hd-faria / rosiChallenge-sbai2019

ROSI Challenge repository for the XIV SBAI

ROSI CHALLENGE - XIV SBAI

This repository contains the ROS package and V-REP files for running the ROSI CHALLENGE simulation.

The final stage of the competition will occur on the XIV SBAI, to be held in Ouro Preto (Brazil) in october 2019. One may find more info about SBAI in www.sbai2019.com.br.

Additionally, the content here may be beneficial for robotics classes. There is a complete mobile robot model with divers actuators and sensor fully communicating with the ROS framework and an industrial belt conveyor scenario. Feel free to use it for spreading the robotics knowledge in your classes! :)

We would like to thank Marc Freese and the Coppelia Robotics team for releasing the V-REP simulator for this challenge. Moreover, these noble gentlemen greatly contributed to this simulator: Amauri Coelho Ferraz, Raphael Pereira Figueiredo da Silva, and Wagner Ferreira Andrade.

DISCLAIMER

V-REP is now CoppeliaSim simulator. The straightforward application of this tutorial may be deprecated.

Remember that your solution has to be implemented in the ROS (Robot Operating System) package/nodes format. The code has to be well structured, readable and cataloged as it will be open-source available.

DO FORK THIS REPOSITORY only if are willing to propose corrections/improvements to the simulator. YOUR SOLUTION to the challenge has to be created on a CLEAN GITHUB REPOSITORY!

On the competition stage, the simulation will run on a server computer, and each teams' package will run on another computer. Be careful not to change any simulation parameters as the official one is the only one considered for the competition.

Be sure you have read and understood all of the challenge's rules available in resources/regulamento_rosiChallenge.pdf

Package description

This repository is structured as a ROS package. The folders organization is as follows:

• config - Contains the .yaml files with simulation parameters. You may change them accordingly to your needs.

• launch - Contains a ROS launch file example. Notice that rosi_joy.launch loads the parameters file from config folder.

• msg - Message files needed to interact with the Rosi simulated model.

• resources - General support files.

• script - Example node in Python to control Rosi using a joystick. Code written with a Xbox 360 wireless joystick.

• urdf - Contains ROSI URDF model.

• vrep_content - Contains the simulation scenes for the challenge. You may load them inside V-REP simulator.

Installation

The simulator was conceived using Ubuntu 18.4.2, ROS Melodic, and V-REP 3.6.2 (rev.0). Another software versions might work, but they are not recommended nor officially supported for the competition.

Installation advices

By covention on this installation steps, all boxes starting with a $ mark means that you should run the command on the terminal.

As an example:

$ sudo apt update

The above line means that you should run the command sudo apt update on your terminal. Notice that the $ stands for a terminal command input and do not belongs to the command line syntax itself.

If you want to directly copy and paste this README.md commands on the terminal (the clever way), remember that you should not copy the $ mark, and the keyboard shortcut to paste on the terminal is CTRL + SHIFT + V.

The common issues for each step are directly addressed in a Troubleshooting remark on the end of the step instructions.

Installation steps

Follow the steps below to configure your V-REP along with ROS:

1. Clone and download this repository package to your ROS Workspace src folder (../catkin_ws/src) folder with the name rosi_defy:

$ git clone https://github.com/filRocha/sbai2019-rosiDefy rosi_defy

2. Download V-REP PRO EDU V3.6.2 rev0 from the Coppelia Robotics website: http://www.coppeliarobotics.com/downloads.html

3. Unzip it (preferentially) to your home folder and rename the folder as vrep.

4. Add both the CATKIN_WS and V-REP folder location to your .bashrc, an alias to run it, and source the .bashrc again:

$ echo "export ROS_CATKIN_WS='<path_to_your_catkin_ws_folder>'" >> $HOME/.bashrc
$ echo "export VREP_ROOT='<path_to_your_vrep_folder>'" >> $HOME/.bashrc
$ echo "source $ROS_CATKIN_WS/devel/setup.bash" >> $HOME/.bashrc
$ echo "alias vrep=$VREP_ROOT/vrep.sh" >> ~/.bashrc
$ source $HOME/.bashrc

Remember to insert the path to your CATKIN_WS and V-REP folder in this command.

