Abstract: Autonomous driving technology for underground rubber-tired vehicles in coal mines plays a crucial role in promoting intelligent mining and enhancing the efficiency and safety of auxiliary transportation. This paper presents a comprehensive review of recent research progress on autonomous driving systems for underground rubber-tired vehicles, systematically analyzing the system architecture and technological evolution of key subsystems, including perception, localization, path planning, decision-making and control, as well as communication and scheduling. In conjunction with representative engineering practices in typical mining areas, the study examines current integration approaches and practical deployment challenges. It further summarizes key issues encountered under harsh underground conditions, such as sensor degradation, positioning difficulties, limited communication reliability, and lack of system standardization. To address these challenges, the paper explores emerging trends in multimodal fusion perception, adaptive perception enhancement, multi-agent collaborative scheduling, modular system design, and platform standardization. The insights provided herein offer valuable references for the development and intelligent upgrading of autonomous transportation systems in coal mines.