Modelling of dynamic coverage process by quadrotor swarms for sudden abnormal crowd evacuation

This study addresses the challenge of achieving real-time and adaptive aerial coverage over crowds during sudden abnormal evacuations -a highly nonlinear and dynamic process.This paper proposes a physics-aware hybrid modelling framework that integrates a microscopic social force model for simulating crowd movements with an enhanced artificial potential field approach for quadrotor swarm guidance.Central to our method is a nonlinear function that maps real-time crowd density into a spatially varying potential field, along with a dynamic balancing strategy to avoid local optima.Extensive simulations based on real pedestrian trajectory datasets show that the proposed framework achieves an average coverage efficiency of 82.4% and a recovery latency of only 4.3 seconds following evacuation onset, outperforming state-of-the-art methods in coverage persistence, responsiveness, and energy efficiency.These results validate the framework's capability to support adaptive monitoring in emergency crowd management scenarios.