keyence_driver: published point clouds have 'made up' timestamp

[gavanderhoorn]

Point clouds published by the driver node have a timestamp, but it is set to the time at which the conversion from profile data to Point cloud began (see driver.cpp, L514).

For accurate matching of profiles against TF data, it is probably necessary to improve this. However, this might prove difficult, as the Keyence protocol does not encode any time information in its messages, nor in the profile data.

A field with a similar function is the encoder_count field, but that is (as the name suggests) only used to affix an encoder count to a profile.

One option might be to move the setting of the ROS hdr.stamp field up to where the initial Keyence data packet is received, and in addition make sure that only POS_CURRENT is used for the byPosMode field.

[phvass]

System currently works as-is. This is a nice-to-have for the current project.