Traffic Generation in OpenCDA¶

Complete Documentation

For comprehensive traffic generation documentation, please visit the Official OpenCDA Documentation.

OpenCDA supports two main approaches for traffic generation:

CARLA Traffic Manager: Native CARLA traffic simulation
SUMO Co-simulation: External traffic simulation with SUMO

CARLA Traffic Manager¶

The CARLA Traffic Manager provides built-in traffic simulation capabilities.

Configuration¶

Configure traffic in your YAML file under carla_traffic_manager:

carla_traffic_manager:
  sync_mode: true
  global_distance: 5.0
  global_speed_perc: -100
  auto_lane_change: false
  random: false
  vehicle_list:
    - spawn_position: [100, 100, 0.3, 0, 0, 0]
    - spawn_position: [150, 100, 0.3, 0, 0, 0]

Parameters¶

sync_mode: Synchronization with CARLA world
global_distance: Minimum distance between vehicles (meters)
global_speed_perc: Speed adjustment (-100 = 60km/h for 30km/h limit)
auto_lane_change: Enable/disable automatic lane changes
random: Randomize vehicle models and colors
vehicle_list: Specific spawn positions and configurations

Vehicle Spawning Methods¶

Method 1: Explicit Positions¶

vehicle_list:
  - spawn_position: [x, y, z, roll, yaw, pitch]
  - spawn_position: [x2, y2, z2, roll2, yaw2, pitch2]

Method 2: Range-based Spawning¶

vehicle_list: ~
range:
  - [x_min, x_max, y_min, y_max, x_step, y_step, vehicle_count]

SUMO Traffic Management (Co-simulation)¶

SUMO provides more sophisticated traffic simulation capabilities.

Prerequisites¶

SUMO installation
SUMO configuration files (.sumocfg, .net.xml, .rou.xml)

Configuration¶

sumo:
  port: 8813
  host: "127.0.0.1"
  gui: true
  client_order: 1
  step_length: 0.05

SUMO Files Structure¶

scenario_folder/
├── Town06.sumocfg    # Main configuration
├── Town06.net.xml    # Road network
└── Town06.rou.xml    # Route definitions

Co-simulation Process¶

Initialize: Start both CARLA and SUMO
Synchronize: Coordinate timesteps between simulators
Exchange: Share vehicle states and commands
Update: Apply changes in both environments

Traffic Flow Configuration¶

Basic Traffic Flow¶

carla_traffic_manager:
  global_distance: 4.0
  global_speed_perc: -200 # 50km/h
  vehicle_list:
    - spawn_position: [-100, 4.8, 0.3, 0, 0, 0]
    - spawn_position: [-150, 4.8, 0.3, 0, 0, 0]

Mixed Traffic Scenarios¶

carla_traffic_manager:
  vehicle_list:
    - spawn_position: [-100, 4.8, 0.3, 0, 0, 0]
      vehicle_speed_perc: -150 # Individual speed setting
    - spawn_position: [-150, 8.3, 0.3, 0, 0, 0]
      vehicle_speed_perc: -250

Dense Traffic¶

carla_traffic_manager:
  global_distance: 2.0
  vehicle_list: ~
  range:
    - [-500, 500, -20, 20, 25, 4, 50] # 50 vehicles in area

Advanced Features¶

Traffic Light Management¶

carla_traffic_manager:
  ignore_lights_percentage: 20 # 20% ignore traffic lights

Dynamic Behavior¶

Aggressive driving: Lower global_distance, higher speeds
Cautious driving: Higher global_distance, lower speeds
Mixed behavior: Individual vehicle configurations

Performance Considerations¶

CARLA Traffic Manager¶

Pros: Easy setup, integrated with CARLA
Cons: Limited behavioral complexity
Best for: Simple scenarios, testing CAV algorithms

SUMO Co-simulation¶

Pros: Realistic traffic patterns, advanced behaviors
Cons: Complex setup, higher computational cost
Best for: Large-scale scenarios, traffic flow studies

Best Practices¶

Start Simple: Use CARLA Traffic Manager for initial testing
Gradual Complexity: Add more vehicles incrementally
Performance Monitoring: Watch simulation frame rate
Scenario Matching: Choose traffic density based on scenario needs

Troubleshooting¶

Common Issues¶

Low frame rate: Reduce vehicle count or disable visualizations
Collisions: Increase global_distance
Unrealistic behavior: Switch to SUMO co-simulation
Sync issues: Check step_length consistency

Debug Tips¶

Use CARLA spectator to observe traffic
Enable traffic manager debug mode
Monitor vehicle spawn success
Check for conflicting spawn positions

For more detailed information, refer to the Official OpenCDA Documentation.