(All instructions consider that you use bash. If you use .zsh, you know what to do ;)

5. Test the V-REP functionality by running directly on your terminal the following command:

$ vrep

(Notice that you have created this command on the last step using the alias command on your .bashrc)

6. Clone recursively the V-REP/ROS interface and V-REP Velodyne2Ros plugin to your catkin_ws/src:

$ cd $ROS_CATKIN_WS/src/
$ git clone --recursive https://github.com/CoppeliaRobotics/v_repExtRosInterface.git vrep_ros_interface
$ git clone https://github.com/filRocha/vrep_plugin_velodyne.git

(More information and credits about this interface can be found on its Github repository

7. Install some support packages:

$ sudo apt install python-catkin-tools xsltproc ros-$ROS_DISTRO-brics-actuator ros-$ROS_DISTRO-tf2-sensor-msgs ros-$ROS_DISTRO-joy ros-$ROS_DISTRO-joint-state-publisher

8. We have just installed a new tool for compiling the catkin workspace: catkin build. If you use catkin_make, you have to clean your workspace and perform a fresh new compilation:

$ cd $ROS_CATKIN_WS
$ catkin clean
$ catkin build
$ source $HOME/.bashrc

Notice that you may not use catkin_make to compile your workspace anymore.

Troubleshooting:

• The catkin clean command just deletes the ..catkin_ws/devel and ..catkin_ws/build folders from your ROS Workspace. If the catkin clean command returns an error (e.g. the workspace was not correctly indentified), you can just remove those folders manually.

9. Some messages from our rosi_defy package should be referenced in the vrep_ros_interface package. To do that:

9.1 Insert their namespace and names in <vrep_ros_interface>/meta/messages.txt file:

$ echo -e "rosi_defy/ManipulatorJoints\nrosi_defy/RosiMovement\nrosi_defy/RosiMovementArray\nrosi_defy/HokuyoReading" >> $ROS_CATKIN_WS/src/vrep_ros_interface/meta/messages.txt

9.2 Tell vrep_ros_interfacethat it depends on the rosi_defy package by adding

<depend>rosi_defy</depend>

to vrep_ros_interface/package.xml.

9.3 Add rosi_defy package dependence on the vrep_ros_interface/CMakeLists.txt:

set(PKG_DEPS
  ... (many many other packages)
  rosi_defy
)

Compile again your catkin_ws using

$ catkin build

10. If your compilation runs well, there is now a ros interface and RosVelodyne libraries called libv_repExtRosInterface.so and libv_repExtRosVelodyne.so, respectively, on <catkin_ws>/dev/lib/ folder. You must copy it to the V-REP folder:

$ cp $ROS_CATKIN_WS/devel/lib/libv_repExtRosInterface.so $VREP_ROOT
$ cp $ROS_CATKIN_WS/devel/lib/libv_repExtRosVelodyne.so $VREP_ROOT

(Notice that, for further events, every time you add new custom ROS messages to the interface, you have to re-compile this library and re-copy it to $VREP_ROOT.

11. Everything should be set up for now. To run the simulation, you should first (always!) run ROS:

$ roscore
$ vrep

Open the scene <rosi_defy>/vrep_content/challenge_scenario.ttt in V-REP and play it. You should be able to see the simulator topics being published with rostopic list.

Additionally, if you have a joystick, you can run the rosi_joy.py example node to see how the communication with the robot works.

NOTICE that you have to run the roscore ALWAYS before vrep in order to work. If you stop ROS master, you have to close V-REP and run it all again.

Hello World!

To first run your simulator, do the following:

1. Open a new terminal and run $ roscore to enable ROSMASTER. You always have to run the ROS master before V-REP.

2. In another terminal window, run $ vrep. This should work if you have created successfully V-REP alias on .bashrc. If the negative case, run it directly using $ <vrep_folder>/vrep.sh.

3. Go to File > Open Scene... and locate the V-REP scenario in <rosi_defy>/vrep_content/challenge_scenario.ttt. You have also available the ROSI model that can be directly loaded in another scenario. It can be also found in the V-REP model browser (simulator's left side column).

4. Run the simulator by going to Simulation > Start simulation. At this point, you should be able to see all ROSI topics running in ROS framework.

5. You can find a ROS script example code in <rosi_defy>/script/rosi_joy.py. This node lets you control ROSI with a joystick (made with Xbox 360 controller). Besides, it is a good way to see the basics of the ROS<->ROSI interaction.

Simulation Parameters

The simulation may run slow due to its complexity and amount of running modules. You can tweak some parameters to better adjust the simulation to your code needs.

Inside the rosi_defy package, you may find the parameters in <rosi_defy>/config/simulation_parameters.yaml. They are all automatic loaded from the launch file in <rosi_defy>/launch/rosi_joy.launch. If you do not load them, all parameters will be set to true. Naturally, you can change parameters while running roscore using rosparam set <param_name> <param_desired_value>.

You can adjust the following flags:

• simulation_rendering - Boolean - Controls the rendering of V-REP visualization. Disable it for better performance.

