This is a ROS workspace made for autonomous navigation and teleoperation of the leo rover, for the ERC 2021.
Keywords: ROS, ERC_2021, python
The source code is released for use at Robocol.
**Author: motion control and power subsystems
Affiliation: Robocol
Maintainer: Motion control team at Robocol, robocol@uniandes.edu.co
The motion_control_pkg package has been tested under ROS Melodic on Ubuntu 18.04. For optimal performance, running a partition with Ubuntu on your machine is recommended, however, a virtual machine is still a good option. This is development code, expect continous changes and improvements.
The package has been made to work with the ERC-Remote-Navigation-Sim package. In order for it to work, you must check their repository here and follow their instructions for building and launching the marsyard simulation. We recomend you build this workspace at ~ in your machine.
- python3 (you probably already have this with your ubuntu and ROS installation)
- Robot Operating System (ROS) (middleware for robotics)
- use pip3 install [module] or python3 -m pip install [module] to install the following modules
pip3 install opencv-python matplotlib rospkg pynput*Troubleshooting Pip install opencv-python fails with ModuleNotFoundError: No module named 'skbuild': your pip version must be >= 19.3. Please upgrade pip with pip install --upgrade pip.
- Clone the repository (recommended location is ~).
cd ~git clone https://github.com/robocol-rem-u/motion_control.git- Move to the root of the workspace.
cd ~/motion_control- Build the workspace
catkin build
source devel/setup.bash- In a new tab (Ctrl+Shift+t) run roscore
roscore- Run the desired node in your previous tab.
rosrun motion_control_pkg planeacion_a.pyDocker is a great way to run an application with all dependencies and libraries bundles together. Make sure to install Docker first.
NOTE If you are using docker on linux, make sure to run the docker commands with sudo. Or, you can configure docker to work without root permisions here.
It is NOT recommended to build the image from windows, since issues can arise due to file format incompatibility. Working on a *nix system (linux, macos) is recommended.
cd to your working repository, then run docker build to create an image named erc.
cd ~/motion_control
docker build -t erc .spin up a container named motion-container from the image
docker run -it --name motion-container ercto start a new terminal inside the container, on a new terminal on your host run:
docker exec -it motion-container bashNow let's try to run a node.
source devel/setup.bash
roscoreOn a new terminal (inside the container)
source devel/setup.bash
rosrun motion_control_pkg planeacion_a.py If you wish, you can echo the route topic:
source devel/setup.bash
rostopic echo /Robocol/MotionControl/ruta Finally, publish to the topic
rostopic pub -1 /Robocol/Inicio_fin geometry_msgs/PoseArray "header:
seq: 0
stamp:
secs: 0
nsecs: 0
frame_id: ''
poses:
- position:
x: 0.0
y: 0.0
z: 0.0
orientation:
x: 0.0
y: 0.0
z: 0.0
w: 0.0
- position:
x: 5.0
y: 0.0
z: 0.0
orientation:
x: 0.0
y: 0.0
z: 0.0
w: 0.0"You should see an output on the node and the route topic.
In order to create a path given two points in a map, send the coordinates to the control node and have the rover move autonomously towards a goal, follow these steps:
Begin by going to your ERC working directory and launching the marsyard simulation
cd ~/ERC-Remote-Navigation-Sim/
source devel/setup.bash
roslaunch leo_erc_gazebo leo_marsyard.launchA new gazebo window should appear, and the ROS master should now be running. Now lets run the necessary nodes from the motion control package:
cd ~/motion_control/
source devel/setup.bashrun the following nodes. Remember to run each one on its own terminal and to source the workspace each time.
rosrun motion_control_pkg planeacion_a.pyrosrun motion_control_pkg teclado.py rosrun motion_control_pkg control_simV2.pyoptionaly, you can echo the /Robocol/Inicio_fin topic, to make sure it's working correctly
rostopic echo /Robocol/Inicio_finNow that all the required nodes are running, let's publish a message to create a path:
rostopic pub -1 /Robocol/Inicio_fin geometry_msgs/PoseArray "header:
seq: 0
stamp:
secs: 0
nsecs: 0
frame_id: ''
poses:
- position:
x: 0.0
y: 0.0
z: 0.0
orientation:
x: 0.0
y: 0.0
z: 0.0
w: 0.0
- position:
x: 5.0
y: 0.0
z: 0.0
orientation:
x: 0.0
y: 0.0
z: 0.0
w: 0.0"Chech the echo terminal, it should display the same message you just published. Then check the terminal running the planeacion_a.py node, it should display an array of points, this is the path we are going to use. Finally, check the terminal running control_simV2.py, it shoud display an updated array of coordinates, the same as before.
Now, head to the terminal running teclado.py and press k twice, until it displays
modo: Autonomo
If you followed the instructions correctly, leo rover should now be moving on its own! See the terminal running control_simV2.py to see leo's pose, error and goal.
Congratulations! You have succesfully made leo navigate autonomously.
runs the three main nodes required for autonomous navigation
cd ~/motion_control
source devel/setup.bash
roslaunch motion_control_pkg autonomous_nav.launchContains the A* algorithm to create a path given two poses.
/Robocol/Inicio_fin([PoseArray])
Robocol/MotionControl/ruta([numpy_msg(Floats)])
Controls the robot with the keyboard. Press k to toggle autonomous navigation. Run the control node (control_simV2) after this one. MANEJA EL ROBOT CON EL TECLADO U OPRIMIENDO K CAMBIA AL CONTROL AUTOMATICO. DESPUES DE CORRER ESTE CORRER EL NODO DE control_sim.
/cmd_vel([Twist])Robocol/MotionControl/flag_autonomo([Bool])
Recieves path and moves the robot. RECIBE LAS INDICACIONES PARA HACER EL RECORRIDO Y MUEVE AL ROBOT - VERSION 4.
cmd_vel([Twist])
zed2/odom([Odometry])Robocol/MotionControl/flag_autonomo([Bool])Robocol/MotionControl/ruta([numpy_msg(Floats)])
Please report bugs and request features with the motion control team at the traction subsystem.