I use a hector slam node with my robot and my framework works well. I configured tf properly and can visualize the map in rviz. I also reset the map once every few minutes because of my application and this doesn't seem to have any ill effects.

However, hector slam often segfaults with the message:

[hector_mapping-1] process has died [pid 9493, exit code -11, cmd /opt/ros/kinetic/lib/hector_mapping/hector_mapping __name:=hector_mapping __log:=/home/12345/.ros/log/df3ebdde-660a-11e7-b4fd-3c4a92d5191e/hector_mapping-1.log].
log file: /home/12345/.ros/log/df3ebdde-660a-11e7-b4fd-3c4a92d5191e/hector_mapping-1*.log

When I go to look for the log file, it isn't there and I am sad.

As I have other nodes running which rely on hectorslam, is there a way to restart it once it segfaults or in general restart a node? Has anyone else had this issue with hector slam?

Thanks

• Ubuntu 16.04
• ROS Kinetic

• hector slam parameters

  PARAMETERS

  • /hectormapping/advertisemap_service: True
  • /hectormapping/baseframe: base_frame
  • /hectormapping/laserzmaxvalue: 3.0
  • /hectormapping/laserzminvalue: -3.0
  • /hectormapping/mapframe: map
  • /hectormapping/mapmultireslevels: 2
  • /hectormapping/mapresolution: 0.1
  • /hectormapping/mapsize: 512
  • /hectormapping/mapstart_x: 0.5
  • /hectormapping/mapstart_y: 0.5
  • /hectormapping/mapupdateanglethresh: 0.06
  • /hectormapping/mapupdatedistancethresh: 0.4
  • /hectormapping/odomframe: base_frame
  • /hectormapping/pubmapodomtransform: True
  • /hectormapping/scansubscriberqueuesize: 1
  • /hectormapping/scantopic: scan
  • /hectormapping/tfmapscanmatchtransformframename: laser
  • /hectormapping/updatefactor_free: 0.4
  • /hectormapping/updatefactor_occupied: 0.9
  • /hectormapping/usetfposestart_estimate: False
  • /hectormapping/usetfscantransformation: False
  • /rosdistro: kinetic
  • /rosversion: 1.12.7

  NODES / hectormapping (hectormapping/hectormapping) laserlinktobaselink (tf/statictransform_publisher)