Tutorial for ROS and Gazebo

This is the code basis for ROS and Gazebo Tutorials and Assignments. We use a simulated differential drive robot equipped with

• IMU
• Odometer
• LiDAR
• Camera

Requirements

• ROS Kinetic (Ubuntu 16.04) or Melodic (Ubuntu 18.04), Installation Instructions can be found here
• Gazebo (recommended: Version 7.0), comes with the ROS full desktop version, otherwise Installation Instructions can be found here
• (optional) ROS-QTC-Plugin for using QT Creator as IDE, Installation Instructions can be found here

Additional Required Packages (installation instructions are given for ROS Melodic)

• gmapping

sudo apt-get install ros-melodic-gmapping

• teleop_twist_keyboard

sudo apt-get install ros-melodic-teleop-twist-keyboard

• map_server

sudo apt-get install ros-melodic-map-server

• amcl

sudo apt-get install ros-melodic-amcl

How to:

• Start by creating a catkin workspace folder, downloading the git repository and compiling the code

mkdir -p ~/tutorial_ws/src
cd ~/tutorial_ws/src 
git clone https://github.com/NRottmann/ROS_Gazebo_Tutorial.git
cd  ..
catkin_make

• Start the simulation

cd ~/tutorial_ws
source devel/setup.bash
roslaunch simulation_environment apartment.launch

• The Gazebo environment should have opened and something similar to the image below should have appeared:

• Now you can move the robot around by opening a new terminal and typing

rosrun teleop_twist_keyboard teleop_twist_keyboard.py

Important Topics (Message Types) provided by the Simulation Environment

• Published
  • /camera/image_raw (sensor_msgs/Image)
  • /imu (sensor_msgs/Imu)
  • /odom (nav_msgs/Odometry)
  • /scan (sensor_msgs/LaserScan)
• Subsrcibed
  • /cmd_vel (geoemtry_msgs/Twist)

Assignment:

1. Generate a Map of the Environment

• Start the simulation environment and gampping (hint: have a look into the mapping.launch file)
• Drive the robot around using the teleop_twist_keyboard until you the map is sufficient accurate (hint: you can have a look onto the map by using rviz)

rosrun teleop_twist_keyboard teleop_twist_keyboard.py

• Save the map using map_server

rosrun map_server map_saver -f myMapName

2. Find the Person in the Building (the robot will start at a random position outside the building)

• Create a ROS Node which subscribes to the pose information from the amcl package (Particle Filter) and searches for the missing person by publishing to the /cmd_vel topic. The robot should stop if it found the missing person. For localization you can upload your generated map (example: localization.launch)
• By running the person_detector (example: detection.launch), a new topic /person_detector will appear which publishes the message pal_person_detector_opencv/Detections2d. This message contains information about detected person in the camera image. For more information about the person detector, we refer to the ros wiki. For simplicity, we included the required parts of the pal repository into our tutorial repository.