CARLA is an open-source simulator for autonomous driving research. It has been developed from the ground up to support development, training, and validation of autonomous urban driving systems. CARLA simulator has been built to serve as a modular and flexible Application Programming Interface (API) to address a range of tasks involved in the problem of autonomous driving.
CARLA is an open platform and its content of urban environments is also free. This content has been created from scratch by a dedicated team of digital artists employed for this purpose. It includes urban layouts, a multitude of vehicle models, buildings, pedestrians, street signs, etc. The simulation platform supports flexible setup of sensor suites and provides signals that can be used to train driving strategies, such as GPS coordinates, speed, acceleration, and detailed data on collisions and other infractions. A wide range of environmental conditions can be specified, including weather and lighting conditions.
CARLA is used to study the performance of three approaches to autonomous driving: a classic modular pipeline, an end-to-end model trained via imitation learning and an end-to-end model trained via reinforcement learning. It is used to stage controlled goal-directed navigation scenarios of increasing difficulty. The approaches are evaluated in controlled scenarios of increasing difficulty, and their performance is examined via metrics provided by CARLA, illustrating the platform’s utility for autonomous driving research.