Dynamic Takeoff and Landing Control for Multi-Rotor eVTOL Aircraft

Urban air mobility (UAM) has gained significant attention due to technological advancements. To reflect the flight characteristics of electric vertical takeoff and landing (eVTOL), a new takeoff and landing control strategy is needed. On one hand, we introduce a novel approach control model termed the air apron-based approach (ABA). On the other hand, we propose a multi-objective optimization model for dynamic collaborative sequencing of arrival and departure flights. By employing the deep Q-network (DQN) algorithm, we develop an optimized data structure and a suitable reward mechanism. Simulation results indicate that, compared to other approach models, the ABA improves spatial utilization by 6.6%. Additionally, ABA demonstrates greater efficiency when simultaneous eVTOL arrivals exceed 16. Data for 300 flights were randomly generated and three methods were used for flight sequencing, including DQN algorithm, genetic algorithm (GA), and first come, first served (FCFS). Simulation has shown that compared to the FCFS, the dynamic collaborative sequencing strategy performs better in hovering time, delay time, and the number of empty flights. Compared to the GA, the DQN algorithm exhibits stronger competitiveness when service capacity exceeds 40% of vertiport. The dynamic takeoff and landing control method proposed in this study can optimize the operational efficiency of UAM.