Leveraging CARLA to create End-to-End Autonomous Driving Solutions

End-to-end solutions are increasingly becoming the standard in autonomous driving, surpassing traditional modular approaches. In this talk, we demonstrate how the CARLA open-source simulator can be employed to train and validate end-to-end autonomous driving solutions. CARLA provides a versatile platform for simulating diverse scenarios, allowing researchers and engineers to develop and test AI models in a safe, controlled environment. Additionally, we discuss the critical process of transferring these solutions from simulation to real-world vehicles, exploring the challenges and successes encountered in bridging the gap between virtual and physical autonomy. Attendees will gain insights into how CARLA accelerates the development cycle for autonomous driving and facilitates the deployment of robust, real-world applications.