Configuration

This overview is a compact reference for all available Docker Compose profiles and .env variables. The basic workflow and concrete examples are described in Getting Started.

The recommended way to edit the configuration is the GUI. It exposes the same variables listed below, applies presets, validates the configuration, and writes valid changes back to .env:

docker compose up gui

Open http://localhost:8501 after the GUI has started.

Environment Variables

Environment variables are key-value settings that configure the Docker Compose setup. Some of them are also passed into the containers, but their main role in this repository is to select profiles, maps, vehicles, scenarios, overrides, and optional runtime behavior without editing the Compose files directly.

All variables below can be set through the GUI or manually in .env.

Variable	Purpose / Impact	Typical Values
COMPOSE_PROFILES	Active service groups for simulation, testing, perception, and planning.	carla,manual-testing,no-perception,planning, sumo,no-perception,planning
COMPOSE_FILE	Compose file chain. Examples override the standard docker-compose.yml with a more specialized file.	docker-compose.yml:examples/ml-planning/docker-compose.override.yml
VEHICLE	Vehicle profile for sensor frames (URDF), parameters, and controller tuning. Mostly relevant for CARLA.	karl, shuttle
SENSORS	Comma-separated list of CARLA sensor layout files to be used when spawning objects. SUMO does not use sensor layouts.	global-sensors.json,karl-minimal.json, global-sensors.json,shuttle-basic.json,rita.json
MAP	Selected prebuilt map. Leave empty when using custom OpenDRIVE mode.	CARLA: aldenhoven, campus, Town10HD; SUMO: campus, rita-intersection
SPAWN_POINT	Ego spawn point, given as x,y,z,heading in map coordinates, or as lat,lon,altitude,heading when suffixed with wgs84. Required for prebuilt maps.	50.787510,6.045938,260.0,90.0,wgs84, -1.0,-154.0,1.0,238.0
CUSTOM_OPENDRIVE	Custom OpenDRIVE (.xodr) map for CARLA standalone map mode. Not supported for SUMO.	TODO/custom.xodr
CUSTOM_LANELET	Matching Lanelet2 .osm map for route planning and map server. Required with custom OpenDRIVE.	TODO/custom.osm
ORIGIN_LAT	GPS origin latitude for a custom lanelet map frame. Set together with ORIGIN_LON or leave both empty.	50.782329
ORIGIN_LON	GPS origin longitude for a custom lanelet map frame. Set together with ORIGIN_LAT or leave both empty.	6.070377
LANELET_RELOAD	Lets the stack automatically select and reload Lanelet2 map data for prebuilt maps. The GUI sets this automatically.	true, false
SCENARIO_FILE	OpenSCENARIO file for CARLA manual-testing or automated-testing. The GUI can prefilter scenarios for the selected map.	TODO/example.xosc
USE_SIM_TIME	Enables ROS simulated time across the stack. Must stay true in this setup.	true
ROS_TRACING	Enables ROS tracing output for analysis.	false, true

Service Profiles

COMPOSE_PROFILES activates service groups from the included Docker Compose files. Pick one compatible value per group.

Group	Profiles	Compatibility / Effect
Simulation	carla, sumo	Selects the simulation backend. CARLA supports the full stack; SUMO is lightweight and more restricted.
Testing	no-testing, random-testing, manual-testing, automated-testing	CARLA only. SUMO does not use a testing profile. Use manual-testing for RViz scenario control.
Perception	no-perception, perception	perception is CARLA only. SUMO must use no-perception.
Planning	planning	CARLA and SUMO support planning.

Common examples:

COMPOSE_PROFILES=carla,random-testing,no-perception,planning
COMPOSE_PROFILES=carla,manual-testing,perception,planning
COMPOSE_PROFILES=sumo,no-perception,planning

Service Overview

The following overview lists the Docker Compose services involved in OpenADSim and shows which profiles activate them. Services marked always have no profile assigned and therefore start with every configuration.

CARLA Simulation

Service	Profiles	Role
carla-server	carla	Offscreen CARLA backend with custom DefaultEngine.ini
ros-middleware-bridge	carla	DDS-to-Zenoh bridge for CARLA ROS topics
carla-ros-bridge	carla	ROS 2 bridge; loads MAP or CUSTOM_OPENDRIVE
carla-spawn-objects	carla	Spawns ego vehicle and sensors from SENSORS; optional SPAWN_POINT
carla-converter	carla	Generates stack-relevant ego, object, and map topics
carla-simulation-adapter	carla	Maps CARLA outputs to /simulation/*; optional Lanelet2 map selection via LANELET_RELOAD
carla-ackermann-control	carla	Applies Ackermann control commands to the ego vehicle
carla-manual-control	carla	Manual teleoperation interface
carla-control-active-bridge	carla	Publishes /control/active from manual override state
carla-environment	random-testing	Adds random traffic and environment variation
carla-scenario-runner-ros-manual-testing	manual-testing	Interactive scenario execution in manual testing
carla-scenario-runner-ros-automated-testing	automated-testing	Batch scenario execution from SCENARIO_FILE

SUMO Simulation

Service	Profiles	Role
sumo-interface	sumo	Generates EgoData, ObjectList, and map information from SUMO
sumo-simulation-adapter	sumo	Maps SUMO outputs to /simulation/*; optional Lanelet2 map selection via LANELET_RELOAD
sumo-control-active-bridge	sumo	Publishes /control/active from manual override state

OpenADStack

Service	Profiles	Role
zenoh-router	always	Zenoh router for ROS 2 communication
robot-state-publisher	always	Publishes the vehicle URDF for VEHICLE
localization.lanelet2-map-server	always	Serves Lanelet2 maps with CUSTOM_LANELET, ORIGIN_LAT, and ORIGIN_LON
understanding.simple-object-tracking	perception	Tracks object motion from fused perception objects
planning.lanelet2-route-planning	planning	Computes a Lanelet2 route from ego data
planning.simple-planner	planning	Generates a reference trajectory from the enriched route
planning.trajectory-optimization	planning	Optimizes the final drivable trajectory
control.ackermann-trajectory-control	planning	Converts the optimized trajectory into Ackermann control commands

Monitoring and Tooling

Service	Profiles	Role
monitoring.rviz-carla	carla	RViz visualization with the CARLA display configuration
monitoring.rviz-sumo	sumo	RViz visualization with the SUMO display configuration
monitoring.bag-recorder	automated-testing	Records ROS bags of the executed scenarios
gui	gui	Web-based configuration GUI on port 8501