Skip to main content

Module 2: Digital Twin Simulation

Overview​

In Module 2 (Weeks 6-7), you'll learn to create digital twins of robotic systems using Gazebo and Unity ML-Agents, enabling safe testing and validation before physical deployment.

Duration: 2 weeks (20 hours estimated) Weeks: 6-7 Prerequisites: Module 1 (ROS 2 Fundamentals), basic 3D modeling concepts

Learning Objectives​

  1. Set Up Simulation Environments - Configure Gazebo for robot simulation with realistic physics
  2. Create Robot Models - Build and import URDF/SDF models with sensors and actuators
  3. Integrate Unity ML-Agents - Use reinforcement learning for robot behavior training
  4. Test in Simulation - Validate navigation and manipulation before hardware deployment

Module Content​

Week 6: Gazebo Simulation​

  • Gazebo architecture and installation
  • Creating simulation worlds
  • Importing robot models (URDF/SDF)
  • Sensor simulation (camera, lidar, IMU)

Week 7: Unity ML-Agents​

  • Unity ML-Agents setup
  • Environment design for reinforcement learning
  • Training navigation policies
  • Sim-to-real transfer techniques

Capstone Integration​

Digital twin simulation enables safe testing of your humanoid robot:

  • Test navigation algorithms without hardware
  • Validate manipulation strategies in simulation
  • Train reinforcement learning policies safely
  • Debug complex behaviors before deployment

Assessment​

Gazebo Simulation Project (Week 7)

  • Robot model with sensors in Gazebo
  • Simulated navigation task
  • ROS 2 integration

Continue to Week 6: Gazebo Basics →