Artificial Intelligence-Driven Autonomous Vehicles: Current Developments and the Future Prospects

Abstract

Artificial intelligence (AI) technology is profoundly transforming the field of autonomous driving, propelling it from theory to practical application. This paper systematically reviews the key technological advancements in AI-driven autonomous driving. Recognition and control algorithms based on deep learning and reinforcement learning have enhanced the safety of real-time decision-making. Multisensor fusion and vehicle-to-everything (V2X) communication technologies have strengthened environmental perception and vehicle–road cooperation capabilities. The combination of computer vision and lidar has enabled high-precision 3D modeling. Currently, the global market is experiencing rapid growth. China, which relies on the “5+6” strategy and policy pilots, is accelerating the implementation of this technology. Levels 2 and 3 (L2/L3) systems have been commercialized, and Level 4 (L4) systems have entered the demonstration operation stage. However, an insufficient perception of complex environments, the “black box” problem of decision-making algorithms, and hardware computing power bottlenecks remain the main challenges for higher-level autonomous driving. In the future, promoting the development of technology toward Level 5 (L5) through the research and development of explainable AI algorithms, breakthroughs in domestic chips, and cross-industry collaboration. At the same time, an ethical framework centered around people and an intelligent transportation ecosystem should be constructed.