摘要
自动导向车(AGV)地面系统是根据接收到的物料搬运任务,合理调度车辆,在智能交通管理控制下,沿规定的路线自动高效地完成物料搬运任务,并及时调度小车自动充电的自动控制系统。针对AGV地面系统应用的复杂性,以往对具体应用项目进行设计的方法,存在系统适用性不强,项目实施周期长,很难形成产品化等问题,为此,本文通过对国内外软件体系结构及进口引进AGV系统的研究,充分利用面向对象技术,建立了一个能满足各种应用要求的AGV地面系统,取得主要研究成果有:
(1)提出了建立在软件平台思想上,将系统按运行平台和开发平台划分,支持AGV系统规划、仿真、控制和诊断集成的AGV整体解决方案的体系结构,以满足各种AGV系统应用需求。设计了包含AGV地面控制系统、地面图形监控系统、车载系统、仿真系统的AGV运行平台和包含系统规划工具、地面程序定义工具、车载程序定义工具的AGV开发平台的功能模型。
(2)提出了基于软件分层技术、模型-视图分离技术、消息总线技术及分布式技术的AGV地面运行平台体系结构。利用软件分层技术将系统分为界面表示层、领域对象层、服务层及数据存储层有利于系统的开发和部署。利用模型-视图分离技术设计了用户界面表示层对象和领域层对象间的交互方式,极大提高了领域对象的可重用性。设计了基于订阅-发布设计模式的消息总线,用于领域层中对象的交互,使系统有了很好的扩展性。设计了基于代理设计模式的分布式模型结构,很好地解决了如何将系统部署在不同计算单元。
(3)使用UML作为建摸语言,设计了各个系统的软件结构,采用类结构设计建立了系统静态结构模型,采用对象交互设计方法建立了系统动态结构模型,详细阐述了任务管理、车辆管理、交通管理、IO管理、通信管理、车辆仿真、路径管理、图形监视等子系统的结构和原理。
(4)提出了基于遗传算法求解的AGV线路调度方法。针对车辆调度的调度效率问题,建立了将任务调度和车辆调度相结合的数学模型,采用遗传算法对模型进行求解,验证了遗传算法是一种有效的AGV线路调度算法。
结合本公司AGV系统的开发实践,结果表明本文对AGV地面系统设计思路及方法是正确的、可靠的,它不仅在AGV系统设计中具有重大的实用价值,而且对设计各类智能搬运设备的调度监控系统也有广泛的应用前景。
A(iV Master system is an auto--contro1 system dimed to di spatch thc ciIrr l crs
acc()rding to the conveying task, to fulfill the task efficiently along the
r()ut e under inte1lectual ized traffic management, and schedule the c%lrri ers
for auto--charging. To cover the shortages as 1ow flexibi l ity, long imp l eInc'n[
tcrfn, di fficult commercial ization, an AGV Master contro1 system thdt ckln be
appl ied in varies condit ion is establ ished adopt ing ()hjccl --()ri en[ L.(l
technoIogy after studying the software frameworks and AGV systems ()1' ()t hur
comPany. Main achievements in this thesi s are as fo1l ows f
(l ) A tota1 solut ion framework is put forward based on s()l'twtlre p1 i1t f'()rm
prl nc ipl e. Thi s frame, which involves perf()rm platform and develop P l ilt I'()rm,
Supr)orts system 1ayout, s imu1ate, control and diagnose. A 1'unct i()na l ll1()du I
is established, in which the perform platform includes AGV milster c()nt r()l
system, mdster moni tor system, carrier control system and s imu 1 at i ng systt'l1l.
*rhe deve lopping p1atform inc1udes system layout uti l ity, mdster nr()griln1
del'ini t ion uti l ity and vehic1e program defini t ion uti l ity.
(2) The AGV perform platform structure 1s estab1 ished on the bi1sls ()I'
software layering technology, model--view separate technology, messdge bus
technol ogy and di str ibuted techno l ogy. The system i s sepdrdted i nto represunt
1ilyer, fi eld ohject layer, service layer and data st()rage ld}cr. l t is
c()nveni cnt to deve1op and organi ze. An a] ternat ion mode betwoen l nt ert'iI()tt
Iilyer and fie1 d object layer is designed on mode1--v iew separdt e techno1 ()g}',
in which re--use of field object is highly improved. A message bus bi1sed on
order--issue model is designed for object alternation in fi eld layer, whi ch
gi ves the system good expansibi l ity. A distributing model i s desi gnc(l ()n
ilgc'nt--de's ign mode so the system can be organized in d i fferent c()InpLltc un l [ s.
(3 ) The system software is designed in UML model ing 1anguage. 'l'he syst em
stLl1 i c model is designed in category structure, and dynamic model is designcd
in ()bject a1ternating structure- Thi s thesis also states the subsysl em
st ruc ture and pr inc iple of task mandger, cdrrier manager, traffic lnilni:tgt.I-,
I () lllilnager, communlcat ion manager, dnd path mandger i n cle1i[ i l s. 'l.he
S l llltl l il t l ng tlnd mon i t()ring suhsys tem are tlI so presented.
II
(4 ) An AGV dispatching technique is put forward based on genet ic
iiIg()ri thms. A matbemat ic model combi ni ng task manager and carri er mdnklger
l s establ ished to improve the di spatching efficiency. A rckI l ll1()dc l
ca1 cu ldt ing val idates that genetic algori thm is dn et'fect ive meth()d l'()r A(i\/
d i s pkl t c h i ng.
Combinat ing the development practice of A(;V in my comDany, resuitrch
results show the design and technique for AGV master system in this thesi s
i s correct and dependab1e. It not only have great practi cal val ue in A(;V syst om
des ign, but a1so have vast appl icat ion perspect ive in designi ng the di spilt c,h
m()n itor ing system of intel1 igence carry equiprnent.
