Improving three-dimensional security for surveillance systems in smart cities using graph neural network model, Barcelona and Singapore as examples

This research investigates the implementation and enhancement of Three-Dimensional (3D) Security and surveillance system within the context of smart cities, by speeding up the defendable security actions and threat detection. These systems are integral components for achieving comprehensive urban safety through integrating spatial data analysis with other city infrastructures. A major challenge in smart city development is ensuring real-time, accurate threat detection and coordinated emergency responses. Existing surveillance systems lack the multidimensional capabilities and integration flexibility needed to adapt to complex urban environments. The study proposed upgrading surveillance technologies by incorporating a Graph Neural Network (GNN) algorithm into traffic monitoring system. Simulations demonstrated that GNN-enhanced system achieved higher prediction accuracy (1–3%), improved inference speed (5 to 10 km/h), and reduce error rates (1-3%) compared to traditional methods. The improved performance is attributed to the ability of GNNs to effectively model relationships within spatial and traffic network data, allowing for more accurate and context-aware decision-making process (speeding up to 60%) in real time. The enhanced system offers data processing using 3D spatial analysis, real-time threat detection, integration with urban systems like traffic control, and learning adaptability via GNNs – together addressing delays and inefficiencies of earlier methods (in Barcelona from 45 minutes to 30 minutes and in Singapore from 30 to 20 minutes). The results are applicable in urban surveillance areas with established digital infrastructure, where spatial data, system integration (e.g. traffic, emergency systems), and computational resources are available. These systems are especially valuable in large areas, critical zones, and event-based surveillance needs, where proactive and precise security management is essential. Defendable security actions speed up to 60% by faster predictions and decision-making process