Environment Creation

Introduction

Realistic simulation environments are essential for testing robots in scenarios that mirror real-world conditions. This chapter covers creating custom worlds, adding obstacles, terrain, and environmental conditions in Gazebo.

Learning Objectives:

• Create custom Gazebo world files (SDF format)
• Add models (buildings, furniture, obstacles)
• Design terrain and ground planes
• Configure lighting and physics parameters
• Build test scenarios for specific robot tasks

Theory

Why Environment Design Matters

Testing Diversity:

• Indoor environments (warehouses, homes, offices)
• Outdoor environments (streets, parks, rough terrain)
• Task-specific scenarios (assembly lines, retail stores)

Algorithm Validation:

• Navigation algorithms need obstacles and narrow passages
• Manipulation requires cluttered tabletops
• Perception algorithms need varied lighting and textures

Sim-to-Real Transfer:

• Realistic environments improve transferability
• Domain randomization reduces reality gap
• Edge case coverage

Gazebo World Structure (SDF)

SDF (Simulation Description Format) defines complete simulation environments:

<?xml version="1.0"?>
<sdf version="1.7">
  <world name="my_world">
    <!-- Physics engine configuration -->
    <physics type="ode">
      <max_step_size>0.001</max_step_size>
      <real_time_factor>1.0</real_time_factor>
    </physics>

    <!-- Lighting -->
    <light type="directional" name="sun">
      <pose>0 0 10 0 0 0</pose>
      <diffuse>1 1 1 1</diffuse>
      <specular>0.5 0.5 0.5 1</specular>
    </light>

    <!-- Ground plane -->
    <include>
      <uri>model://ground_plane</uri>
    </include>

    <!-- Custom models -->
    <!-- Add buildings, obstacles, etc. -->
  </world>
</sdf>

Creating a Basic World

Step 1: Create World File

Create warehouse.world:

<?xml version="1.0"?>
<sdf version="1.7">
  <world name="warehouse">
    <!-- Physics -->
    <physics type="ode">
      <max_step_size>0.001</max_step_size>
      <real_time_factor>1.0</real_time_factor>
      <real_time_update_rate>1000.0</real_time_update_rate>
    </physics>

    <!-- Scene -->
    <scene>
      <ambient>0.4 0.4 0.4 1</ambient>
      <background>0.7 0.7 0.7 1</background>
      <shadows>true</shadows>
    </scene>

    <!-- Sun -->
    <light type="directional" name="sun">
      <cast_shadows>true</cast_shadows>
      <pose>0 0 10 0 0 0</pose>
      <diffuse>0.8 0.8 0.8 1</diffuse>
      <specular>0.2 0.2 0.2 1</specular>
      <attenuation>
        <range>1000</range>
        <constant>0.9</constant>
        <linear>0.01</linear>
        <quadratic>0.001</quadratic>
      </attenuation>
      <direction>-0.5 0.1 -0.9</direction>
    </light>

    <!-- Ground Plane -->
    <include>
      <uri>model://ground_plane</uri>
    </include>

    <!-- Warehouse Floor -->
    <model name="warehouse_floor">
      <static>true</static>
      <link name="floor">
        <visual name="visual">
          <geometry>
            <plane>
              <normal>0 0 1</normal>
              <size>50 50</size>
            </plane>
          </geometry>
          <material>
            <ambient>0.5 0.5 0.5 1</ambient>
            <diffuse>0.5 0.5 0.5 1</diffuse>
          </material>
        </visual>
        <collision name="collision">
          <geometry>
            <plane>
              <normal>0 0 1</normal>
              <size>50 50</size>
            </plane>
          </geometry>
        </collision>
      </link>
    </model>
  </world>
</sdf>

Step 2: Launch World in Gazebo

# Launch Gazebo with custom world
gazebo warehouse.world

Or via ROS 2 launch file:

