This repository is a curated list of resources for the CARLA simulator, targeting developers and researchers in autonomous driving. It provides a comprehensive collection of tutorials, code examples, and related projects to facilitate the development, training, and validation of self-driving systems.

How It Works

The list categorizes resources by application area within autonomous driving, including Reinforcement Learning, Imitation Learning, Multi-Agent systems, Perception (Detection, Segmentation), Data Collection, and Control. This structure allows users to quickly find relevant tools and examples for their specific research or development needs within the CARLA ecosystem.

Quick Start & Requirements

• CARLA simulator installation is required. Specific CARLA versions are often noted for compatibility with listed projects.
• Many projects require Python and specific libraries like TensorFlow or PyTorch.
• Links to official CARLA documentation and community resources are provided.

Highlighted Details

• Extensive coverage of Reinforcement Learning and Imitation Learning approaches for CARLA.
• Resources for data collection and generation, including KITTI dataset format compatibility.
• Numerous projects focusing on perception tasks like object detection and semantic segmentation.
• Includes examples for multi-agent systems and various control strategies.

Maintenance & Community

• The list is community-driven, with contributions welcomed via pull requests.
• Guidelines are provided for suggesting deprecations of outdated resources.
• A link to the CARLA Discord server is available for community interaction.

Licensing & Compatibility

• The list itself is licensed under Creative Commons CC0.
• Compatibility with specific CARLA simulator versions (e.g., 0.8.x, 0.9.x) is frequently noted for individual projects.

Limitations & Caveats

Many listed projects are tied to older CARLA versions (e.g., 0.8.x), potentially requiring significant adaptation for current CARLA releases. The maintenance status of individual linked repositories is not consistently indicated.