This paper provides a comprehensive survey of deep learning techniques used in autonomous driving, covering the current state-of-the-art in AI-based self-driving architectures, driving scene perception, path planning, behavior arbitration, and motion control. The authors discuss both modular perception-planning-action pipelines and End2End systems, highlighting the strengths and limitations of deep learning approaches. Key topics include deep convolutional neural networks (CNNs), recurrent neural networks (RNNs), and deep reinforcement learning (DRL). The paper also addresses challenges such as safety, training data sources, and computational hardware requirements. Additionally, it explores the debate between camera and LiDAR sensing technologies, object detection and recognition, semantic segmentation, localization, and the use of occupancy maps for perception. The survey concludes with an overview of motion controllers, including learning controllers and End2End control systems, and discusses the integration of deep learning in these systems.This paper provides a comprehensive survey of deep learning techniques used in autonomous driving, covering the current state-of-the-art in AI-based self-driving architectures, driving scene perception, path planning, behavior arbitration, and motion control. The authors discuss both modular perception-planning-action pipelines and End2End systems, highlighting the strengths and limitations of deep learning approaches. Key topics include deep convolutional neural networks (CNNs), recurrent neural networks (RNNs), and deep reinforcement learning (DRL). The paper also addresses challenges such as safety, training data sources, and computational hardware requirements. Additionally, it explores the debate between camera and LiDAR sensing technologies, object detection and recognition, semantic segmentation, localization, and the use of occupancy maps for perception. The survey concludes with an overview of motion controllers, including learning controllers and End2End control systems, and discusses the integration of deep learning in these systems.