• velodyne_processing - Boolean - If your code do not rely on Velodyne data, you can disable this sensor here and speed up the simulation.

• kinect_processing - Boolean - If your code do not rely on Kinect data, you can disable this sensor here and speed up the simulation.

• hokuyo_processing - Boolean - If your code do not rely on Hokuyo data, you can disable this sensor here and speed up the simulation.

• hokuyo_lines - Boolean - Turn this ON if you want to see hokuyo detection lines.

ROSI2ROS

This section shows how to interact with the simulated ROSI in ROS framework. It shows the published/subscribed topics, along with its message type and brief comment. The standard is:

• /topic_namespace/topic_name - <topic_message/Type> - A brief description of its content.

As a rule of thumb, all variables are mapped in the International System of Units.

ROSI publishes to (you receive information from the robot):

• /rosi/arms_joints_position - <rosi_defy/RosiMovementArray> - Rosi tracked arms position in [radians].

• /rosi/kinect_joint - <std_msgs/Float32> - Kinect joint position in [radians].

• /sensor/gps - <sensor_msgs/NavSatFix> - Emulated GPS sensor output.

• /sensor/imu - <sensor_msgs/Imu> - Emulated IMU sensor output.

• /sensor/kinect_depth - <sensor_msgs/Image> - Emulated kinect depth image output.

• /sensor/kinect_rgb - <sensor_msgs/Image> - Emulated kinect rgb image output.

• /sensor/kinect_info - <sensor_msgs/CameraInfo> - Emulated kinect information.

• /sensor/velodyne - <sensor_msgs/PointCloud2> - Emulated Velodyne output.

• /sensor/hokuyo - <rosi_defy/HokuyoReading> - Emulated hokuyo output. It gives a vector of 3D coordinates of detected point with respect to hokuyo.

• /sensor/ur5toolCam - <sensor_msgs/Image - Emulated camera on UR5 tool.

• /simulation/time - <std_msgs/Float32> - V-REP simulation time in [seconds]

• /ur5/jointsPositionCurrentState - <rosi_defy/ManipulatorJoints> - UR-5 robotic manipulator joints current position.

• /ur5/forceTorqueSensorOutput - <geometry_msgs/TwistStamped> - UR-5 Force/Torque sensor output. It gives two vector of linear and angular forces and torques, respectively. Axis order is x, y, z.

ROSI subscribes to (you send commands to the robot):

• /rosi/command_arms_speed - <rosi_defy/RosiMovementArray> - Sets the tracked arms angular velocity in [radians/s]. Command limits in [-0.52, 0.52] rad/s

• /rosi/command_traction_speed - <rosi_defy/RosiMovementArray> - Sets the traction system joint angular velocity in [radians/s]. The traction system drives simultaneously both the wheel and the tracks. Command limits in [-37.76, 37.76] rad/s.

• /rosi/command_kinect_joint - <std_msgs/Float32>- Sets the kinect joint angular position set-point. Joint limits are [-45°,45° ]. It has a built-in PID controller with maximum joint speed of |0.35| rad/s.

• /ur5/jointsPosTargetCommand - <rosi_defy/ManipulatorJoints> - Sets the UR-5 joints desired angular position. Each joint has a built-in PID controller. One may find more UR-5 info in here.

Teams' Codes

The following are the ROSI Challenge finalists' codes:

• 1st - Time IFES Vitória: https://github.com/AntMol8/Time_Ifes_Vitoria
• 2nd - Pra Valê: https://github.com/vinicius-r-silva/Pra_Vale
• 3rd - AAI Robotics: https://github.com/ara1557/AAI_robotics
• 4th - Hofs: https://github.com/Gustavo-Hofs/hofs_rosi_challenge_2019
• 5th - PPGEAS - UFSC: https://github.com/feressalem/rosi_challenge_ppgeas_ufsc
• 6th ForROS: https://github.com/raphaellmsousa/ForROS
• 7th Pyneapple: https://github.com/DiegoFr75/pyneapple
• 8th Taura Bots: https://github.com/alikolling/taura_Rosi_challenge
• 9th Titãs da Robótica: https://github.com/jeanpandolfi/TitasdaRoboticaRosiChallenge

We strongly recommend the teams' to maintain their repositories online and public. However, we are not responsible for their content and situation.

I have found something wrong!

Great! Please, open a new issue directly on this repository, or send your pull request.

If you have any suggestions, comments or compliments, you can reach the official challenge contact: [e-mail]

Doubts about the simulator MUST be treated as an issue! They will not be answered via e-mail.

Have FUN!