1. 車体: メガローバー ver3.0
2. PC: Jetson AGX Orin
3. Lidar: Livox Mid360
4. カメラ: Realsense D455
5. DCDCコンバータ: YDS-812

1. clone submoduleごとclone

git clone --recursive https://github.com/YasuChiba/megarover-ros2.git

1. build

cd workspace
colcon build
colcon build --cmake-args -DBUILD_TESTING=OFF
source install/setup.sh
pip3 install transforms3d

2. run

ホスト側で下記を実行

`xhost +local:` 
sudo chmod 777 /dev/video* 

必要に応じて以下をコンテナ内で実行

ros2 launch c_megarover msg_MID360_launch.py
ros2 launch c_megarover create_3dmap_launch.py rviz:=true
ros2 launch c_megarover create_2dmap_launch.py simulator:=false rviz:=false
ros2 launch c_megarover create_3dmap_rosbag_launch.py  rviz:=true rosbag_path:=/home/user/workspace/rosbag/rosbag2_2024_04_09-03_27_26
ros2 launch c_megarover create_3dmap_rosbag_launch.py  rviz:=true rosbag_path:=/home/user/workspace/rosbag/rosbag2_2024_04_10-01_35_06/ | grep -v "Failed to find match for fiel"

micro ros agentの開始.これを実行することで、車体側とROS2経由で通信が出来るようになる。
dockerコンテナをUSB接続前から立ち上げていると制御基板がコンテナ内から見えないかも.
permissionで怒られたらsudo chmod 666 /dev/ttyUSB0を実行。

ros2 run micro_ros_agent micro_ros_agent serial --dev /dev/ttyUSB0 --baudrate 115200

見えるトピックは以下のはず

• /rover_odo
• /rover_sensor
• /rover_twist

3. record/play

   ros2 bag record -a
   ros2 bag play hoge
   

4. export pcd SLAMの結果の点群を、rosbagから取り出すには....

   ros2 run c_megarover_common pointcloud_to_pcd_node --ros-args -r input:=/Laser_map
   

1. record
   start_recording.shをコンテナ外から実行。もしくはコンテナ内でros2 launch c_megarover record_all_launch.py simulator:=false rviz:=trueを実行。 workspace/rosbagにデータが格納。
2. slam
   1. 以下の２つをそれぞれ同時に実行
      • ros2 launch c_megarover create_3dmap_launch_rosbag.py
      • ros2 bag play rosbag/rosbag_2024-xxxx/ --topics /livox/imu /livox/lidar
   2. マッピングの終盤に以下を実行
      • ros2 run c_megarover_common pointcloud_to_pcd_node --ros-args -p prefix:=/home/user/workspace/pcd/ -r input:=/Laser_map
   3. pcdファイルが出力されたあとに以下を実行
      • ros2 run c_megarover_common pcd_to_occupancygrid_tool --ros-args -p pcd_file_path:=/home/user/workspace/pcd/1713279228.582533121.pcd -p output_file_path:=/home/user/workspace/maps/map2
      • ros2 run c_megarover_common pcd_to_occupancygrid_tool --ros-args -p pcd_file_path:=/home/user/workspace/maps/map2.pcd -p output_file_path:=/home/user/workspace/maps/map2 -p thres_point_count:=2

コンテナ外でstart_2d_navigation.shを実行. 参照する地図を変える場合workspace/container_entrypoints/start_2d_navigation.shを編集。

1. run

   ros2 launch c_megarover 3d_localization_launch.py simulator:=true rviz:=true map_file_path:=/home/user/workspace/pcd/sendagi.pcd
   
   ros2 bag play rosbag/rosbag2_2024_04_12-03_10_57/ --topics /livox/lidar /livox/imu --clock
   

   --clock is required when the simulator:=true.

1. run

   ros2 launch c_megarover navigation_launch.py simulator:=true rviz:=false map_file_path:=/home/user/workspace/maps/sendagi.pcd map_2d_file_path:=/home/user/workspace/maps/sendagi.yaml
   
   ros2 bag play rosbag/rosbag2_2024_04_12-03_10_57/ --topic /livox/imu /livox/lidar /rover_odo --clock
   

2. run(2d)

   ros2 launch c_megarover 2d_localization_launch.py simulator:=true rviz:=true map_file_path:=/home/user/workspace/maps/sendagi.pcd map_2d_file_path:=/home/user/workspace/maps/sendagi.yaml
   

slam_toolboxはodomが必須なので、create_2dmap_launch.pyはodomがある環境で試す。

(maybe) need to run ./setup_udev_rules.sh in librealsense2 on the host machine
ros2 launch realsense2_camera rs_launch.py initial_reset:=true
colcon build --cmake-args '-DBUILD_ACCELERATE_GPU_WITH_GLSL=ON'

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