基于多AGV的智能仓储调度系统研发

发布时间：2018-04-02 10:23

本文选题：自动导航车　切入点：调度系统　出处：《浙江大学》2017年硕士论文

【摘要】：随着土地资源紧缺和人力成本的急剧上升,仓储物流业的发展面临严重制约,因此设计适用于多机器人多目标的自动化智能仓库系统,实现企业由自动化向全智能转型,成为研究热点。本文重点研究智能仓储调度系统。从实际应场景用出发,分析多AGV智能仓库系统运转低效且拥堵死锁状况频发的成因,从仓库系统建模和调度控制策略两方面进行研究设计。首先对系统运作核心单元AGV进行关键技术研究,实现二维码定位和混合导航的框架设计。随后对立体仓库进行建模,将连续动态多目标仓库模型栅格化,转化为离散的多目标面运动控制问题,极大的减少了系统的状态空间规模和数据传输量。同时采用区域划分实现仓库模型中资源的合理配置,利用独占区为仓库复杂冲突路段建立监控,以便于控制中心后续调度。然后确定了多AGV智能仓储物流调度系统的最终解决方案。系统初始化时,控制中心调用路径算法生成区域路径哈希表并分离存储,同时通过实现跨区域跨楼层路径规划,将多楼层物流仓库从三维空间降维至二维平面。对原始任务按照业务流程进行拆分整合,通过改进的市场竞争机制算法下进行任务分配,实现小车任务执行的灵活性和高效性。控制中心通过调度算法对独占区进行监测,实现多AGV并行无碰撞作业。本系统通过无线局域网实现控制中心与AGVs之间的大容量数据传输,利用XML技术对任务和地图模型数据进行格式化存储和数据传输。最后将该调度系统应用于AGV仿真实验平台,模拟真实场景进行建模、车辆配置和订单动态添加,实验过程中未出现碰撞、死锁和长时间滞留现象,所有AGV均顺利完成系统下发的任务,从而验证了该调度系统的可行性、有效性和鲁棒性。
[Abstract]:With the shortage of land resources and the sharp rise of manpower cost, the development of warehousing logistics industry is facing serious constraints, so the design of automated intelligent warehouse system, which is suitable for multi-robot and multi-target, can realize the transformation from automation to full intelligence. This paper focuses on intelligent warehouse scheduling system. From the perspective of practical scenarios, the causes of inefficient operation and frequent deadlocks of multi-#en0# intelligent warehouse systems are analyzed. The system modeling and scheduling control strategy are studied and designed. Firstly, the key technology of AGV, the core unit of system operation, is studied to realize the frame design of two-dimensional code location and hybrid navigation. Then, the three-dimensional warehouse is modeled. The continuous dynamic multi-objective warehouse model is rasterized and transformed into a discrete multi-objective surface motion control problem. The scale of state space and the amount of data transmission are greatly reduced. At the same time, the rational allocation of resources in the warehouse model is realized by using the regional partition, and the monitoring of the complex conflict section of the warehouse is established by using the exclusive area. In order to facilitate the subsequent scheduling of the control center, the final solution of the multi-#en0# intelligent warehouse logistics scheduling system is determined. When the system is initialized, the control center invokes the path algorithm to generate the regional path hash table and store it separately. At the same time, the multi-floor logistics warehouse is reduced from three-dimensional space to two-dimensional plane through cross-regional cross-floor path planning. The original task is split and integrated according to the business process, and the task allocation is carried out under the improved market competition mechanism algorithm. The control center monitors the exclusive area through the scheduling algorithm and realizes the parallel collision free operation of multiple AGV. The system realizes the large capacity data transmission between the control center and the AGVs through the wireless local area network (WLAN). Finally, the scheduling system is applied to the AGV simulation experiment platform to simulate the real scene modeling, vehicle configuration and order dynamic addition, using XML technology to format and store the task and map model data and data transmission. There is no collision, deadlock and long time detention in the experiment. All the AGV successfully complete the task of the system, which verifies the feasibility, effectiveness and robustness of the scheduling system.
【学位授予单位】：浙江大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TH692.3

【相似文献】