Add parameter to os_cloud_node to control the sensor frames publishing

[abAgriRobot]

Related Issues & PRs

• resolves: Running the ROS2 driver with intra-process communication when using composition #197

Summary of Changes

By default the os_cloud_node publishes the static transformation for the sensor_frame, lidar_frame and imu_frame.
In some cases this transformations are defined in a higher level URDF file, and having the node publishing the frames may create conflicts in the tf tree.
We added a parameter to control if the node publishes the frames.

    ros__parameters:
        sensor_frame: 'os_sensor'
        lidar_frame: 'os_lidar'
        imu_frame: 'os_imu'
        timestamp_mode: "TIME_FROM_ROS_TIME" # this value needs to match os_sensor/timestamp_mode
        publish_sensor_frames: True

The parameter defaults to True so the behavior of the node doen't change unless explicitly define

       publish_sensor_frames: False

Validation

[Samahu]

Renaming the names of the frames to some other names doesn’t suffice?

[abAgriRobot]

The pointcloud and imu data use the defined frames as frame_id, renaming them would cause the data to be published in the wrong frame.

My config file looks similar to this:

ros__parameters:
        sensor_frame: 'my_vehicle_lidar_front/laser'
        lidar_frame: 'my_vehicle_lidar_front'
        imu_frame: 'my_vehicle_lidar_front/imu'
        timestamp_mode: "TIME_FROM_ROS_TIME" # this value needs to match os_sensor/timestamp_mode
        publish_sensor_frames: True

And I want the pointcloud data to have frame_id=my_vehicle_lidar_front/laser and the imu data frame_id=my_vehicle_lidar_front/imu. In the current version the node will do this and in addition will publish the static transforms
In our application, we define this frames in a higher level URDF since we still need them for simulation and replaying rosbag data. So the frames that the lidar need will be available and the node doesn't require to initialize them.