Authors

Abstract

Both cruising ability and safety should be considered in the 3D inspection path planning of agricultural unmanned aerial vehicles (UAVs). Specific to a complex working environment, the 3D inspection environment of agricultural UAVs was simulated through terrain modeling and threat modeling. First, the dynamic constraints of flight approaching rate and response time were added to the threat cost, and the 3D mission space model and flight path cost function were constructed considering the influence of UAVs’ turning performance. Second, the offset estimation strategy, variable spiral search strategy, quasi-reverse learning strategy and dimension-by-dimension mutation strategy were introduced into the dung beetle optimizer (DBO) algorithm to improve the global optimization ability and convergence rate of the algorithm. By establishing a three-dimensional trajectory planning model for unmanned aerial vehicles, the trajectory planning is transformed into a multi-objective function optimization problem, and an improved algorithm is used to solve the three-dimensional trajectory planning of unmanned aerial vehicles. The fitness is evaluated by considering the objective function of trajectory cost, terrain cost, and danger level, and the trajectory planning is iteratively optimized. The results indicate that the proposed improved dung beetle algorithm for trajectory planning has lower overall cost and stability in adapting to different complex terrain environments.

Abstract in Chinese

农业无人机的三维巡检路径规划不仅要考虑续航能力，还要考虑安全性。针对复杂的作业环境，通过地形建模和威胁建模来模拟农业无人机的三维巡检环境。首先，将飞行接近率和响应时间的动态约束添加到威胁成本代价中，并考虑无人机转弯性能的影响，建立三维任务空间模型与航迹代价函数；其次，在蜣螂算法中引入偏移估计策略、变螺旋搜索策略、准反向学习策略和逐维变异策略，提高算法的全局寻优能力和收敛速度。通过建立无人机三维航迹规划模型，将航迹规划转化为多目标函数优化问题，并利用改进算法求解无人机三维航迹规划，以综合考虑航迹代价、地形代价和危险程度的目标函数评估适应度，对航迹规划迭代寻优。结果表明，所提改进蜣螂算法规划的航迹具有更低的总代价和适应不同复杂地形环境的稳定性。