| 一种基于改进蚁群算法的载人潜水器全局路径规划 |
| |
| 引用本文： | 史先鹏,解方宇.一种基于改进蚁群算法的载人潜水器全局路径规划[J].海洋技术,2019,38(2). |
| |
| 作者姓名： | 史先鹏 解方宇 |
| |
| 作者单位： | 国家深海基地管理中心,山东青岛266237;杭州电子科技大学自动化学院,浙江杭州310018;杭州电子科技大学自动化学院,浙江杭州,310018 |
| |
| 基金项目： | 国家自然科学基金;国家重点研发计划 |
| |
| 摘 要： | 当前关于使用蚁群算法解决载人潜水器路径规划问题的研究,往往只注重路径的长度和算法收敛速度,容易忽略路径点与障碍物之间的距离和路径的平滑度等要素。载人潜水器过于靠近障碍物航行时容易产生碰撞;按照不平滑路径行驶时,频繁地转向会降低航行效率。为解决这些问题,受人工势场法启发,文中在蚁群算法的概率选择环节引入障碍物惩罚因子φ和转向惩罚因子ψ,对路径点的选择加以限制。仿真测试表明,相比于传统蚁群算法和Dijkstra算法,该算法规划的路径与障碍物之间保持安全距离且转向次数更少,因此载人潜水器按照此路径航行时,安全性和航行效率更高。
|
| 关 键 词： | 载人潜水器 路径规划 蚁群算法 人工势场法 |
A Global Path Planning of Manned Submersible Based on Improved Ant Colony Algorithm |
| |
| SHI Xian-Peng and XIE Fang-Yu.A Global Path Planning of Manned Submersible Based on Improved Ant Colony Algorithm[J].Ocean Technology,2019,38(2). |
| |
| Authors: | SHI Xian-Peng and XIE Fang-Yu |
| |
| Abstract: | Current research on ant colony algorithm (ACO) in solving the path planning problem of manned submersible often pays attention to the length of the path and the convergence speed of the algorithm, but easily ignores the distance between the path points and obstacles and the smoothness of the path. When a manned submersible is too close to an obstacle to navigate, it is prone to collision; when driving on an unsmooth path, frequent steering will reduce navigation efficiency. In order to solve these problems, this paper combines the artificial potential field method to introduce the obstacle penalty factor and the steering penalty factor in the probability selection of the ant colony algorithm to limit the choice of path points. Simulation tests show that compared with Dijkstra algorithm and traditional ant colony algorithm, the proposed algorithm is far away from obstacles and has fewer turns. Manned submersibles are safer and more efficient when sailing on this route. |
| |
| Keywords: | manned submersible path planning ant colony optimization algorithm artificial potential field method |
| 本文献已被 CNKI 万方数据 等数据库收录！ |
| 点击此处可从《海洋技术》浏览原始摘要信息 |
| 点击此处可从《海洋技术》下载免费的PDF全文 |
|
