基于分步控制算法的多AGV路径规划  

Path Planning of Multiple AGVs Based on Step Control Algorithm

作　　者：李惠光[1] 贾建成[1] 冷春辉[1] 

机构地区：[1]燕山大学工业计算机控制工程河北省重点实验室,河北省秦皇岛066004

出　　处：《控制工程》2010年第S2期93-96,共4页Control Engineering of China

基　　金：河北省自然科学基金资助项目(F2008000860)

摘　　要：在自动化车间中多AGV系统路径规划是影响工作效率的关键因素之一,针对其中存在的NP问题和多运输任务情况下系统效率降低的问题,在两阶段控制策略的基础上提出了一种基于多参数时间窗和路径表最优搜索策略的分步控制算法。离线阶段使用改进的Dijkstra删除算法生成由多个路径表组成的路径库,并加载在各个下位机上。在线阶段使用分步控制算法将多AGV同时规划的NP问题转变为对单个AGV的分步调度,对每个工作AGV的最优路径和最小运行时间进行近似求解,降低了在线运算量,提高了系统的实时性。通过实例分析证明,在多任务和多工作AGV的情况下,分步控制算法仍然可以保持系统稳定性和工作效率。The work efficiency of automatic workshop is related to multiple AGVs path planning.To solve the problems of NP and systematic efficiency reduced in the system,a step control algorithm based on two-stage traffic control scheme is presented for the AGV system. In the off-line stage,the improved Dijkstra deletion algorithm generates the path base.It is composed of several path-tables and loaded in lower machine.In the on-line stage,the step control algorithm based on multi-parameter time window and path table optimal search strategy changes the NP simultaneous problem of multiple AGV paths planning into the problem of single AGV step scheduling, which generates the approximate solutionre of optimal path and minimum run time.A case study proves that this algorithm retains the stability and work efficiency of multi-AGV and multitasking system.

关 键 词：多AGV系统 路径规划 两阶段控制策略 分步控制算法 

分 类 号：TP273[自动化与计算机技术—检测技术与自动化装置]