A Risk-aware Spatio-temporal Scheduler for Large-scale UAV Operations

Spatio-temporal scheduling of Unmanned Aerial Vehicles (UAV) is a major challenge for the UAV Traffic Man- agement (UTM) system due to dense traffic and high mobility. To ensure safety and efficiency, UAV schedulers assign both geographic paths and operational time windows to prevent conflicts and minimize delays. However, scheduling also influences UAV operational risks by determining the potential exposure of UAV failures to third parties on the ground. Since operational risk has not been systematically considered in existing schedulers, this work integrates systematic risk assessment into a hybrid spatio-temporal scheduling framework that combines greedy search with Mixed-Integer Linear Programming. The proposed scheduler is evaluated in large-scale simulations of 5,000, 10,000, and 15,000 flights over central London during 24 hours. Results show that the risk-aware solver successfully schedules 13.24%- 30.70% more flights while reducing operational risk by 85.71%- 96.67% compared to the benchmark. These improvements be- come more pronounced at larger scales, demonstrating the scalability and potential of the proposed approach for real-world UTM deployment.