CARLA

CARLA is an open-source simulator for autonomous driving research. As such CARLA data is synthetic and can be generated with varying sensor and environmental conditions. The following documentation is largely incomplete and merely describes the provided demo data.

Quick Links

Paper	CARLA: An Open Urban Driving Simulator
Website	carla.org/
Code	github.com/carla-simulator/carla
License	MIT License
Available splits	n/a

Available Modalities

Name	Available	Description
Ego Vehicle	✓	Depending on the collected dataset. For further information, see EgoStateSE3.
Map	✓	We included a conversion method of OpenDRIVE maps. For further information, see MapAPI.
Bounding Boxes	✓	Depending on the collected dataset. For further information, see BoxDetectionWrapper.
Traffic Lights	X	n/a
Pinhole Cameras	✓	Depending on the collected dataset. For further information, see PinholeCamera.
Fisheye Cameras	X	n/a
LiDARs	✓	Depending on the collected dataset. For further information, see LiDAR.

Download

n/a

Installation

n/a

Dataset Specific

Box Detection Labels

class py123d.conversion.registry.DefaultBoxDetectionLabel

Default box detection labels used in 123D. Common labels across datasets.

EGO = 0

VEHICLE = 1

TRAIN = 2

BICYCLE = 3

PERSON = 4

ANIMAL = 5

TRAFFIC_SIGN = 6

TRAFFIC_CONE = 7

TRAFFIC_LIGHT = 8

BARRIER = 9

GENERIC_OBJECT = 10

to_default()

Inherited, see superclass.

Return type:

DefaultBoxDetectionLabel

LiDAR Index

class py123d.conversion.registry.DefaultLiDARIndex

Default LiDAR indices for XYZ point clouds.

X = 0

Y = 1

Z = 2

Dataset Issues

n/a

Citation

If you use CARLA in your research, please cite:

@article{Dosovitskiy2017CORL,
  title = {{CARLA}: {An} Open Urban Driving Simulator},
  author = {Alexey Dosovitskiy and German Ros and Felipe Codevilla and Antonio Lopez and Vladlen Koltun},
  booktitle = {Proceedings of the 1st Annual Conference on Robot Learning},
  year = {2017}
}