Based on Ubuntu 20.04 LTS (tested with WSL2)
Install ROS Noetic:
sudo apt install ros-noetic-desktop-full
Source ROS environment by adding to ~/.bashrc:
source /opt/ros/noetic/setup.bash
Install dependencies
sudo apt install python3-rosdep python3-rosinstall python3-rosinstall-generator python3-wstool build-essential
sudo apt-get install ros-noetic-tf2-geometry-msgs ros-noetic-ackermann-msgs ros-noetic-joy ros-noetic-map-server git g++
Setup rosdep
sudo rosdep init
rosdep update
Clone this repository
git clone https://github.com/sevjaeg/autonomous-racing-cars.git
Download modified f1tenth simulator (our changes fix the tf tree to work smoothly with AMCL)
cd autonomous-racing-cars
git clone https://github.com/sevjaeg/f1tenth_simulator
Create catkin workspace
mkdir -p ~/catkin_ws/src
cd ~/catkin_ws/src
Create symbolic link
ln -s <REPO_PATH>/autonomous-racing-cars/f1tenth_simulator
Build the simulator package
cd ~/catkin_ws
catkin_make
Run the setup
. ~/catkin_ws/devel/setup.bash
Start the simulator with (keyboard mode can be enabled by pressing k, then w, a, s, and d can be used)
roslaunch f1tenth_simulator simulator.launch
Create a symbolic link from the repo to your source
cd ~/catkin_ws/src
ln -s <REPO_PATH>/autonomous-racing-cars/<PATH_TO_YOUR_NODE>
And build it and don't forget to source everything with source devel/setup.bash.
To be able to debug nodes they need to be built with:
catkin_make -DCMAKE_BUILD_TYPE=DEBUG
There is a Setup Guide available online. Do not use your standard catkin_ws but e.g. cartographer_ws to avoid problems with the catkin_make command. Besides that follow all steps in the tutorial.
After compiling, the following steps are required for setup:
cd ~/catkin_ws/src
ln -s <REPO_PATH>/autonomous-racing-cars/cartographer_config/f110_description/
cp <REPO_PATH>/autonomous-racing-cars/cartographer_config/f110_2d.lua ~/cartographer_ws/install_isolated/share/cartographer_ros/configuration_files
cp <REPO_PATH>/autonomous-racing-cars/cartographer_config/f110_2d.launch ~/cartographer_ws/install_isolated/share/cartographer_ros/launch
To run the node, the following command is required
source ~/cartographer_ws/install_isolated/setup.bash
In a first terminal run the simulator (TODO or the car setup), e.g. with
roslaunch wall_follow slam.launch
Note that RVIZ might show you some errors due to missing tf transformations as the SLAM node is not running yet.
To run the cartographer, open a second terminal and run
and
roslaunch cartographer_ros f110_2d.launch
Then, the SLAM node should generate a map in the map_slam topic. For correct laser scan visualisation you might have to change the fixed frame in rviz to odom or map_slam.
To save the map, run
rosrun map_server map_saver -f <MAP_NAME> map:=/map_slam
in a third terminal.
Jinja2 version 3.2.X might cause problems
pip install jinja2==3.0.0
python pointing to python2 might cause problems
sudo apt install python-is-python3
sudo apt-get install ros-noetic-amcl
roslaunch wall_follow filter.launch
sudo apt-get install ros-noetic-robot-localization
roslaunch wall_follow fusion.launch
Requires scikit-image version 0.16.x or newer (tested with 0.19.2)
pip install scikit-image
Calculate a path with
roslaunch pure_pursuit plan_path.launch
Make sure to configure a SLAM map as such in the launch file as the image format is slighly different. The computed path is written to a bag file which is automatically read by pure pursuit.
Recorded runs from INFHS (lap6 with best quality so far)
rosbag play bags/infhs_lap6_left_auto_2022-04-07-13-38-33.bag
SLAM output: infhs_slam_2 (best map based on lap6)
Running AMCL and sensor fusion including visualisation
cd lab4
roslaunch filter_fusion_hw.launch
Run roslaunch timer timer.launch. Be sure to discard the first lap time.
- Generate map with Cartographer SLAM
convert map.pgm -fuzz 33% -fill black -opaque gray map.pgmto make hard wallsconvert map.pgm map.pngfor easier visualisation (optional)- Manually refine image with GIMP (if required)
- Update launch file (if required)
- Generate path