RTABMAP (Real-Time Appearance-Based Mapping) is a RGB-D Graph-Based SLAM approach based on an incremental appearance-based global Bayesian loop closure detector.
RTABMAP combines gmapping with 3D SLAM in order to do autonomous navigation (in the next tutorials) while 3D mapping. Along with 3D point cloud data from the depth sensor, RTABMAP also collects 2D and 3D laser scan data.
These instalation steps are to be done on BOTH the Turtlebot and Master computers.
sudo apt-get install ros-kinetic-rtabmap-ros -y
In order to start and configure RTABMAP properly, we need to create our own launch file on the Turtlebot.
mkdir ~/workspace
cd ~/workspace
cd ~/workspace
roslaunch turtlebot_bringup minimal.launch
cd ~/workspace
roslaunch rtabmap.launch database_path:=~/workspace/rtabmap.db
database_path:=
is the location of the database file
roslaunch rtabmap_ros demo_turtlebot_rviz.launch
roslaunch turtlebot_teleop keyboard_teleop.launch
In order to collect data fully, you will need to stop the turtlebot periodically and slowly turn the turtlebot in circles
/camera/rgb/image_raw
RTABMAP provides their own tool to visualize and manipulate data.
With the rtabmap-databaseViewer
tool, you can:
cd ~/workspace
scp turtlebot@IP_OF_TURTLEBOT:~/workspace/rtabmap.db .
rtabmap-databaseViewer
to view the rtabmap data
cd ~/workspace
Coming soon, we are working out the best parameters to use with this Turtlebot setup