Revision history [back]

As far as I know the IBEOs have those channels at different pitch angles, right?

In that case you have two possibilities (besides creating your own message). The first would be to convert the laser data to a pointcloud as Ivan Dryanovski suggested. The second would be to publish the four laser channels as four laser scan topcis and use the tf library. Each scan will have a different frame_id associated in the header and you use a static transform publisher that publishes the transforms according to the relative pose of the scans to the sensor.

Both versions will display correctly in rviz. The second has four different topcis, though. The advantage is that it publishes LaserScan messages making it compatible with software that uses LaserScans.

The best would probably be to follow an approach similar to the kinect driver: You advertise every possibility in your driver, but you will only publish (and convert laser to pointcloud) if a publisher has subscribers.