOURCE_PATH=/Your/Path/catkin_ws/src/mybot_gazebo/models:/Your/Path/catkin_ws/src/mybot_gazebo/worlds:${GAZEBO_RESOURCE_PATH}' >> ~/.bashrc

Reload your .bashrc file.

$ source ~/.bashrc

Each time you compile your catkin_ws with $ catkin_make -j 1, you need to source your project:

$ source ./devel/setup.bash

We are finally ready for our first try!

2. Execution

Run roscore and gazebo

$ ./src/scripts/run_server_gazebo.sh

This command launches the roscore and gazebo with the world HPeC_sandbox.world. You should see in gazebo GUI a drone with sensors (range in blue) and some objects and surfaces.

Execute and use a node of the project

If you have at least launched a ROS Master with $ roscore or used the script run_server_gazebo.sh, you can in another terminal launch every nodes compiled by the HPeC project. In this example we will use the node of rosnode_model package.

First of all, go to your catkin_ws and source it with $ source ./devel/setup.bash. Then, you should be able to run:

$ rosrun rosnode_model rosnode_model_node 

If it worked the terminal printed this [INFO] [1528472534.171299944]: [TASK WRAPPER][RUNNING].

In the HPeC project, most of the nodes created can be run in several modes. Indeed, they can be executed in their software version or in the hardware one. Thus, when you launch the node with rosrun, it does nothing until you tell it in which mode start.

To control the node mode, you have to send him an integer in the right ROS Topic. To do so, you can check which topic is using the node in its source files or by using $ rostopic list.

If you only run the roscore and the rosnode_model_node, rostopic list gives this output:

/rosnode_model_mgt_topic
/rosnode_model_notification_topic
/rosout
/rosout_agg

This is the topic /rosnode_model_mgt_topic where we have to publish a message. On another terminal run:

$ rostopic pub  -1  /rosnode_model_mgt_topic std_msgs/Int32 "1"

• 0 to stop the application
• 1 to start the software version
• 2 to start the "hardware" version

The terminal where the node has been launched should print this kind of output,

[INFO] [1528474077.010573417]: [THREAD][RUNNING][SW]: rosnode_model SOFTWARE VERSION 
[INFO] [1528474077.039458231]: Could not read rosnode_model activation rate. Setting 1 Hz
Hello World!
SOFTWARE rosnode_model Processing time : 0.051
Hello World!
SOFTWARE rosnode_model Processing time : 0.027
...

Note that in this case the hardware version is not really a hardware mode, just another version of the program to give a ROS node model.

Run the simulation Loop ROS + GAZEBO + SITL + MAVROS

In order to run the complete simulation loop, we need at least 3 terminals. In each terminal is used one specific part. Follow commands for each one of them.

Terminal 1: ROS+Gazebo

$ cd /Your/Path/catkin_ws
$ source ./devel/setup.bash
$ ./src/scripts/run_nodes_gazebo.sh

It run the ROS Master, gazebo and several nodes which are essential to run the first scenario.

Terminal 2: SITL

$ cd /Your/Path/SITL/ardupilot/ArduCopter
$ python ../Tools/autotest/sim_vehicle.py -v ArduCopter -f gazebo-iris -m --mav10 --map --console -I0

This command should create 3 windows:

• xterm Terminal output of the autopilot
• console GUI with drone parameters
• map GUI map that locates the drone

/!\ IT IS VERY IMPORTANT TO WAIT THAT APM finishes the drone setup before to start MAVROS. It requires around one minute. The information is displayed in the console window and the following messages indicate when the configuration is over:

APM: EKF2 IMU0 is using GPS

APM: EKF2 IMU1 is using GPS

Terminal 3: MAVROS

$ roslaunch mavros px4.launch fcu_url:="udp://:14550@127.0.0.1:14551"

This application finishes the loop and allow ROS nodes to communicate navigation orders to the emulated autopilot SITL.

Now you should see the drone flying in the virtual world created by gazebo. :)

Troubleshooting

/!\ Work in progress

1. catkin_ws

CMake Error, error: could not find svn for checkout of opentld

You need to install subversion:

$ sudo apt-get install subversion

Then compile again with

$ catkin_make -j 1

2. Gazebo

Missing libArduPilotPlugin.so ... etc

In case you see this message when you launch gazebo with demo worlds, check you have no error after sudo make install the plugin ardupilot_gazebo. If no error use "ls" on the install path given to see if the plugin is really here. If this is correct, check with "cat /usr/share/gazebo/setup.sh" the variable GAZEBO_PLUGIN_PATH. It should be the same as the install path. If not use "cp" to copy the lib to right path.

For Example

sudo cp -a /usr/lib/x86_64-linux-gnu/gazebo-7.0/plugins/ /usr/lib/x86_64-linux-gnu/gazebo-7/

path mismatch is confirmed as ROS's glitch. It'll be fixed.

3. MAVROS

4. SITL

Additional links to learn how to use Git

Github link

https://github.com/

Online tutorials

Try github

Gitreal tutorial

A really good French git lesson

Openclassroom git

Git documentation

Using branches

How to merge your branch

Good practices with git

Best practices

Writing good commits

Git command summary

Git Cheat sheet

Paper

Hardware-in-the-loop simulation with dynamic partial FPGA reconfiguration applied to computer vision in ROS-based UAV