ROS2 Digital Twin of Agro Rover

A high-fidelity simulation of the agro rover described by a URDF model, running the same ROS 2 navigation and perception stack as the physical robot. Gazebo provides the physics and sensor simulation (LiDAR, camera, GPS), while RViz visualizes state and planning. The twin enables safe, repeatable testing of new behaviours and regression-checking changes before field deployment.

Testing autonomy changes directly on field hardware is slow, risky, and hard to reproduce. A digital twin is needed so navigation, perception, and control can be iterated and regression-tested safely before deployment.

The rover is modelled in URDF and simulated in Gazebo with physics and sensor models (LiDAR, camera, GPS) that mirror the real platform. The same ROS 2 navigation and perception stack runs against the twin, with RViz for visualization, enabling repeatable scenario testing and sim-to-real validation.

Planned — targets: a twin faithful enough to catch regressions before hardware testing, a quantified sim-to-real gap, and a library of reusable test scenarios.

1. Phase 1

   URDF model

   Describe the rover's links, joints, and sensors in URDF.

2. Phase 2

   Gazebo world

   Build a Gazebo environment with simulated LiDAR, camera, and GPS.

3. Phase 3

   Stack integration

   Run the real ROS 2 navigation/perception stack against the twin with RViz.

4. Phase 4

   Sim-to-real

   Validate behaviours in sim and quantify the gap against the physical rover.