# warehouse_world.launch.py
from launch import LaunchDescription
from launch.actions import IncludeLaunchDescription
from launch.launch_description_sources import PythonLaunchDescriptionSource
import os
from ament_index_python.packages import get_package_share_directory

def generate_launch_description():
    world_file = os.path.join(
        get_package_share_directory('my_simulation'),
        'worlds',
        'warehouse.world'
    )

    gazebo = IncludeLaunchDescription(
        PythonLaunchDescriptionSource([
            os.path.join(get_package_share_directory('gazebo_ros'),
                         'launch', 'gazebo.launch.py')
        ]),
        launch_arguments={'world': world_file}.items()
    )

    return LaunchDescription([gazebo])

Adding Models to the Environment

Method 1: Using Gazebo Model Database

Gazebo provides pre-built models:

<!-- Add a construction cone -->
<include>
  <uri>model://construction_cone</uri>
  <pose>2 3 0 0 0 0</pose>
</include>

<!-- Add a table -->
<include>
  <uri>model://table</uri>
  <pose>5 0 0 0 0 1.57</pose>  <!-- x y z roll pitch yaw -->
</include>

<!-- Add multiple boxes -->
<include>
  <uri>model://cardboard_box</uri>
  <name>box_1</name>
  <pose>-2 -2 0.5 0 0 0</pose>
</include>

<include>
  <uri>model://cardboard_box</uri>
  <name>box_2</name>
  <pose>-2 2 0.5 0 0 0</pose>
</include>

Available Models:

• Browse at: https://github.com/osrf/gazebo_models
• Or use Gazebo GUI: Insert tab → Select model

Method 2: Creating Custom Models

Create a custom obstacle:

<model name="custom_wall">
  <static>true</static>
  <link name="wall_link">
    <pose>0 0 1 0 0 0</pose>
    <collision name="collision">
      <geometry>
        <box>
          <size>5 0.2 2</size>  <!-- length width height -->
        </box>
      </geometry>
    </collision>
    <visual name="visual">
      <geometry>
        <box>
          <size>5 0.2 2</size>
        </box>
      </geometry>
      <material>
        <ambient>0.8 0.2 0.2 1</ambient>
        <diffuse>0.8 0.2 0.2 1</diffuse>
      </material>
    </visual>
  </link>
</model>

Terrain Generation

Heightmap Terrain

For outdoor robots, create terrain from heightmap images:

<model name="terrain">
  <static>true</static>
  <link name="link">
    <collision name="collision">
      <geometry>
        <heightmap>
          <uri>file://terrain.png</uri>
          <size>100 100 10</size>  <!-- x y max_height -->
          <pos>0 0 0</pos>
        </heightmap>
      </geometry>
    </collision>
    <visual name="visual">
      <geometry>
        <heightmap>
          <uri>file://terrain.png</uri>
          <size>100 100 10</size>
          <texture>
            <diffuse>file://grass.jpg</diffuse>
            <normal>file://grass_normal.jpg</normal>
            <size>10</size>
          </texture>
        </heightmap>
      </geometry>
    </visual>
  </link>
</model>

Heightmap Image:

• Grayscale PNG/JPEG
• White = high elevation, Black = low elevation
• Recommended: 512x512 or 1024x1024 pixels

Uneven Floors

<model name="uneven_floor">
  <static>true</static>
  <link name="floor">
    <collision name="collision">
      <geometry>
        <mesh>
          <uri>file://uneven_terrain.dae</uri>
        </mesh>
      </geometry>
    </collision>
    <visual name="visual">
      <geometry>
        <mesh>
          <uri>file://uneven_terrain.dae</uri>
        </mesh>
      </geometry>
    </visual>
  </link>
</model>

Environment Lighting

Multiple Light Sources

<!-- Overhead lights (typical warehouse) -->
<light type="point" name="ceiling_light_1">
  <pose>10 10 5 0 0 0</pose>
  <diffuse>1 1 1 1</diffuse>
  <specular>0.1 0.1 0.1 1</specular>
  <attenuation>
    <range>20</range>
    <constant>0.5</constant>
    <linear>0.01</linear>
  </attenuation>
</light>

<light type="point" name="ceiling_light_2">
  <pose>-10 10 5 0 0 0</pose>
  <diffuse>1 1 1 1</diffuse>
  <attenuation>
    <range>20</range>
  </attenuation>
</light>

