Autonomous Vehicles and AI: A Review of Perception, Decision-Making, and Safety Challenges

Abstract

Autonomous vehicles rely on AI-driven perception, planning, and control systems to navigate complex environments. This paper reviews advancements in deep learning, sensor fusion, and reinforcement learning techniques used in self-driving cars. We analyze the role of AI in object detection, trajectory prediction, and collision avoidance. Challenges such as adversarial attacks, edge computing constraints, and regulatory concerns are discussed. The review concludes with future directions for improving robustness and ethical considerations in autonomous mobility.