Skip to main content

Module 2: Robot Simulation (Weeks 6-7)

📚 Learning Outcomes​

By the end of this module, you will be able to:

  1. Create realistic simulation environments in Gazebo with physics, sensors, and lighting
  2. Model robots using URDF (Unified Robot Description Format) with accurate kinematics and dynamics
  3. Integrate simulated sensors (cameras, LiDAR, IMU) with ROS 2 perception pipelines
  4. Configure physics engines for accurate contact dynamics and collision detection
  5. Bridge Gazebo simulations with Unity for photorealistic visualization
  6. Understand the sim-to-real gap and techniques to minimize it

🎯 Module Overview​

Duration: 2 weeks (14 days) Prerequisites: ROS 2 Fundamentals (Module 1), basic Python, URDF/XML syntax

Simulation is critical for robotics development:

  • Safety: Test dangerous scenarios (stair falls, collisions) without risking hardware
  • Speed: Iterate 10-100x faster than real-world testing
  • Scale: Test thousands of scenarios in parallel
  • Cost: Zero cost for robot damage or sensor replacement

This module teaches you to build production-quality simulations that transfer to real robots.

Weekly Breakdown​

Week 6: Gazebo Fundamentals​

Topics:

  • Gazebo architecture (client-server, plugins)
  • Creating worlds (terrain, lighting, objects)
  • URDF robot modeling (links, joints, sensors)
  • SDF (Simulation Description Format) vs. URDF
  • Physics engines: ODE, Bullet, DART

Hands-On:

  • Build a warehouse environment
  • Model a differential drive mobile robot
  • Add RGB camera and 2D LiDAR sensors
  • Configure physics parameters (gravity, friction, contact)

Assessment: Complete Gazebo world with mobile robot navigating obstacles

Week 7: Advanced Simulation & Sim-to-Real​

Topics:

  • Sensor noise models (Gaussian, ray noise)
  • Unity integration for photorealistic rendering
  • Gazebo plugins for custom behavior
  • Domain randomization for robust policies
  • Sim-to-real transfer techniques

Hands-On:

  • Add sensor noise to camera and LiDAR
  • Create custom Gazebo plugin
  • Integrate Unity visualization
  • Test navigation policy in randomized environments

Assessment: Mobile manipulator simulation with realistic sensor noise and Unity visualization

Code Example: URDF Mobile Robot​

<?xml version="1.0"?>
<robot name="simple_robot">
  <!-- Base Link -->
  <link name="base_link">
    <visual>
      <geometry>
        <box size="0.6 0.4 0.2"/>
      </geometry>
      <material name="blue">
        <color rgba="0 0 0.8 1"/>
      </material>
    </visual>
    <collision>
      <geometry>
        <box size="0.6 0.4 0.2"/>
      </geometry>
    </collision>
    <inertial>
      <mass value="10.0"/>
      <inertia ixx="0.133" ixy="0" ixz="0" iyy="0.333" iyz="0" izz="0.433"/>
    </inertial>
  </link>

  <!-- Left Wheel -->
  <link name="left_wheel">
    <visual>
      <geometry>
        <cylinder radius="0.1" length="0.05"/>
      </geometry>
      <material name="black">
        <color rgba="0 0 0 1"/>
      </material>
    </visual>
    <collision>
      <geometry>
        <cylinder radius="0.1" length="0.05"/>
      </geometry>
    </collision>
    <inertial>
      <mass value="1.0"/>
      <inertia ixx="0.0026" ixy="0" ixz="0" iyy="0.0026" iyz="0" izz="0.005"/>
    </inertial>
  </link>

  <joint name="left_wheel_joint" type="continuous">
    <parent link="base_link"/>
    <child link="left_wheel"/>
    <origin xyz="-0.2 0.225 0" rpy="-1.5707 0 0"/>
    <axis xyz="0 0 1"/>
  </joint>

  <!-- Camera Sensor -->
  <link name="camera_link">
    <visual>
      <geometry>
        <box size="0.05 0.05 0.05"/>
      </geometry>
    </visual>
  </link>

  <joint name="camera_joint" type="fixed">
    <parent link="base_link"/>
    <child link="camera_link"/>
    <origin xyz="0.3 0 0.15" rpy="0 0 0"/>
  </joint>

  <gazebo reference="camera_link">
    <sensor name="camera" type="camera">
      <update_rate>30</update_rate>
      <camera>
        <horizontal_fov>1.39626</horizontal_fov>
        <image>
          <width>640</width>
          <height>480</height>
        </image>
        <clip>
          <near>0.1</near>
          <far>100</far>
        </clip>
        <noise>
          <type>gaussian</type>
          <mean>0.0</mean>
          <stddev>0.007</stddev>
        </noise>
      </camera>
      <plugin name="camera_controller" filename="libgazebo_ros_camera.so">
        <ros>
          <namespace>/robot</namespace>
          <remapping>image_raw:=camera/image_raw</remapping>
        </ros>
      </plugin>
    </sensor>
  </gazebo>
</robot>

Launch in Gazebo:

ros2 launch my_robot_description gazebo.launch.py

Key Takeaways​

✅ Gazebo is production-ready for testing complex behaviors ✅ URDF defines robot structure—links (rigid bodies) + joints (connections) ✅ Sensors in simulation != reality—always add noise models ✅ Physics tuning matters—ODE for speed, Bullet for accuracy ✅ Unity for visuals—when you need photorealism or human interaction testing

What's Next?​

👉 Module 3: NVIDIA Isaac Platform — GPU-accelerated simulation and RL training