scenic.simulators.carla

Scenic world model for the CARLA driving simulator.

This model is designed to be used with the CARLA interface to the VerifAI toolkit. See the VerifAI repository for further documentation and examples.

The model currently supports vehicles, pedestrians, and props. Vehicles have an agent parameter, which specifies the agent to be used to control the vehicle.

In addition, the model uses several global parameters to control weather (descriptions are from the CARLA Python API reference):

  • cloudiness (float): Weather cloudiness. It only affects the RGB camera sensor. Values range from 0 to 100.

  • precipitation (float): Precipitation amount for controlling rain intensity. It only affects the RGB camera sensor. Values range from 0 to 100.

  • precipitation_deposits (float): Precipitation deposits for controlling the area of puddles on roads. It only affects the RGB camera sensor. Values range from 0 to 100.

  • wind_intensity (float): Wind intensity, it affects the clouds moving speed, the raindrop direction, and vegetation. This doesn’t affect the car physics. Values range from 0 to 100.

  • sun_azimuth_angle (float): The azimuth angle of the sun in degrees. Values range from 0 to 360 (degrees).

  • sun_altitude_angle (float): Altitude angle of the sun in degrees. Values range from -90 to 90 (where 0 degrees is the horizon).

Submodules

model

Scenic world model for traffic scenarios in CARLA.

map

Stub to allow changing the map without having to change the model.

interface

Support code for the CARLA world model.

car_models

prop_models