基于动态多种群自定义变种粒子群算法的无人机探索路径规划  

Unmanned aerial vehicle exploration path planning based on dynamic multi-swarm customized variant particle swarm optimization

作　　者：蒋文彬 杨忠[1] 卓浩泽 廖禄伟 朱泽堃 王先坦 JIANG Wenbin;YANG Zhong;ZHUO Haoze;LIAO Luwei;ZHU Zekun;WANG Xiantan(College of Automation Engineering,Nanjing University of Aeronautics and Astronautics,Nanjing 211106,China;Electric Power Science Research Institute of Guangxi Power Grid Co.,Ltd,Guangxi 530023,China)

机构地区：[1]南京航空航天大学自动化学院,江苏南京211106 [2]广西电网有限责任公司电力科学研究院,广西南宁530023

出　　处：《应用科技》2024年第5期263-271,共9页Applied Science and Technology

基　　金：广西电网公司2023年科技创新专业科技项目(GXKJXM20230169);贵州省科技计划项目(黔科合支撑[2020]2Y044号)。

摘　　要：针对传统粒子群算法(particle swarm optimization,PSO)存在的粒子更新思路单一、随机性受限、收敛速度慢和易陷入局部最优等问题,提出一种动态多种群自定义变种粒子群算法(dynamic multi-swarm customized variant particle swarm optimization,DMCVPSO),旨在更高效地完成复杂未知环境下的无人机(unmanned aerial vehicle,UAV)快速探索任务。首先,在综合考虑诸多限制因素后构建聚合适应度函数。其次,该算法根据每代粒子的适应度动态划分多种群,针对各子种群的特点引入不同的并行更新策略,引入莱维飞行、贪婪策略有利于优势群进行更加细密有效的搜索;采用概率性混合变异的策略降低劣势群探索的盲目性;融合余弦函数和自适应策略用于平衡混合群的局部开采和全局勘探能力。最后,通过数值仿真和三维可视仿真平台对该算法进行可行性验证。结果表明,所提出的优化算法有助于解决收敛速度过慢、陷入局部最优等问题,提高无人机在复杂未知环境中的探索效率。To address the issues in traditional Particle Swarm Optimization(PSO),such as single particle update methodology,limited randomness,slow convergence speed,and a tendency to get trapped in local optima,this paper proposes a dynamic multi-swarm customized variant particle swarm optimization(DMCVPSO)algorithm.The aim is to efficiently accomplish rapid exploration tasks for unmanned aerial vehicle(UAV)in complex unknown environments.First,an aggregated fitness function is constructed after considering various constraints.Then,the algorithm dynamically divides the population into multiple swarms based on the fitness of particles in each generation.Different parallel update strategies are introduced for each sub-population,Lévy flight and greedy strategy enhance fine-grained and effective searches for the dominant swarm,while a probabilistic hybrid mutation strategy reduces the exploration randomness of the inferior swarm.Furthermore,the cosine function and adaptive strategy are employed to balance local exploitation and global exploration capabilities of the mixed swarm.Finally,the feasibility of this algorithm is validated through numerical simulations and a 3D visual simulation platform.Results show that the proposed optimization algorithm facilitates to solve problems such as slow convergence and local optimum trapping,significantly improving UAV exploration efficiency in complex and unknown environments.

关 键 词：路径规划 多旋翼无人机 粒子群优化 自主探索 聚合适应度 动态多种群 自定义变种 参数自适应 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]