Gazebo Basics & Physics

Introduction

Gazebo is the industry-standard robot simulator. It runs physics simulations, renders 3D graphics, and integrates seamlessly with ROS 2.

This chapter teaches:

• What Gazebo is and why robots use it
• Physics engines (ODE, Bullet, DART)
• Creating simulation worlds
• Spawning robots and objects
• Physics properties (gravity, friction, collisions)
• Integrating Gazebo with ROS 2

By the end, you'll have a Gazebo world running with a robot controlled via ROS 2 commands.

---

Learning Outcomes

By the end of this chapter, you will:

1. Understand Gazebo architecture and simulation loop
2. Choose appropriate physics engines for your use case
3. Create Gazebo worlds (SDF format)
4. Spawn robots and objects into simulation
5. Configure physics (gravity, friction, collision)
6. Launch Gazebo from ROS 2 launch files
7. Debug Gazebo simulations

---

Part 1: What is Gazebo?

Gazebo Overview

Gazebo is a robot simulator that:

• Simulates physics (gravity, collisions, friction, inertia)
• Renders 3D environments
• Publishes sensor data (camera, LiDAR, IMU) to ROS 2 topics
• Receives motor commands from ROS 2 nodes
• Runs at high frequency (~1000 Hz physics, ~60 Hz graphics)

Gazebo vs. Other Simulators

Simulator	Strength	Weakness	Use Case
Gazebo	ROS 2 integration, physics	Learning curve	Research, robotics education
Isaac Sim	High fidelity, synthetic data	Requires powerful GPU	Industrial, high-quality data
CoppeliaSim	User-friendly, flexible	Smaller community	Education, prototyping
Webots	Full learning suite	Less ROS integration	Education

For this course: We use Gazebo (free, open-source, ROS 2 native).

---

Part 2: Gazebo Architecture

The Simulation Loop

Gazebo runs a physics loop at ~1000 Hz:

┌─────────────────────────────────────┐
│ 1. Apply ROS 2 commands             │
│    (motor velocities, joint forces) │
└─────────────────────────────────────┘
        ↓
┌─────────────────────────────────────┐
│ 2. Physics Step                     │
│    (Newton's laws, collisions)      │
└─────────────────────────────────────┘
        ↓
┌─────────────────────────────────────┐
│ 3. Compute Sensor Data              │
│    (camera renders, lidar casts rays)│
└─────────────────────────────────────┘
        ↓
┌─────────────────────────────────────┐
│ 4. Publish to ROS 2 Topics          │
│    (/camera/image, /scan, /tf)      │
└─────────────────────────────────────┘
        ↓
┌─────────────────────────────────────┐
│ 5. Render Graphics (60 Hz)          │
│    (display to screen)              │
└─────────────────────────────────────┘

Key Components

• Physics Engine: Computes dynamics (ODE, Bullet, DART)
• Rendering Engine: OpenGL graphics
• Sensor Simulation: Camera, LiDAR, IMU plugins
• ROS 2 Bridge: Publishes/subscribes to ROS 2 topics
• GUI: 3D visualization and controls

---

Part 3: Physics Engines

Physics Engine Choices

Gazebo supports three physics engines. Each has tradeoffs:

ODE (Open Dynamics Engine)

• Pros: Stable, good for legged robots, default
• Cons: Slower than others
• Use when: Humanoid locomotion, complex contacts

Bullet

• Pros: Fast, good for rigid bodies
• Cons: Less stable with complex constraints
• Use when: Speed matters, simple objects

DART

• Pros: Very stable, handles complex structures
• Cons: Slower, requires compilation
• Use when: Precision dynamics, grasping

Recommendation: Use ODE for this course (default, stable for humanoids).

Setting Physics in SDF

<?xml version="1.0"?>
<sdf version="1.10">
  <world name="robot_world">
    <!-- Physics engine -->
    <physics type="ode">
      <!-- Gravity (m/s²) -->
      <gravity>0 0 -9.81</gravity>

      <!-- Simulation step size (seconds) -->
      <max_step_size>0.001</max_step_size>

      <!-- Iterations per step (higher = more accurate) -->
      <real_time_update_rate>1000</real_time_update_rate>

      <!-- ODE-specific settings -->
      <ode>
        <solver type="quick">
          <!-- Constraint solver iterations -->
          <iters>50</iters>
          <!-- Pre-constraint force calculation -->
          <precon_iters>0</precon_iters>
          <!-- Use SSOR or Projected Gauss-Seidel -->
          <sor>1.3</sor>
        </solver>
      </ode>
    </physics>

    <!-- Rest of world definition -->
  </world>
</sdf>

---

Part 4: Creating Gazebo Worlds

SDF Format (Simulation Description Format)

Gazebo worlds are defined in SDF (XML-based):

<?xml version="1.0"?>
<sdf version="1.10">
  <world name="robot_world">
    <!-- Physics -->
    <physics type="ode">
      <gravity>0 0 -9.81</gravity>
      <max_step_size>0.001</max_step_size>
    </physics>

    <!-- Sun for lighting -->
    <light type="directional" name="sun">
      <cast_shadows>true</cast_shadows>
      <pose>10 10 10 0 0 0</pose>
      <diffuse>0.8 0.8 0.8 1</diffuse>
    </light>

