Carla allows for ros agents to initialize before the simulation is run.
Additionally, ROS-based agents must communicate to the Leadeboard when the stack initialization is finished. This communication should be done by publishing in the following topic /carla/hero/status std_msgs/Bool.
https://leaderboard.carla.org/get_started/
Currently this status "ok" message is published as soon as acting vehicle controller node starts up. Different sub-components of our code do however need more time than this to initialize. E.g. #272 (Perception loading models).
We could define an initialization checklist, which waits for all components to be properly initialized before sending this status ok.
This could be done adding a topic onto which all initialized components have to publish. If a message from all components was received we send the carla ok status to start the run.
Effort estimate
7
Definition of Done
[ ] Define an initialization sequence and protocol
Description
Carla allows for ros agents to initialize before the simulation is run.
Currently this status "ok" message is published as soon as acting vehicle controller node starts up. Different sub-components of our code do however need more time than this to initialize. E.g. #272 (Perception loading models).
We could define an initialization checklist, which waits for all components to be properly initialized before sending this status ok. This could be done adding a topic onto which all initialized components have to publish. If a message from all components was received we send the carla ok status to start the run.
Effort estimate
7
Definition of Done