I'm experimenting with ROS Kinetic on an Odroid Xu4, running Ubuntu 16.04, together with an iRobot Create2 and a RPLIDAR A1.
My current goal is to build a simple self-navigating robot with the [create_autonomy](https://github.com/autonomylab/create_autonomy) package. For that im using [move_base](http://wiki.ros.org/move_base), [rplidar](http://wiki.ros.org/rplidar) and [gmapping](http://wiki.ros.org/gmapping?distro=kinetic).
With gmapping installed via `sudo apt-get install ros-kinetic-gmapping` I managed to get reasonable outputs for laser, map and costmaps in rviz. But when I tried to set a 2D Nav Goal I got `transform require extrapolation into the future` and `global frame: odom plan frame size 2: map` errors.
Following the tips [here](https://answers.ros.org/question/227694/tf-error-transform-require-extrapolation/) and [here](https://github.com/ros-industrial/industrial_training/issues/77) I purged the gmapping package and reinstalled it via `git clone` from [github](https://github.com/ros-perception/slam_gmapping) to my catkin ws.
Before building with catkin it I changed the `ros::Time(0)` parts in `slam_gmapping.cpp`.
As a result I now have no information in the costmap-topics, while `/map` and `/rplidar_scan` are working perfectly fine.
I did no changes to my [launchfile](https://pastebin.com/UBsJm0Bn) or my other configs, my tf-tree hasn't changed too.
I have absolutely no idea whats going wrong and just overwhelmed by the complexity of ROS (I'm playing only since a few months with it). Has someone an idea whats the problem with my setup? Or did a make a general mistake?
↧
