融合多策略的蜣螂算法的三维无人机航迹规划
Three-dimensional UAV path planning based on Dung Beetle Algorithm fused with multiple strategies
针对蜣螂算法在无人机航迹规划中存在的种群多样性匮乏、收敛迟缓和全局探索力不足等问题,提出了一种融合多策略的更优蜣螂优化算法(BDBO).该算法将动态学习的佳点集初始化、基于时间选择的主动并行双边搜索、自然衰减种群机制及融合Sigmoid的边界收敛策略嵌入航迹规划全过程:在全局阶段迅速扩大可行航迹空间,在局部阶段精细调整航迹最优点,从而同步提升收敛速度与求解精度.以航迹长度、飞行安全、平滑度及高度代价的加权和为目标函数,在CEC2017及三维山地环境进行验证.结果表明:BDBO 所得航迹的适应度值较对比的四种算法平均降低19.47%、25.51%、20.85%和3.97%,充分证明了BDBO在无人机三维航迹规划中的有效性与优越性.
To address the deficiencies of the Dung Beetle Optimizer (DBO) in UAV path planning—namely, insufficient population diversity, slow convergence, and weak global exploration, a multi-strategy-enhanced variant termed BDBO is proposed. The algorithm seamlessly integrates four synergistic components throughout the entire planning process: dynamic-learning-based good-point set initialization, time-triggered active parallel bilateral search, a natural population-decay mechanism, and a boundary convergence strategy fused with a Sigmoid function. Collectively, these strategies rapidly expand the feasible flight-corridor during the global phase and refine waypoint positions in the local phase, thereby accelerating convergence and improving solution accuracy simultaneously. A composite objective function aggregating path length, flight safety, trajectory smoothness, and altitude cost is minimized on both CEC2017 benchmarks and a realistic 3-D mountainous scenario. Experimental results reveal that BDBO reduces the best fitness value by 19.47 %, 25.51 %, 20.85 %, and 3.97 % compared with four state-of-the-art counterparts, unequivocally demonstrating its effectiveness and superiority for three-dimensional UAV path planning.
| [1] |
于少猛, 闫铭, 王鹏飞, |
| [2] |
唐东旭. 无人机技术在森林生态遥感监测中的应用与探讨[J]. 智慧中国, 2024(12): 80-81. |
| [3] |
王芸. 基于模拟退火算法的末端车载无人机物流配送路径规划研究[J]. 自动化与仪器仪表, 2025(2): 247-251. |
| [4] |
汪国安, 王红军, 马宁, |
| [5] |
宋俊辉, 刘宇庭, 郭世杰. 动态环境下AIP-RRT*与DGF-APF融合的机器人路径规划[J]. 仪器仪表学报, 2025, 46(3): 51-64. |
| [6] |
孙波, 周健康, 赵玉清, |
| [7] |
毛文平, 李帅永, 谢现乐, |
| [8] |
|
| [9] |
蒋翱徽, 刘文红. 基于改进蜣螂优化算法的无人机三维路径规划[J]. 电子测量技术, 2024, 47(13): 128-135. |
| [10] |
陈琦, 王亚杰, 孙云飞, |
| [11] |
李为瀚, 刘媛华. 融合多策略改进的蜣螂优化算法[J/OL]. 重庆工商大学学报(自然科学版), 2025: 1-10. (2025-02-24). |
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
李书群, 陈钰, 杨雨婷, |
| [16] |
郝武帮, 施磊. 融合自然选择机理和遗传算法的多峰MPPT优化控制[J]. 舰船电子工程, 2024, 44(11): 186-191, 203. |
| [17] |
李俊毅, 邓志祥. 双曲正切函数的新型变步长LMS算法[J]. 计算机与数字工程, 2024, 52(8): 2272-2278. |
| [18] |
毛清华, 赵冰, 李阳. 融合多项式差分学习与逐维变异的混沌蜉蝣算法[J/OL]. 北京航空航天大学学报, 2024: 1-15. (2024-06-28). |
| [19] |
杨桂芹, 刘志琦, 张国庆, |
| [20] |
朱凯鹏, 王全政, 杨文政, |
中国高校产学研创新基金资助项目(2021ZYA04004)
/
| 〈 |
|
〉 |