Abstract: Nowadays,most of Automated Guided Vehicle (AGV) systems adopt cloud computing-based network framework,which has a poor performance in real time.The tasks requiring high real-time such as obstacle avoidance,path planning and obstacle identification cannot response timely cause of the large time delay.For this reason,a multi-AGV control system including cloud,edge,and device three layers was developed based on edge computing.To reduce the latency of task process and improve real-time performance,the real time tasks were deployed in edge nodes that near the AGVs.Experiments showed that the system based on edge computing reduced the latency by 46.4%~58.8% compared with that based on cloud computing framework when performing the same tasks.

Key words: automated guided vehicle, edge computing, logistics system, Internet of things

摘要： 目前大多数自动导引车控制系统使用云计算框架,实时性差。云端与自动导引小车之间有较大的网络延迟,这使得自动导引小车中如避障、路径规划以及障碍物识别等高实时性任务不能得到及时响应。由此提出一种基于边缘计算的自动导引小车控制系统,具有云服务层、边缘层和设备层3层架构。实时性高的计算任务部署在边缘节点,靠近自动导引小车,计算发生在数据生产者附近,可以减少任务处理延迟时间,提高实时性。实验证明,在执行相同任务时,基于边缘计算的控制系统比基于云计算的控制系统减少46.4%~58.8%的计算延时。

关键词: 自动导引车, 边缘计算, 物流系统, 物联网