3D激光导航无人叉车多车路径规划方法  

Multi-vehicle path planning method for unmanned forklift guided by 3D laser

作　　者：黄国胜 孙振勇 姜喆华 孟令广 伏松平 Huang Guosheng;Sun Zhenyong;Jiang Zhehua;Meng Lingguang;Fu Songping

机构地区：[1]中国铁建电气化局集团有限公司,北京100043 [2]北京起重运输机械设计研究院有限公司,北京100007

出　　处：《起重运输机械》2025年第7期65-69,共5页Hoisting and Conveying Machinery

摘　　要：文中通过阐释3D激光导航无人叉车的基本概念、应用场景、SLAM导航方式、常见冲突类型以及数种经典的多车路径规划算法,结合使用场景,对遗传算法、蚁群算法和深度强化学习算法原理及优缺点进行对比分析,选定蚁群算法结合深度强化学习作为其路径规划方法。同时,设计了基于冲突预测的路径调整策略,有效避免了车辆之间的碰撞和死锁问题。最后,通过实际项目应用验证了所提方法的有效性和可行性,为3D激光导航无人叉车的多车路径规划提供了理论依据和技术支持,对推动无人叉车在自动化物流系统中的广泛应用具有意义。The basic concept,application scenario,SLAM navigation mode,common conflict types and several classic multi-vehicle path planning algorithms of 3D laser navigation unmanned forklift are explained.Combined with the application scenario,the principles,advantages and disadvantages of genetic algorithm,ant colony algorithm and deep reinforcement learning algorithm are compared,and ant colony algorithm combined with deep reinforcement is selected for path planning.At the same time,a path adjustment strategy based on conflict prediction is proposed,which effectively avoids collision and deadlock between vehicles.Finally,the effectiveness and feasibility of the proposed method are verified by application.This study provides theoretical basis and technical support for multi-vehicle path planning of 3D laser navigation unmanned forklift,and is of great significance to promote the wide application of unmanned forklift in automated logistics system.

关 键 词：无人叉车 路径规划 冲突类型 蚁群算法 3D激光导航 

分 类 号：TH242[机械工程—机械制造及自动化]