Adding a new way to replay rosbags

facontidavide commented 4 years ago

The slam_gmapping_replay node is not very easy to debug using RViz.

Users may find more convenient this new approach that:

executes gmapping at maximum speed allowed by the CPU.
all scans are used, none is discarted. This makes the mapping more deterministic.
Relevant messages, such as TF and scan, are republished.
/clock is republished too.

Usage:

        rosrun gmapping slam_gmapping scan:=/your_scan _rosbag:=/path/your/rosbag.bag

I think it is beneficial because is solve the old question that many people asked themselves: "How much can I increase the --rate of rosbag play, without affecting the quality of the map"?

With this new mode, rosbag is replayed at the "optimal" speed. in my case, it goes 4-5 times faster than realtime on a pretty receint I7 processor.

k-okada commented 3 years ago

@facontidavide thanks for enhancement, could you update the https://wiki.ros.org/gmapping too? Thansk in advance.

k-okada commented 3 years ago

@facontidavide sorry coudl you re-open this PR? with add tf2 related stuff on package.xml and CMakeLists.txt

ros-perception / slam_gmapping

Adding a new way to replay rosbags #80