/feup-ri-city-slam

Duckietown SLAM project for Intelligent Robotics, FEUP.

Primary LanguagePythonGNU General Public License v3.0GPL-3.0

feup-ri-city-slam

SLAM project with Duckietown for Intelligent Robotics (M.EIC041), FEUP.

Directory Structure

  • catkin_workspace: ROS workspace for the ORB SLAM 3 ROS package
    • src: ROS packages
      • duckie_bot_pkg: ROS package for a translator from CompressedImage to Image
      • duckie_sim_pkg: ROS package for the DuckieTown Gym Simulator, with a script to publish images from the simulator to a ROS topic
      • orb_slam3_ros: ROS package for ORB-SLAM 3
  • assets: Assets used in the project
  • lib: Libraries used in the project
    • gym-duckietown: DuckieTown Gym Simulator

Prerequisites for running ORB SLAM 3 locally with the DuckieTown Gym Simulator

  • python==3.8
  • Ubuntu==20.04.6
  • Running xhost + local:docker on host machine (for GUI support)

Install simulator dependencies and setup:

From the root directory of the repo:

./setup_sim.sh

Install ROS Noetic

Follow the instructions in the ROS Noetic Website. Make sure to also install dependencies for package building.

Install catkin

sudo apt install -y python3-catkin-tools

Installing ORB SLAM 3 Dependencies

As in the orb_slam3_ros repository:

  1. Install dependencies:

    sudo apt install -y libeigen3-dev libopencv-dev ros-noetic-rviz

  2. Install Pangolin:

    cd ~
git clone --recursive https://github.com/stevenlovegrove/Pangolin.git
cd Pangolin

# Next command requires you to confirm the installation of dependencies
sudo ./scripts/install_prerequisites.sh recommended

mkdir build && cd build
cmake ..
make
sudo make install

  3. Verify OpenCV is installed (required at least 3.0):

    pip3 install opencv-python
python3 -c "import cv2; print(cv2.__version__)"

  4. (Optional) Hector trajectory server

    sudo apt install -y ros-[DISTRO]-hector-trajectory-server

Running ORB SLAM 3 with DuckieTown Gym Simulator

Running without docker

From the root directory of the repo, build the ROS workspace:

source /opt/ros/noetic/setup.bash
cd catkin_workspace
rosdep install --from-paths src --ignore-src -r -y
catkin build # This will most likely fail the first time, if it does, run the next line
# catkin build -j1

In a terminal:

source /opt/ros/noetic/setup.bash
cd catkin_workspace
source devel/setup.bash
roslaunch orb_slam3_ros duckie_sim.launch

In another terminal:

source /opt/ros/noetic/setup.bash
cd catkin_workspace
source devel/setup.bash
rosrun duckie_sim_pkg ros_script.py

Running with docker

docker buildx build -f Dockerfile.orbslam3 -t orbslam3 .
docker buildx build -f Dockerfile.duckiesim -t duckiesim .
docker run --rm -t --net=host ros:noetic roscore
# Allow docker to access the host's display
xhost + local:docker

docker run --gpus all --rm -it --ipc=host --net=host --privileged \
  --env="DISPLAY=unix$DISPLAY" \
  -v /tmp/.X11-unix:/tmp/.X11-unix:rw \
  --runtime=nvidia \
  duckiesim
# Allow docker to access the host's display
xhost + local:docker

# Run the ORB SLAM 3 container
docker run --gpus all --rm -it --ipc=host --net=host --privileged \
  --env="DISPLAY=unix$DISPLAY" \
  -v /tmp/.X11-unix:/tmp/.X11-unix:rw \
  --runtime=nvidia \
  orbslam3

Run in Real Life with duckiebot

docker buildx build -f Dockerfile.orbslam3 -t orbslam3 .
docker buildx build -f Dockerfile.duckiebot-translator -t duckiebot-translator .
docker run --rm --net=host \
    -e ROS_MASTER_URI=http://pato.local:11311 \
    duckiebot-translator
# Allow docker to access the host's display
xhost + local:docker

# Run the ORB SLAM 3 container
docker run --gpus all --rm -it --ipc=host --net=host --privileged \
  --env="DISPLAY=unix$DISPLAY" \
  -v /tmp/.X11-unix:/tmp/.X11-unix:rw \
  --runtime=nvidia \
  -e ROS_MASTER_URI=http://pato.local:11311 \
  orbslam3
# Start the line following demo
dts duckiebot demo --demo_name lane_following --duckiebot_name pato

# Start the keyboard control demo with the line following control
dts dts duckiebot keyboard_control pato

Authors