How is tf data handled by velodyne_driver?
Hello,
I am going to use a Velodyne VPL16 along with the velodyne_driver to publish a PoinCloud2 message to ROS.
I have seen that the velodyne_node has the velodyne's frame_id as a parameter and I wanted to know if this is used just to include the frame_id in the header or if any transformation is being performed while the packages are being read from the raw data provided by the actual Velodyne.
Let me explain: if the Velodyne turns at 10Hz, it means that it takes 0.1s to perform a full revolution. If the Velodyne moves at just 20 Km/h (~5.6m/s), by the time it finishes the revolution it will have moved about 0.56m since it started.
So, if points are passed to ROS via the velodyne_points topic after having completed a 360º sweep, what odometry information should I use? The sensor's position at the beginning of the revolution, at the end, in the middle?
That's why I asked if any transformation was being performed by the velodyne_driver to compensate its own displacement.
I think it just reads the relative coordinates and publishes them converted to PointCloud2, but isn't the movement of the Velodyne introducing a big error?