Multi-source data-driven intelligent analysis and decision optimization for high-density pedestrian flows in urban public spaces

Abstract

Managing high-density pedestrian flows in urban public spaces via Information Technologies (IT) is crucial for safety and efficiency. Despite advancements in sensing, AI-driven prediction, and control, a critical gap persists: lacking the systematic integration needed for robust automated crowd management systems, an issue intensified by AI/IoT growth. To address this challenge, a comprehensive review of the literature from 2014 to 2024 has been conducted, analyzing and synthesizing IT-driven decision support approaches for automated crowd management. The field is organized around three core technological pillars: (1) multi-source data fusion architectures for comprehensive real-time monitoring; (2) intelligent prediction systems using deep learning for accurate forecasting and anomaly detection; and (3) advanced decision optimization platforms enabling dynamic, multi-objective control strategies. In addition, the review explores key emerging trends such as edge computing, digital twins, and human-machine collaboration. The findings offer theoretical insights, practical guidelines, an overview of persistent challenges, and strategic directions for future research in intelligent crowd management within the broader context of smart cities and resilient infrastructure.