Track taken from Georgia Tech AutoRally project: https://github.com/AutoRally
Robot taken from Andres Garcia: https://github.com/AndresGarciaEscalante/Real-Time-Appearance-Based-Mapping
Robot modified to use Intel D455 parameters in RGBD camera. mapping.launch file modified to load configurations from .yaml and localization mode added for benchmarking purposes. Added a ground truth for benchmarking purposes as well
Can be paired with https://github.com/sisaha9/slamevaluations to run benchmarks on generated data
- Run uncluttered_world_proper tag from dockerhub: https://hub.docker.com/repository/docker/sisaha9/180asims/general
- Everything has been sourced already. Open 4 more terminals and connect them to the same container instance with docker exec
- Run roslaunch autorally_gazebo autoRallyTrackWorldMine.launch in the main terminal
- To open teleop type rosrun teleop_twist_keyboard teleop_twist_keyboard.py in terminal 2
- To start RTABMAP mapping type roslaunch autorally_gazebo mapping.launch localization:=false in terminal 3
- To generate ground truth data first run rosrun rtabmap_ros odom_msg_to_tf odom:=ground_truth/state _frame_id:=robot_footprint_gt in terminal 4
- In terminal 5 run rosrun tf static_transform_publisher 0 0 0 0 0 0 world map 100
- Move around the track till you feel you have gained sufficient map data
- Save your rtabmap.db file and kill all the processes
- Follow steps 3-7 again except when loading the mapping.launch you will set localization:=true. Move around the terminal till you feel like the robot has localized
- Run rtabmap-report --poses /root/.ros/rtabmap.db. This will generate 3 text files in the .ros folder. Copy them out of the container and use the slamevaluations tool linked above to generate metrics and plots on it
- To hyperparameter tune edit the .yaml file in catkin_ws/src/autorally_gazebo/params/param1.yaml. Redo all the steps above to generate the data. Soon a library to process all this information will be released so that you can efficiently hyperparameter tune these
- Install ros-melodic-desktop-full
- Install other dependencies from the Dockerfile
- source /opt/ros/melodic/setup.bash
- Run catkin_make in the catkin_ws folder
- source the devel/setup.bash in catkin_ws (or run setup.sh)
- Open 4 more terminals and source the melodic and devel again (or run setup.sh)
- Follow steps 3-12 above