Application of a Simultaneous Localization and Mapping (SLAM) algorithm inside a simulated environment in Gazebo with the RTAB-Map package.

RTAB-Map (Real-Time Appearance Based Mapping) is an RGB-D Graph-SLAM method (that solves the Full-SLAM problem). It uses a loop closure detector to determinate whether a new image is obtained from a previously visited location or a new, unvisited location. When a loop closure hypothesis is accepted, a new constraint is added to the map’s graph and then a graph optimizer minimizes the errors in the map. RTAB-Map utilizes a memory management technique to limit the number of locations considered as candidates during loop closure detection, which allows it to be performed in real-time.

Note: For best performance, the features in the environment need to be distinguishable (not visually similar).

The robot uses an RGB-D camera and a LiDAR to localize itself and create a map of the environment.

The project consists of the following parts:

• A ROS package that launches a custom robot model in a custom Gazebo world. The robot model was extended from the previous project RoboticsND-Where-Am-I and now employs an RGB-D camera.
• Interface with the ROS RTAB-Map package to perform SLAM.
• Teleoperation of the robot to explore and create a map of the environment.
• Database analysis to explore the generated maps, loop closures and other information about the mapping.

There is only one package in this project.

• my_robot: This package holds the robot, the Gazebo world and the launch files to interface with the RTAB-Map package.

The directory structure is depicted below:

.RoboticsND-Map-My-World        
├── my_robot                     # my_robot package
│   ├── CMakeLists.txt           # compiler instructions
│   ├── launch
│   │   ├── localization.launch  # interface with RTAB-Map for localization with an already created map
│   │   ├── mapping.launch       # interface with RTAB-Map to create a map
│   │   ├── robot_description.launch
│   │   └── world.launch
│   ├── meshes
│   │   └── hokuyo.dae
│   ├── package.xml
│   ├── rviz
│   │   └── rvizconfig_where_am_i.rviz
│   ├── urdf
│   │   ├── my_robot.gazebo
│   │   └── my_robot.xacro
│   └── worlds
│       └── myworld_project4.world
└── images                       # simulation images
    ├── 2D_map.png
    ├── 3D_map.png
    └── rtab-map.png

The project was developed on Ubuntu 20.04 LTS with:

• ROS Noetic
• Gazebo 11.11.0

The following dependencies need to be installed:

$ sudo apt-get update && sudo apt-get upgrade -y
$ sudo apt-get install ros-noetic-rtabmap-ros
$ sudo apt-get install ros-noetic-teleop-twist-keyboard

To run this project, you must have ROS and Gazebo installed.

$ mkdir -p /catkin_ws/src/
$ cd catkin_ws/src/
$ catkin_init_workspace

$ git clone https://github.com/elena-ecn/RoboticsND-Map-My-World.git

$ cd /catkin_ws/
$ catkin_make

$ source devel/setup.bash

$ roslaunch my_robot world.launch

In a new terminal (Ctrl+Shift+T), run:

$ source devel/setup.bash
$ roslaunch my_robot mapping.launch

In a new terminal (Ctrl+Shift+T), run:

$ source devel/setup.bash
$ rosrun teleop_twist_keyboard teleop_twist_keyboard.py

Teleoperate the robot through your keyboard to explore the environment. Once done, feel free to stop all running nodes. The rtabmap-ros package will automatically save the created map along with the localized trajectory in a database file ~/.ros/rtabmap.db.

In a terminal, run:

$ rtabmap-databaseViewer ~/.ros/rtabmap.db

To view the relevant information:

• Select 'yes' to using the database parameters
• View -> Constraints View and Graph View

To see the 3D Map:

• Edit -> View 3D Map

If you would like to perform localization with the map you created, there are some parameters that need to be changed in the rtabmap-ros package. So, instead of the mapping.launch, use:

$ roslaunch my_robot localization.launch

The contents of this repository are covered under the MIT License.