carla
carla copied to clipboard
How do I import newly added sensors into a Carla Package
Hey Guys,
I have added a new lidar sensor in carla. This sensor is working in the source carla i.e. when I do make launch and run the python3 manual_control.py
(I can see its data in rviz). Then, I went on to package this sensor using the command make package
, which creates a new directory Dist and a packaged carla inside it. My Dist folder looks like this,
I went inside CARLA_Shipping_e0e00321-dirty/LinuxNoEditor/ directrory, and in one terminal I ran the carla server using ./CarlaUE4.sh
and in anoher terminal I ran a python client and spawned my new lidar in it. But I am getting the segmentation fault error. Look at video below,
sensor testing in packaged carla
I thought there maybe an issue with my new lidar sensor implementation (though it is working with source build carla). I went on to check the safe distance sensor as described here create new sensor tutorial. I packaged it using make package
and ran the server and the python client but I am getting the same segmentation fault error.
Can anyone help me with this?
Note: Before this I have also added a new vehicle and its working perfectly fine in the packaged version of Carla. so, I am not sure whats going wrong here.
CARLA version: 0.9.13
Platform/OS: Ubuntu 18
Problem you have experienced: new sensor added not working in a packaged carla
What you expected to happen: the newly added sensor should be working in the packaged carla
Steps to reproduce: Do make package
in source carla and try to test the newly added sensor in the packaged carla.
Other information (documentation you consulted, workarounds you tried):
You are using the official python package, but your new sensor API doesn't exist in it. So you will get a segmentation error when you try to use it.
You should use the egg or whl file build by yourself, not the official package.
This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.