Digital twin and behaviour validation of the Eddie mobile robot

Final degree project aimed at creating and validating a digital twin of the Eddie mobile robot. It covers measuring the real robot, CAD modelling, generating a URDF model of its structure and kinematics, tuning its behaviour in Gazebo to match the physical system, and integrating sensors such as Lidar and Kinect for visualisation and analysis in RViz.

Robotics
Digital twin
ROS Noetic
Gazebo
RViz
URDF
SolidWorks
Mechanical design
C++
Python
Linux

Simulation of the Eddie robot in Gazebo

Final degree project focused on characterising and validating the behaviour of a mobile robot by creating its digital twin. Starting from the physical Eddie robot, the project develops a digital replica with enough fidelity to simulate its movement, adjust its behaviour and compare the model with the real system in a controlled environment.

The project was developed in several stages:

  • Measuring the real robot and creating a CAD replica in SolidWorks as the geometric foundation of the model.
  • Generating the URDF model, defining the physical structure of the robot, the relationships between its components and its kinematics within the ROS ecosystem.
  • Integration and simulation in Gazebo, tuning the model parameters to approximate the behaviour of the physical robot.
  • Incorporating Lidar and Kinect sensors into the digital twin to work on perception and navigation in simulation.
  • Visualisation and debugging in RViz, checking states, reference frames and signals relevant to validating the system as a whole.

As a complementary part of the work, a 3D-printed enclosure or mount was also designed to integrate a screen into the real robot, making it easier to display information and interact with the robot in potential practical applications.

Project resources

Physical robot

The real Eddie mobile robot