Authors

Abstract

Multi-unmanned aerial vehicle (UAV) collaborative task planning and distribution path planning are the core content of agricultural UAV logistics distribution. In this study, the multi-UAV collaborative task planning and the distribution path planning were discussed, and such constraint conditions as UAV load capacity, battery capacity and flight time were comprehensively considered, aiming to reduce the number of UAVs and their power consumption. To ensure the safe and efficient completion of multi-UAV logistics distribution tasks, 3D agricultural ultralow space was subjected to environment modeling, and a bilevel planning model for collaborative planning of UAV distribution route and flight path was constructed. Then, an improved particle swarm optimization (PSO) algorithm with the improved learning factor and inertia coefficient was designed on the basis of PSO framework, and the global optimal solution in the current iteration was improved using variable neighborhood descent search. The feasibility of the proposed algorithm was verified by analyzing a practical case. With the central city area of XX City as the study area, 1 logistics & freight transportation center was taken as the central warehouse (coordinates: 50, 50, unit: km) and 50 intelligent express cabinets as the express cabinets of UAVs. The obtained results were comparatively analyzed with those acquired through the basic PSO algorithm. The results manifest that the proposed algorithm performs better than the compared algorithms. The improved PSO algorithm is superior to the basic PSO algorithm in aspects of total UAV flight distance, number of UAVs used and algorithm convergence time, indicating that the model and algorithm established in this study are feasible and effective.

Abstract in Chinese

多无人机任务协同规划与配送路径规划是农业无人机物流配送的核心内容，探讨多无人机任务协同规划与配送路径规划的农业无人机群配送路径的规划问题, 综合考虑无人机载重量、无人机电池容量、无人机飞行时间等约束条件, 目标为降低无人机数量及耗电量。为保障安全、高效完成多农业无人机物流配送任务，首先对三维农业超低空间进行环境建模，构建了一种无人机配送线路以及航迹协同规划的双层规划模型。基于粒子群算法框架设计了一种改进学习因子与惯性系数的改进粒子群算法, 利用变邻域下降搜索对当前迭代中的全局最优解进行改进。通过实际案例分析, 验证了该算法的可行性，以XX市中心城区为研究区域。选取1个物流货运中心作为中心仓库位置坐标(50,50),单位(km)、50个智能快递柜作为无人机快递柜位置。将所得结果与基础粒子群算法进行对比分析，结果表明, 本算法性能优于对比算法, 改进PSO得到的无人机飞行总距离，启用无人机数量和算法收敛时间等方面均优于基础粒子群算法，说明本文构建的模型与算法是可行的和有效的。