ROS robotics projects

Robotics Project A.Y. 2021-2022
Professor: Prof. Matteo Matteucci
Supervisors: Paolo Cudrano and Simone Mentasti

Team

• Riccardo Izzo (@RiccardoIzzo)

• Valentina Abbattista (@valeabba)

• Ossama El Oukili (@ossaleon)

Robot characteristics

Robot image	Robot scheme

• Omnidirectional robot
• 4 mecanum wheels with rollers at 45°
• Encoders on each wheel (RPM or ticks)
• Geometric parameters: wheel radius (r), wheel position along x (l) and wheel position along y (w)
• Data:
  • r = 0.07 m
  • l = 0.200 m
  • w = 0.169 m
  • Gear ratio (T) = 5:1
  • Encoders resolution (N) = 42 CPR (Counts Per Rev.)

First project - Odometry with ROS

Goals:

• Compute odometry using appropriate kinematics ✔️
  • Compute robot linear and angular velocities v,w from wheel encoders
  • Compute odometry using both Euler and Runge-Kutta integration
  • Calibrate robot parameters to match ground truth
• Compute wheel control speeds from v,w ✔️
• Add a service to reset the odometry to a specific pose (x, y, ϑ) ✔️
• Use dynamic reconfigure to select between integration method ✔️

Fig.1 - Odometry comparison with rviz (green: reference odometry, red: our approximation)

Fig.2 - Odometry comparison with rqt_plot (front right wheel in the example)

Second project - Mapping and Localization with ROS

Goals:

• Create the map ✔️
• Perform amcl based localization ✔️
• Write a service to save an image with the map and the trajectory of the robot ✔️

Fig.3 - Map creation with gmapping

Sensors

• OptiTrack Markers
• YDLIDAR G4

Development environment and Tools

• Ubuntu 18.04
• ROS Melodic 1.0
• rviz
• gmapping
• amcl
• OpenCV 3.2.0