改进粒子群权重的水下机器人路径规划  

Path planning of underwater vehicle based on improved particle swarm weight

作　　者：石博博 朱嘉晟 杨天慈 袁浩宇 郭阳 衣正尧 SHI Bobo;ZHU Jiasheng;YANG Tianci;YUAN Haoyu;GUO Yang;YI Zhengyao(Navigation and Engineering College,Dalian Ocean University,Dalian 116023,China)

机构地区：[1]大连海洋大学航海与船舶工程学院,辽宁大连116023

出　　处：《舰船科学技术》2025年第5期159-165,共7页Ship Science and Technology

基　　金：辽宁省教育厅基本科研资助项目(JYTMS20230472)。

摘　　要：为提高水下机器人路径规划的收敛速度和行驶路径的光滑性,本文提出改进粒子群算法的最优路径规划。建立改进的粒子群算法数学模型,对传统粒子群算法进行3种权重系数改进研究;建立水下环境场景数学模型,模拟水下复杂地形;融合贝塞尔曲线。在迭代10次的最优适应度,收敛速度提高44%;在迭代20次时,收敛速度提高15%;在迭代30次时,收敛速度提高2%。改进算法相比于传统算法,实现前期全局寻优能力强,易得到合适的种子;后期局部寻优能力强,易提高收敛精度;并且在复杂环境下可以更快收敛到最优路径,缩短水下机器人的作业周期。改进的算法可以对水下环境进行全局搜索,且不易陷入最优解,供水下机器人领域参考。In order to improve the convergence speed and smoothness of the path planning of the underwater vehicle,the optimal path planning of the improved particle swarm optimization algorithm is proposed.The mathematical model of the improved particle swarm optimization algorithm is established,and the three weight coefficients of the traditional particle swarm optimization algorithm are improved.The mathematical model of underwater environment scene is established to simulate underwater complex terrain.Blending Bessel Curves.At the optimal fitness of 10 iterations,the convergence speed is increased by 44%.When the iteration is 20 times,the convergence speed is increased by 15%.When the iteration is 30 times,the convergence speed is increased by 2%.Compared with the traditional algorithm,the improved algorithm has strong global optimization ability in the early stage,and it is easy to get suitable seeds.In the later stage,the local optimiza-tion ability is strong,and the convergence accuracy is easy to be improved.In addition,it can converge to the optimal path faster in complex environment and shorten the operation cycle of underwater robot.The improved algorithm can search the underwater environment globally and is not easy to fall into the optimal solution.It can be used as a reference for the field of robots underwater supply.

关 键 词：粒子群算法 地形建立 水下机器人 权重函数 路径优化 

分 类 号：U676[交通运输工程—船舶及航道工程]