Quantitative Stress Testing Using Scalable Digital Twin Simulation with MobileX Pole for Intelligent Mobile Surveillance

In future smart cities, ensuring urban safety requires data-driven decision-making through real-time monitoring tailored to dynamic, complex environments. Such surveillance relies on diverse mobile sensor devices, including drones, robots, patrol vehicles, and portable sensors. However, scaling and validating these systems directly in the real world is constrained by high costs, safety risks, and limited reproducibility across operating conditions. A scalable Digital Twin (DT) model can overcome these constraints by reproducing real-world mobile surveillance in a virtual environment, enabling large-scale simulations of sensor deployment, communication scenarios, and high-density visual data processing. Nevertheless, digital twins still face well-known limitations such as the reality gap, construction costs, limited coverage of behavioral and social variables, biased learning in AI models, and the need for continuous updates. Many of these issues are expected to be mitigated in the near future as generative AI increasingly automates the construction of virtual environments and objects. Despite these advancements, the systemic resource constraints of integrating large-scale physical sensor streams with virtual rendering remain underexplored. To address this gap, this paper proposes a scalable DT framework for the quantitative stress testing of intelligent mobile surveillance systems. The proposed framework collects real-world visualization data from multiple cameras mounted on MobileX Poles, and supports quantitative stress testing in both virtual and physical environments. It systematically analyzes how computing resource usage varies with the number of smart poles and the total number of camera streams under rendering conditions, thereby quantifying the resource limits of real-world, multi-camera DT simulations.