NUS ME5413 Autonomous Mobile Robotics Final Project
- System Requirements:
- Ubuntu 20.04 (18.04 not yet tested)
- ROS Noetic (Melodic not yet tested)
- C++11 and above
- CMake: 3.0.2 and above
- This repo depends on the following standard ROS pkgs:
roscpp
rospy
rviz
std_msgs
nav_msgs
geometry_msgs
visualization_msgs
tf2
tf2_ros
tf2_geometry_msgs
pluginlib
map_server
gazebo_ros
jsk_rviz_plugins
jackal_gazebo
jackal_navigation
velodyne_simulator
teleop_twist_keyboard
- And this gazebo_model repositiory
This repo is a ros workspace, containing three rospkgs:
interactive_tools
are customized tools to interact with gazebo and your robotjackal_description
contains the modified jackal robot model descriptionsme5413_world
the main pkg containing the gazebo world, and the launch files
Note: If you are working on this project, it is encouraged to fork this repository and work on your own fork!
After forking this repo to your own github:
# Clone your own fork of this repo (assuming home here `~/`)
cd
git clone https://github.com/<YOUR_GITHUB_USERNAME>/ME5413_Final_Project.git
cd ME5413_Final_Project
# Install all dependencies
rosdep install --from-paths src --ignore-src -r -y
# Build
catkin_make
# Source
source devel/setup.bash
To properly load the gazebo world, you will need to have the necessary model files in the ~/.gazebo/models/
directory.
There are two sources of models needed:
-
# Create the destination directory cd mkdir -p .gazebo/models # Clone the official gazebo models repo (assuming home here `~/`) git clone https://github.com/osrf/gazebo_models.git # Copy the models into the `~/.gazebo/models` directory cp -r ~/gazebo_models/* ~/.gazebo/models
-
# Copy the customized models into the `~/.gazebo/models` directory cp -r ~/ME5413_Final_Project/src/me5413_world/models/* ~/.gazebo/models
This command will launch the gazebo with the project world
# Launch Gazebo World together with our robot
roslaunch me5413_world world.launch
If you wish to explore the gazebo world a bit, we provide you a way to manually control the robot around:
# Only launch the robot keyboard teleop control
roslaunch me5413_world manual.launch
Note: This robot keyboard teleop control is also included in all other launch files, so you don't need to launch this when you do mapping or navigation.
After launching Step 0, in the second terminal:
# Launch GMapping
roslaunch me5413_world mapping.launch
After finishing mapping, run the following command in the thrid terminal to save the map:
# Save the map as `my_map` in the `maps/` folder
roscd me5413_world/maps/
rosrun map_server map_saver -f my_map map:=/map
Once completed Step 2 mapping and saved your map, quit the mapping process.
Then, in the second terminal:
# Load a map and launch AMCL localizer
roslaunch me5413_world navigation.launch
- You may use any SLAM algorithm you like, any type:
- 2D LiDAR
- 3D LiDAR
- Vision
- Multi-sensor
- Verify your SLAM accuracy by comparing your odometry with the published
/gazebo/ground_truth/state
topic (nav_msgs::Odometry
), which contains the gournd truth odometry of the robot. - You may want to use tools like EVO to quantitatively evaluate the performance of your SLAM algorithm.
-
We have provided you a GUI in RVIZ that allows you to click and generate/clear the random objects in the gazebo world:
-
From the starting point, move to one of the four given destination boxes at the end of the map:
- Count the number of occurance of each type of box (e.g. box 1, 2, 3, 4, the box numbers are randomly generated)
- Cross the bridge (the location of the bridge is randomly generated)
- Unlock the blockade on the bridge by publishing a
true
message (std_msgs/Bool
) to the/cmd_open_bridge
topic - Dock at the destination box with the least number of occurance
You are welcome contributing to this repo by opening a pull-request
We are following:
The ME5413_Final_Project is released under the MIT License