引文
[1]瑞典NDC公司AGV系统资料.内部资料,1999
[2]董平,赵海伶.AGV及AGVS方案研究.组合机床与自动化加工技术,2002年第2期
[3]郭峰,袁星军,余达太等.自动导引车系统关键技术的研究.北京科技大学学报.1999.21(2):220
[4]白小波,王宏玉等.AGVS系统技术应用.物流技术,2000(2):9
[5]张晓萍;颜永年,吴耀华,荆明.现代生产物流及仿真.清华大学出版社,1998年
[6]汪达开.AGV控制系统.机械工业自动化,1997年第3期
[7]Takakuwa,S. Efficient.module-based modeling forA large-scaleAS/RS-AGV system. Winter Simulation Conference Proceedings 1996.IEEE,Piscataway,NJ,USA.p 1141-1148
[8]Lwin,Myat.Developmen ofA simulatedAGV material handling system. Industrial Engineering (Norcross.Georgia) v 26 n 9 Sept 1994. p 48-49 Publication Year:1994
[9]Johnson,M. Eric; Brandeau,Margaret L.. Designing multiple-loadAutomated guided vehicle systems for delivering material fromA central depot. Journal of Engineering for Industry,Transactions of the ASME v 117 n 1 Feb 1995. p 33-41
[10]JonesAlbert. A multi-level/multi-layerArchitecture for intelligent shopfloor control.INT JCIM,1990, 3(1):60~70
[11]Shiizuka, Hisao; Suzuki, Misa. Modeling of AGV networks in flexible manufacturing systems.Computers & Industrial Engineering v 27 n 1-4 Sep 1994. p 81-86
[12]Langer,J.; Sagar,VK. Multiple capacity single vehicle job scheduling forAGV's.IEE Conference Publication n 395 1994.IEE,Stevenage,Engl.p 348-353
[13]Lee.Jeong-Hoon;Lee,Bum Hee; Choi,Myoung Hwan;Kim,Jung Duk.Joo,Kwang-Taek; Park, Hyon. Real time traffic control scheme forA multipleAGV system.Proceedings - IEEE International Conference on RoboticsAndAutomation v 2 1995.IEEE,Piscataway,NJ,USA, 95CB3461-1. p 1625-1630
[14]Jing,Hsu Wen;Ying,Huang Shell. Route planning of Automated guided vehicles. Intelligent Vehicles Symposium,Proceedings 1994. IEEE,Piscataway,NJ,USA,94TH8011. p479-484
[15]Yu, Feng;Chen,Hong;Zhang, Lanjun. ApproximateAnalysis of traffic congestion inAGV system. Proceedings of the IEEE International Conference on Industrial Technology 1906.IEEE,Piscataway,NJ,USA,96TH8151. p 867-869
[16]Lee, Jung Hoon; Lee,Beom Hee; Choi, Myoung Hwan.Real-time traffic control scheme of multipleAGV systems for collision free minimum time motion:A routing tableApproach.IEEE Transactions on Systems, Man, And Cybernetics PartA:SystemsAnd Humans.v 28 n 3 May 1998. p 347~358
[17]Ulusoy G,Bilge U. Simultaneous Scheduling of Ma-chinesAndAutomated Guided Vehicles.Interna-tional Journal of Production Research,1993,31 (12):2857~2873
[18]Hsieh, Ling-Feng; Sha,D.Y.. HeuristicAlgorithm for the design of facilities layoutAndAGV routes in tandemAGV systems.International Journal of Industrial Engineering v 4 n 1 Mar 1997.p 52-61
[19]吴耀华,肖际伟,高洪峰.AGV任务与路线调度问题的解决方案.生产自动化,2000年第11期
[20]袁颖河,吴智铭.FMS-AGV传输系统自学习调度方法.上海交通大学学报,1998年6月第32卷第6期
[21]李岩,吴智铭,甘泉.柔性加工环境中机器和AGV的集成调度.中国机械工程第12卷第1期2001年4月
[22]张宏远,方宗德,傅卫平,王会丽.基于Pelri网的Agv路径规划方法.组合机床与自动化加工技术,2001年第7期
[23]李岩,吴智铭.基于GA和机器学习的启发式规则调度方法.控制与决策,1999(增刊):
[24]钟江生.用神经网络方法处理FMS中多个AGV运送请求问题.制造业自动化,1999年12月第21卷第6期
[25]冯文镛,杨灿军,陈鹰.AGV原型系统的全局路径规划遗传算法研究.控制理论与应用,2002年第19卷第2期
[26]王小平,曹立明.遗传算法-理论、应用与软件实现.西安交通大学出版社,2002年
[27]Craig Larman.UMI.和模式应用 面向对象分析与设计导论.机械工业出版社,2002年
[28]Crady Booch,James Rumbaugh,Ivar Jacobson.EML用户指南.机械工业出版社,2002年
[29]刘锐东.UML对象设计与编程.北京希望电子出版社,2001年
[30]陈莉君.Linux操作系统内核分析.人民邮电出版社,2000年
[31]周之英.现代软件工程.科学出版社,2000年
[32]宛延阎.定海.面向对象分析和设计.清华大学出版社.2001年
[33]Wayne Wolf,嵌入式计算机系统设计原理,机械工业出版社,2002年
[34]Douglas E.Comer,David L.Stevens.用TCIP/IP进行网际互连.电子工业出版社,1988年
[35]Erich Gamma,Richard Helm,Ralph Johnson,John Vlissides.设计模式-可复用面向对象软件的基础.机械工业出版社,2000年