Closed Roosstef closed 3 years ago
Does anyone else have this problem? Does anyone have any idea if this is a carla or rosbridge problem?
Hi @Roosstef. Sorry for the delay. I think this is definitely a bug in the ROS bridge side. It seems that there is a race condition between the publication of the clock data and the generation of the header for some sensors. I've pushed some changes in https://github.com/carla-simulator/ros-bridge/tree/joel-mb/fix_timestamp that should solve the synchronization problem. Could you verify that it solves the issue for your use case?
@joel-mb your solution fixed the issue. Thank you!
Hey all,
i think the same or similar issue appears again in master dac9e72 with carla 0.9.12
My setup:
Ubuntu 20.04 Ros noetic Carla 0.9.11 Ros bridge master (dac9e72)
sequence numbers dont match time stamps for carla/objects (gt) and /carla/ego_vehicle/lidar0 (pc)
@joel-mb do you think we should reopen this issue?
Hey all,
I noticed a possible problem with the syncronization between sensor data and /carla/objects.
My setup:
In my example I start the ros_bridge with 2 radar sensors , "sensor.pseudo.tf", "sensor.pseudo.objects" and "sensor.pseudo.markers".
I use a simple script to print the two topics of the sensors and the /carla/objects
In the attached file there are different timestamps between the sensor data and the /carla/objects associated with header.seq
In previous Carla version 0.9.10.1 the same script causes no issues.