Planning method for three-dimensional air route network of urban low-altitude logistics drones

To regulate the large-scale operational order of urban low-altitude logistics drones, a planning method for the three-dimensional air route network of urban low-altitude logistics drones was proposed. Based on the urban low-altitude layered airspace structure, the grid method was adopted for the discretization modeling of urban low-altitude airspace. Combined with the "one-network, two-layer, and three-node" route network architecture, a transfer node location model and a route network planning model were established. An algorithm framework based on a self-organizing map neural network and multi-objective simulated annealing was developed to optimize the spatial layout of transfer nodes and network topology. A simulation experiment was conducted in a certain area of Nanjing City. The results show that when the number of demand nodes is fixed, the average node degree of the transfer network is greater than that of the delivery network due to the cargo circulation function undertaken by the transfer layer. Compared with the genetic algorithm and grey wolf optimization algorithm, the proposed multi-objective simulated annealing algorithm can better balance the relationship among objectives, and the comprehensive score increases by over 17%. As the scale of demand nodes expands, the number of transfer nodes shows an increasing trend, which activates more optional route nodes and expands the dimension of drone route selection. Compared with the single-layer network, the proposed route network reduces the non-linear coefficient by 14.09% and the average route flow by 52.43%, which can effectively disperse route loads. The proposed method enables route network planning in scenarios with dense user demands and improves the operational feasibility of logistics drones in complex urban environments.