THEORETICAL FOUNDATIONS AND IMPLEMENTATION OPPORTUNITIES OF ARTIFICIAL INTELLIGENCE AND COMPUTER VISION TECHNOLOGIES IN AUTOMATING METROPOLITAN ESCALATORS.

This study examines real-time monitoring and automatic management of passenger flow on metropolitan escalators using artificial intelligence and computer vision technologies. The aim is to develop the conceptual foundations of an automatic hazard detection system based on deep learning algorithms (CNN, YOLO, transformer). Systematic literature analysis based on PRISMA principles, comparative analysis, and conceptual modeling were employed, with 25 international sources analyzed. A three-layered system architecture integrating Edge AI and Fog Computing is proposed. The daily passenger flow of Tashkent Metropolitan exceeding one million in 2025 confirms the urgency of implementation.