Authors

Abstract

To address the contradiction between obstacle avoidance safety and irrigation operation quality in 3D path planning of fixed-wing UAVs for mountainous agricultural crop irrigation, this paper proposes a multi-strategy improved dung beetle optimization (MDBO) algorithm. First, safe turning radius and pitch angle constraints are embedded into the optimization objective to restrict flight attitude, ensure flight stability under complex mountain terrain, reduce collision risk, and maintain consistent spray deposition. Second, four improvement strategies, including Bernoulli chaotic mapping initialization, golden sine position update mechanism, dynamic adaptive weight, and adaptive Gaussian-Cauchy hybrid mutation perturbation, are integrated into the basic DBO algorithm to enhance global search capability, local exploitation accuracy and convergence performance. Comparative tests on 9 benchmark functions show that MDBO outperforms the basic DBO in convergence speed and solution stability. Furthermore, three irrigation performance indicators (coverage rate, overlap ratio and coverage uniformity) are constructed for posterior evaluation of spray distribution quality. Simulation experiments under three mountainous terrain scenarios verify that MDBO achieves better path accuracy, convergence performance and irrigation effect than traditional algorithms, and can meet the efficient irrigation path requirements of mountainous agricultural crops.

Abstract in Chinese

为解决固定翼无人机在山区农业作物灌溉三维路径规划中避障安全性与灌溉作业质量之间的矛盾，本文提出一种多策略改进的蜣螂优化（MDBO）算法。首先，将安全转弯半径和俯仰角约束嵌入优化目标函数中，以限制飞行姿态，确保复杂山地地形下的飞行稳定性，降低碰撞风险，并保持喷洒沉积的一致性。其次，将伯努利混沌映射初始化、黄金正弦位置更新机制、动态自适应权重以及自适应高斯-柯西混合变异扰动等四种改进策略融入基本DBO算法，以提升全局搜索能力、局部开发精度和收敛性能。在9个基准函数上的对比实验表明，MDBO在收敛速度和解的稳定性方面优于基本DBO算法。此外，构建了三个灌溉性能指标（覆盖率、重叠率和覆盖均匀性），用于对喷洒分布质量进行后评估。在三种山地地形场景下的仿真实验验证了MDBO算法在路径精度、收敛性能和灌溉效果方面优于传统算法，能够满足山地农作物高效灌溉路径的需求。