<!-- Spotlights -->
<light type="spot" name="spotlight">
  <pose>0 0 10 0 0 0</pose>
  <diffuse>1 1 0.8 1</diffuse>
  <direction>0 0 -1</direction>
  <attenuation>
    <range>50</range>
  </attenuation>
  <spot>
    <inner_angle>0.6</inner_angle>
    <outer_angle>1.0</outer_angle>
    <falloff>1.0</falloff>
  </spot>
</light>

Dynamic Lighting (Day/Night)

Use Gazebo plugins for time-of-day simulation:

<plugin name="sun_position" filename="libgazebo_ros_sun.so">
  <ros>
    <namespace>/environment</namespace>
  </ros>
  <update_rate>0.1</update_rate>
</plugin>

Physics Configuration

Tuning Physics Parameters

<physics type="ode">
  <max_step_size>0.001</max_step_size>  <!-- 1ms timestep -->
  <real_time_factor>1.0</real_time_factor>
  <real_time_update_rate>1000.0</real_time_update_rate>

  <!-- Solver configuration -->
  <ode>
    <solver>
      <type>quick</type>
      <iters>50</iters>
      <sor>1.3</sor>
    </solver>
    <constraints>
      <cfm>0.0</cfm>
      <erp>0.2</erp>
      <contact_max_correcting_vel>100.0</contact_max_correcting_vel>
      <contact_surface_layer>0.001</contact_surface_layer>
    </constraints>
  </ode>
</physics>

Key Parameters:

• max_step_size: Smaller = more accurate, slower
• real_time_factor: 1.0 = real-time, >1 = faster than real-time
• iters: Solver iterations (more = stable, slower)

Example Scenarios

Navigation Maze

<!-- Create maze walls -->
<model name="maze">
  <static>true</static>
  <!-- Wall 1 -->
  <link name="wall_1">
    <pose>5 0 1 0 0 0</pose>
    <collision name="collision">
      <geometry><box><size>0.2 10 2</size></box></geometry>
    </collision>
    <visual name="visual">
      <geometry><box><size>0.2 10 2</size></box></geometry>
      <material><ambient>0.3 0.3 0.3 1</ambient></material>
    </visual>
  </link>

  <!-- Add more walls to create maze pattern -->
  <!-- ... -->
</model>

Manipulation Test (Cluttered Table)

<!-- Table with random objects -->
<include>
  <uri>model://table</uri>
  <pose>0 0 0 0 0 0</pose>
</include>

<include>
  <uri>model://coke_can</uri>
  <pose>0.2 0.1 1.0 0 0 0</pose>
</include>

<include>
  <uri>model://beer</uri>
  <pose>-0.1 0.2 1.0 0 0 0</pose>
</include>

<include>
  <uri>model://bowl</uri>
  <pose>0.3 -0.2 1.0 0 0 0</pose>
</include>

Domain Randomization

For robust sim-to-real transfer, randomize environment properties:

# Python script to generate randomized worlds
import random
import xml.etree.ElementTree as ET

def randomize_world():
    # Randomize object positions
    for i in range(10):
        x = random.uniform(-5, 5)
        y = random.uniform(-5, 5)
        # Generate SDF with random poses

    # Randomize lighting
    ambient = random.uniform(0.2, 0.8)

    # Randomize friction
    friction = random.uniform(0.5, 1.5)

    # Save world file
    # ...

Exercises

1. Create a warehouse environment with shelves, boxes, and a navigation path
2. Design an outdoor scenario with terrain, trees, and uneven ground
3. Build a manipulation testbed with a table and 5 graspable objects
4. Implement domain randomization - script that generates 10 variations of a world
5. Test lighting effects - create worlds with different lighting conditions and observe camera sensor output

Summary

Environment creation in Gazebo enables testing robots in diverse, realistic scenarios before real-world deployment. By designing custom worlds with obstacles, terrain, lighting, and physics tuning, developers can validate navigation, manipulation, and perception algorithms comprehensively.

Further Reading

• Gazebo SDF Specification
• Gazebo Model Database
• Building a World Tutorial
• Domain Randomization for Sim-to-Real Transfer