A Generative City Digital Twin System for Flooding Emergency Management in Smart City

Emergency Management (EM) is a critical component in maintaining life quality and enhancing public safety in smart cities. City EM is now often triggered by extreme weather conditions, particularly urban flooding resulting from Localized Heavy Rain (LHR), as it requires timely multi-service collaboration, planning and decision-making among all stakeholders, this in turn requires the systematic support of multimodal information model fusion, timely visual situation awareness, edge-cloud collaborative analysis and virtual-real interactive control. The challenge is to provide the above requirements and collaborate multi-domain real world system with its multimodal virtual system in a bidirectional platform. Digital twin (DT), serving as a digital platform architecture, can enable dynamic synchronization, visualization, and virtual-real interaction across the full life cycle of physical space in virtual space. Building upon this concept, we developed an easy-to-use generative system for 'select area of interest - CFDT'. That is, a dynamic flooding information model-driven, auto-generative City Flooding Digital Twin (CFDT) system that supports near-real-time visual situational awareness, XR-based interaction, and smart decision-making. The developed prototype has undergone demonstration and tested at city, regional and street levels across different countries to assess its interpretability, usability, timely ability, and performance quantitatively and qualitatively. The results show that it is efficacy as a valuable tool for various CDT developments, providing a service fusion/integration platform across multiple domains. This work emphasizes the opportunities that CDT brings to smart city and EM.