    <!-- Ground plane -->
    <model name="ground_plane">
      <static>true</static>
      <link name="link">
        <collision name="collision">
          <geometry>
            <plane>
              <normal>0 0 1</normal>
              <size>100 100</size>
            </plane>
          </geometry>
        </collision>
        <visual name="visual">
          <geometry>
            <plane>
              <normal>0 0 1</normal>
              <size>100 100</size>
            </plane>
          </geometry>
          <material>
            <ambient>0.8 0.8 0.8 1</ambient>
            <diffuse>0.8 0.8 0.8 1</diffuse>
          </material>
        </visual>
      </link>
    </model>

    <!-- Add objects, walls, robot URDF here -->
  </world>
</sdf>

---

Part 5: Spawning Models

Method 1: Include URDF in World

<!-- Inside your world -->
<include>
  <uri>model://my_robot</uri>
  <pose>0 0 0.5 0 0 0</pose>
</include>

Method 2: Spawn via ROS 2 Service

# Convert URDF to SDF and spawn
ros2 run gazebo_ros spawn_entity.py -entity robot \
  -file /path/to/robot.urdf \
  -x 0 -y 0 -z 0.5

Method 3: Spawn via Launch File

<!-- In your launch file -->
<node pkg="gazebo_ros" exec="spawn_entity.py"
      args="-entity robot -file $(find-pkg-share my_pkg)/urdf/robot.urdf
             -x 0 -y 0 -z 0.5"
      output="screen"/>

---

Part 6: Physics Properties

Friction

Friction affects how objects slide:

<link name="wheel">
  <collision name="collision">
    <geometry>
      <sphere>
        <radius>0.05</radius>
      </sphere>
    </geometry>
    <!-- Friction properties -->
    <surface>
      <contact>
        <collide_bitmask>0xffff</collide_bitmask>
      </contact>
      <friction>
        <ode>
          <mu>0.7</mu>  <!-- Coefficient of friction -->
          <mu2>0.7</mu2>
          <fdir1>1 0 0</fdir1>  <!-- Friction direction -->
        </ode>
      </friction>
    </surface>
  </collision>
  <visual name="visual">
    <geometry>
      <sphere>
        <radius>0.05</radius>
      </sphere>
    </geometry>
    <material>
      <ambient>0 0 0 1</ambient>
      <diffuse>0 0 0 1</diffuse>
    </material>
  </visual>
</link>

Friction coefficient guide:

• 0.0: Very slippery (ice)
• 0.5: Normal surface
• 1.0+: Rough surface (rubber on concrete)

Inertia

Inertia affects how objects respond to forces:

<link name="body">
  <inertial>
    <mass>5.0</mass>  <!-- kg -->
    <!-- Inertia matrix (usually auto-computed) -->
    <inertia>
      <ixx>0.1</ixx>
      <iyy>0.1</iyy>
      <izz>0.1</izz>
      <ixy>0.0</ixy>
      <ixz>0.0</ixz>
      <iyz>0.0</iyz>
    </inertia>
  </inertial>
</link>

---

Part 7: Running Gazebo with ROS 2

Basic Launch

<?xml version="1.0"?>
<launch>
  <!-- Start Gazebo server (headless) -->
  <include file="$(find-pkg-share gazebo_ros)/launch/gazebo.launch.py">
    <arg name="world" value="$(find-pkg-share my_robot)/worlds/robot_world.sdf"/>
  </include>

  <!-- Start RViz for visualization (optional) -->
  <node pkg="rviz2" exec="rviz2" name="rviz2" output="screen">
    <param name="rviz_config" value="$(find-pkg-share my_robot)/config/rviz.rviz"/>
  </node>

  <!-- Spawn robot in Gazebo -->
  <node pkg="gazebo_ros" exec="spawn_entity.py"
        args="-entity robot -file $(find-pkg-share my_robot)/urdf/robot.urdf
               -x 0 -y 0 -z 0.5"
        output="screen"/>
</launch>

Run it:

ros2 launch my_robot robot.launch.xml

---

Part 8: Troubleshooting

Robot Stability Issues

Problem	Cause	Solution
Robot oscillates	Physics step too large	Reduce max_step_size
Robot falls through ground	Bad collision shapes	Check URDF collision geometry
Robot too slow	Gravity wrong	Verify <gravity>0 0 -9.81</gravity>
Joint wobbles	Low inertia	Increase mass or adjust inertia

Performance Issues

Gazebo running slowly?:

1. Reduce physics update rate (lower accuracy, faster)
2. Lower ODE solver iterations
3. Reduce sensor update rates
4. Close GUI, run headless

Command:

# Run Gazebo without GUI (faster)
GAZEBO_GUI_PLUGIN_PATH="" gazebo --headless my_world.sdf

---

Summary

Gazebo Basics:

• Physics loop runs at ~1000 Hz
• ODE engine is default and stable
• Worlds defined in SDF (XML)
• Models spawned via URDF or SDF
• Integrates with ROS 2 via plugins

Physics Configuration:

• Gravity, step size, solver iterations
• Friction (mu coefficient)
• Inertia (mass, moment of inertia)
• Contact properties

ROS 2 Integration:

• Launch files start Gazebo
• Spawn entities via service
• Sensor plugins publish to ROS 2 topics
• Motor commands received from ROS 2

---

Next Steps

Next chapter: URDF & Robot Descriptions - defining your robot's structure for Gazebo.

---

Navigation

• Previous: Module 2 Intro
• Next: Chapter 2: URDF & Robot Descriptions
• Lab: Lab 2.1: Load Robot in Gazebo