CARLA Autonomous Driving Simulator

CARLA (Car Learning to Act) is an open-source autonomous driving simulator built to support the development, training, and validation of autonomous driving systems. Developed by the Computer Vision Center (CVC) at Universitat Autonoma de Barcelona and the Barcelona Supercomputing Center, with support from Intel, CARLA was first released in 2017. Key features: built on Unreal Engine 4 for photorealistic rendering with advanced lighting, weather effects, and post-processing. Asset library: over 40 vehicle models, 50+ pedestrian models, and various props. Town environments: 8 pre-built maps including urban, suburban, rural, and highway scenarios. Sensor suite: RGB cameras, depth cameras, semantic segmentation cameras, LiDAR (configurable channels, rotation rate, range, points per second), radar, GNSS/GPS, IMU, collision detection. Weather control: programmable rain, fog, wind, cloudiness, sun position, and time-of-day. Traffic manager: programmable traffic generation with behavior parameters. Scenario Runner integration for defining and executing test scenarios in Python. OpenDRIVE map support for importing real road networks. Leaderboard: official benchmark for evaluating autonomous driving agents. ROS bridge for testing Autoware and other autonomous stacks. Python API for programmatic control. Dataset generation for training ML models. Cross-platform: Linux, Windows. MIT.