V2X in Carla

V2X module integrated the autonomous system between multiple vehicles and the traffic light manager.

Map Info

map_info.json describes the traffic lights at each intersection used by the V2X module. The structure is:

{
  "intersections": {
    "<intersection_id>": [
      {
        "autoware_lane": <int>,
        "autoware_traffic_light": <int>,
        "traffic_light_position": { "x": <float>, "y": <float> }
      }
    ]
  }
}

Field meanings:

intersection_id — intersection label.
autoware_lane — Lane ID.
autoware_traffic_light — Traffic Light ID.
traffic_light_position — (x, y) in the Autoware map frame.

Running single vehicle scenario

Step 1: Running CARLA simulator

./CarlaUE4.sh -quality-level=Epic -world-port=2000 -RenderOffScreen -prefernvidia

Step 2: Run the zenoh_carla_bridge with V2X component (In Carla bridge container)

# Go inside "Carla bridge container"
./container/run-bridge-docker.sh
# Run zenoh_carla_bridge and Python Agent with V2X component
cd autoware_carla_launch
source env.sh
# Option A: rmw_zenoh
./script/bridge_rmw_zenoh/run-bridge-v2x-with-rmw_zenoh.sh
# Option B: ros2dds
./script/bridge_ros2dds/run-bridge-v2x.sh

Note

You can determine whether the bridge is running properly and verify that the V2X component is successfully retrieving the required information from the Carla simulator by checking if the terminal displays the following output…

INFO  zenoh_carla_bridge > Running Carla Autoware Zenoh bridge...
...
INFO: [Traffic Manager] Declaring Subscriber on 'vehicle/pose/**'...
...
INFO: [Intersection Manager] Declaring Queryable on 'intersection/**/traffic_light/**'...
...

Step 3: Run Autoware with the V2X light client (In Autoware container)

# Go inside "Autoware container"
./container/run-autoware-docker.sh
# Run Autoware and the V2X light client in one shot
cd autoware_carla_launch
source env.sh
# Option A: rmw_zenoh
# Format: ./script/autoware_rmw_zenoh/run-autoware-v2x-with-rmw_zenoh.sh <vehicle_id>
./script/autoware_rmw_zenoh/run-autoware-v2x-with-rmw_zenoh.sh
# Option B: ros2dds
# Format: ./script/autoware_ros2dds/run-autoware-v2x.sh <vehicle_id>
./script/autoware_ros2dds/run-autoware-v2x.sh

Note

<vehicle_id> must match CARLA agent’s rolename. (default is “v1”)

Step 4: Wait for Autoware to localize the vehicle, then set the 2D Goal Pose.

Step 5: Press the “Auto” button in Rviz and let Autoware autopilot the vehicle.

Running multiple vehicles scenario

Step 1: Running CARLA simulator

Step 2: Entering bridge container then executing…

cd autoware_carla_launch
source env.sh
# Option A: rmw_zenoh
./script/bridge_rmw_zenoh/run-bridge-two-vehicles-v2x-with-rmw_zenoh.sh
# Option B: ros2dds
./script/bridge_ros2dds/run-bridge-two-vehicles-v2x.sh

Step 3: Running Autoware container for 1st vehicle…

cd autoware_carla_launch
source env.sh
# Option A: rmw_zenoh
./script/autoware_rmw_zenoh/run-autoware-v2x-with-rmw_zenoh.sh v1
# Option B: ros2dds
./script/autoware_ros2dds/run-autoware-v2x.sh v1

Step 4: Running another Autoware container for 2nd vehicle…

cd autoware_carla_launch
source env.sh
# Option A: rmw_zenoh
./script/autoware_rmw_zenoh/run-autoware-v2x-with-rmw_zenoh.sh v2
# Option B: ros2dds
./script/autoware_ros2dds/run-autoware-v2x.sh v2

Step 5: Wait for Autoware to localize two vehicles, and then both set the 2D Goal Pose.

Step 6: Press the “Auto” button in Rviz and let two Autoware autopilot the vehicles