Urban Air Mobility Risk Assessment and Safety Control over Large-Scale Public Events: A City Marathon Case Study

With the rapid growth of the low-altitude economy, ensuring safe unmanned aerial vehicle (UAV) operations over large public events has become a critical issue for urban air mobility. This study proposes a dynamic risk identification and mitigation framework that integrates UAV inherent risk, aerial traffic density, and ground crowd density into a risk evaluation model. To address the absence of real urban air-route data, a simulated low-altitude network was constructed using ArcGIS, K-means clustering, and Delaunay triangulation, while flight paths were optimized through the ripple-spreading algorithm. Based on this model, a risk-aware control mechanism combining rerouting and hovering strategies was implemented to adaptively respond to varying ground risk levels. A total of 412 UAV missions were simulated over a 6.5 h city marathon scenario, followed by an extended evaluation with 1873 missions to assess scalability. The results show that over 20% of UAVs required detouring or hovering under dynamic risk conditions, leading to a 35–50% reduction in high-risk exposure time while maintaining acceptable operational efficiency. The proposed framework demonstrates good adaptability and scalability for risk-aware UAV operations in complex urban environments.