ROS Projects at Politecnico di Milano - course of Robotics [A.A. 2024/2025]

Authors

• Matteo Pompilio
• Piero Burigana
• Merve Rana Kizil

---

First Project

In robotics/catkin_ws/src/first_project/ you can find the implementation of odometry and sector timing for a mobile car robot using ROS, working directly with raw vehicle and GPS data, at the Monza circuit.

Project Structure

• Task 1: Odometry from Vehicle Data

  • Process messages from vehicle sensors to compute odometry.
  • Subscribes to /speedsteer topic (geometry_msgs/PointStamped):
    • y: speed (km/h)
    • x: steering wheel angle (deg)
  • Publishes odometry (nav_msgs/Odometry) to /odom and broadcasts a TF transform from odom to vehicle.

• Task 2: Odometry from GPS

  • Subscribes to /swiftnav/front/gps_pose topic (sensor_msgs/NavSatFix).
  • Converts latitude, longitude, altitude to ECEF, then to local ENU coordinates.
  • Publishes odometry (nav_msgs/Odometry) to /gps_odom and broadcasts a TF transform from odom to gps.
  • Allows manual configuration of reference point via launch file parameters (lat_r, lon_r, alt_r).

• Task 3: Sector Times and Speed

  • Subscribes to both vehicle and GPS topics.
  • Publishes sector time and average speed (custom message on /sector_times).

How to Run

1. Preparation

   • All launch and config files are in the launch/ directory.

2. Start the System

   • Use the included launch file (automatically starts all nodes and RViz visualization):

     roslaunch first_project launch.launch

3. Play the ROS Bag File

   • Run the bag file containing recorded data:

     rosbag play --clock data/project.bag

4. Visualization

   • Use RViz for visualizing odometry topics and TFs.
   • The arrows represent different odometry sources (blue: vehicle odom, with expected huge bias after the first straigth sector with respect to the red track, from GPS odom).

Requirements

• ROS Noetic
• RViz for visualization
• Bag file: project.bag

Provided Folders

• src/ – Source code for all nodes.
• launch/ – Launch files to start the system and RViz.
• cfg/ – RViz configuration files.

---

Second Project

In robotics/catkin_ws/src/second_project/ you can find the implementation of a mapping and navigation pipeline for a mobile robot using ROS.

Project Structure

• Task 1: Mapping

  • Use two 2D LiDAR scans and robot odometry from a ROS bag to reconstruct an environment map
  • Data topics: /scan_back, /scan_front, /odometry, /tf, /tf_static
  • Merge the two lasers for 360° coverage and filter out points belonging to the robot
  • At first, we used the gmapping package, which works on a Grid-based Rao-Blackwellized Particle Filter SLAM; then, we tried to improve our mapping by the slam-toolbox package, which exploits the power of Pose Graph SLAM, with slightly better results, tuning the numerous parameters by trial-and-error
  • Export the map as pgm, .png images and as .data, .posegraph and .yaml files

• Task 2: Navigation

  • Simulate the robot in Stage using the generated map.
  • Setup navigation stack to localize and plan using the static map.
  • Drive to goals loaded from a CSV file, published via a custom node using ROS actions.
  • Visualize in RViz: robot, map, TFs, particle cloud (AMCL), paths, goals (you can also manually send goals via the Navigation toolset).

How to Run

1. Mapping

   • Edit and use the launch files in launch/:

     roslaunch second_project mapping.launch

   • Play the bag file:

     rosbag play --clock data/robotics2.bag

   • Save the map using map_server and the serial formats using serialize_map from slam-toolbox:

     rosrun map_server map_saver -f ./catkin_ws/src/second_project/map/map
     rosservice call /slam_toolbox/serialize_map "{filename: '/home/robotics/catkin_ws/src/second_project/map/map'}"

   • Lastly, crop and clean the map using GIMP:

     

2. Navigation

   • Use the included launch files in launch/:

     roslaunch second_project navigation.launch

   • The map from Task 1 must be placed in the map/ folder.

   • Run with simulated time enabled.

   • Goals are read automatically from csv/goals.csv.

   • The robot will navigate to each goal sequentially:

     

3. Visualization

   • On Windows, connect to http://localhost:8080/vnc.html
   • RViz configuration files are provided in rviz/.
   • View the map, robot, sensors, and navigation status.
   • Alternatively, you can visualize everything by using Floxglove, and by running it:

   roslaunch foxglove_bridge foxglove_bridge.launch port:=8765 address:=0.0.0.0

Requirements

• ROS Noetic
• Mapping package: slam-toolbox
• Stage simulator
• Additional ROS packages: map_server, move_base, amcl
• RViz for visualization
• Bag file: robotics2.bag, in the /robotics/data folder

Provided Folders

• mapping/ – Source and launch files for Task 1.
• navigation/ – Source and launch files for Task 2.
• map/ – Output map files.
• csv/ – File with navigation goals.
• rviz/ – RViz configs.
• config/ – Configuration files for move_base and local/global planners.
• launch/ – Launch files to start the system and RViz.
• stage/ – Stage simulation files.