集装箱码头桥吊作业的强化学习优化  

Reinforcement Learning Optimization of Bridge Craner Opearations at Container Terminals

作　　者：解静修 吴一亮[1] XIE Jingxiu;WU Yiliang(School of Ocean Information Egineering,Jimei University,Xiamen 361021,China)

机构地区：[1]集美大学海洋信息工程学院,福建厦门361021

出　　处：《集美大学学报（自然科学版）》2024年第4期329-335,共7页Journal of Jimei University：Natural Science

基　　金：福建省科技厅重大专项“集装箱码头装备群体智能协同作业关键技术研发及产业化”(2022HZ 022019);福建省自然科学基金项目“沉浸式视频流的低时延视频编码技术研究”(2021J01868)。

摘　　要：对码头桥吊装卸操作进行强化学习优化,通过Unity建立码头桥吊的模型及训练环境,用PPO算法对码头桥吊进行装卸操作训练,最终使桥吊学习在海浪及大风对集装箱船和吊具造成晃动影响时的平稳下落行为,提高桥吊抓取准确性并优化桥吊路径,实现更高效智能的码头装卸作业。In this paper,reinforcement learning was used to optimize the loading and unloading operations of quay bridge cranes,a model and training environment was built for cranes through Unity,and PPO algorithm was used to train the loading and unloading operations of cranes,so that the cranes can learn to lower smoothly when waves and high winds affect the container ship and lifting gears.Eventually this can improve the accuracy of crane grasping and optimize the path of cranes to achieve more efficient and intelligent loading and unloading operations at terminals.

关 键 词：码头桥吊 UNITY 强化学习 集装箱装卸作业 

分 类 号：TP302.1[自动化与计算机技术—计算机系统结构]