Abstract: Under the background of the rapid growth of semiconductor production demand and intelligent manufacturing,the path planning and scheduling problem of Automated Guided Vehicles (AGVs) was studied in the semi-conductor workshop with the consideration of collision avoidance.On this basis,a two-stage algorithm was proposed.An exact solution based on mathematical model and a task assignment algorithm based on task urgency were used to obtain the tasks carrying sequence of semiconductors transferred by AGVs.Then,the initial path was generated according to the walking rules of AGVs and the tasks carrying sequence.A collision detection and avoidance algorithm was further designed to obtain the global AGV conflict-free path planning.A sensitivity analysis for AGV number was done to suggest the optimal number of AGVs in semiconductor intelligent workshop.The experimental results showed that the proposed algorithm could solve the handling of 480 tasks on 4 shelves and between 4 machines.The algorithm was within reasonable time and showed good practicability,which could provide theoretical basis and practical reference for realizing multi-AGVs collaborative scheduling of AGV material handling system in intelligent semiconductor workshop.

Key words: intelligent manufacturing, automated guided vehicle, task scheduling, path planning, collision avoidance

摘要： 在半导体市场需求高速增长及智能制造的背景下,对半导体生产车间中采用自动化物料运输系统的自动导引小车(AGV)路径规划与调度问题进行研究,该问题需同时考虑每台AGV的路径规划及AGV群组间的碰撞避免。基于此提出了两阶段算法,首先分别采用基于数学模型求精确解和基于时间优先级的任务分配算法得到AGV运送半导体工件的任务序列,并根据行走规则生成每台AGV的初始路径。然后设计碰撞检测及避免算法对初始路径中可能发生的碰撞点进行检测,并采取相应的避撞策略,从而获得全局AGV无冲突路径。最后对AGV数量进行灵敏度分析,给出半导体生产车间最优的AGV数量配置建议。实验结果表明,所提算法能够调度AGV在具有4个货架和4个机台的半导体车间中完成480个工件运送任务,并且在算法效率和求解时间上均具有很好的实用性。研究结论可为半导体生产车间采用AGV自动化物料运输系统实现多AGV协同调度提供理论依据和实际参考。

关键词: 智能制造, 自动导引小车, 任务调度, 路径规划, 碰撞避免