摘要
针对控制器局域网(ControllerAera Network,CAN)消息集的固定优先级抢占式调度(Fixed Priority Non-premptive Scheduling,FPNS)的可调度性分析问题,CAN消息修正的最差响应时间(Worst Case Response Time,WCRT)分析方法提出了CAN消息集的FPNS模型,并修正了基于单处理器系统任务集的固定优先级非抢占式调度(FixedPriority Non-preemptive Scheduling,FPNS)模型的CAN消息传统的WCRT分析方法的误判情况。但CAN消息修正的WCRT分析方法所引入的过差估计,缩小了CAN消息集可调度的优先级序列的解空间。这将可能造成对某些可调度的消息集做出不可调度的错误判断,并提升求解所需的计算量与相应的时间开销。为此,本文基于CAN消息集的FPNS模型重新分析了CAN消息响应时间可能出现的最差情况。其分析方法消除了CAN消息修正的WCRT分析方法的过差估计,保证了CAN消息集的FPNS的可调度性分析可在应有的解空间中求解可调度的优先级序列。另一方面,CAN消息的WCRT作为其可调度性的判定指标,虽然可作为衡量CAN通信系统实时性的重要依据。但其反映的仅为消息所可能经历的延时的上限,无法作为定量地衡量CAN通信系统实时性的指标。为此,本文通过对CAN消息响应时间构成要素的分析,提出CAN数据帧排队延时的概率分布可作为定量衡量CAN通信系统实时性的指标。通过对CAN总线传输活动的分析,根据Markov链理论分别建立了CAN数据帧的阻塞延时模型与直接经历的干扰延时模型,以推导CAN消息所经历的阻塞延时的概率分布与直接经历的干扰延时的概率分布,并最终得出了CAN数据帧排队延时的概率分布。
CAN作为被广泛用于实现车辆网络控制系统通信网络的总线技术,至今仍在汽车工业领域内发挥着不可替代的作用。然而,由于CAN所采用的载波监听多路访问/冲突避免(Carrier Sense Multiple Access/Collision Resolution,CSMA/CR)机制,在应用OSEK/VDX直接NM(Network Management,NM)机制的CAN网络管理系统中,网络控制系统的应用通信与NM通信会存在部分兼容性问题。在逻辑环的形成过程中,参与直接NM的全体节点可能在较短的时间内触发大量的alive报文与ring报文的发送请求。这种情况难于分析并有效调度,严重时甚至会影响网络控制系统的有效性与稳定性。为此,本文提出了逻辑环的顺序形成机制,有效地控制了单位时间内可能触发的alive报文与ring报文的发送请求数量,降低其与应用报文的发送请求产生冲突的可能。该机制的应用可为车辆提供更加安全、可靠的操作平台。此外,针对应用OSEK/VDX直接NM机制的CAN网络管理系统的理论研究与应用开发缺乏有效的软件仿真平台的问题。本文通过对NS2环境下已有模拟实现的网络协议的研究,结合对CAN总线协议的分析与抽象,在NS2环境下设计实现了CAN网络的直接NM系统的软件仿真平台。
面向媒体的系统传输(Media Oriented System Transport,MOST)作为车载信息娱乐网络通信的新一代的解决方案,虽然在网络传输带宽、数据的传输方式、资源共享等方面满足了车载信息娱乐系统全新的需求。然而,随着车载信息娱乐系统中信息娱乐设备数量的不断增加,启动过程耗时过长且待机过程耗电量过大已成为车载信息娱乐系统亟待解决的问题。同时,随着车载信息娱乐系统所提供的音视频资源种类的不断丰富,用户对各种音视频资源的动态选择与共享已成为必然的需求。因此,MOST需根据数据目的对各数据源所提供的多种音视频资源需求的变化,动态管理数据源与数据目的间的流数据连接及相应的流数据连接通道。然而,MOST规范仅描述了单对数据源与数据目的间流数据连接的建立与释放过程;数据源、数据目的,以及连接管理功能块与之相关的函数功能。为此,本文提出了适用于MOST规范所定义的全部网络实现标准的支持ECL唤醒的MOST网络的睡眠唤醒策略,用于解决车载信息娱乐系统启动过程耗时过长且待机过程耗电量过大的问题。还提出了以MOST网络的网络管理机制为基础MOST网络流数据连接管理的方案。该方案通过流数据连接管理相关信息的管理机制,以及流数据连接的建立、释放与切换策略,使连接管理主节点(Connection Master,CM)实现了在数据目的对数据源所提供的音视频资源的需求发生变化时,对可能存在的各路流数据连接的有效管理,从而为车载信息娱乐系统提供更为完善的服务。以上研究为MOST相关应用的产业化开发与商业推广提供了可靠的保障。
In the research of Fixed Priority Non-preemptive Scheduling (FPNS) for Controller AreaNetwork (CAN) message set, as the traditional Worst Case Response Time (WCRT) analysismay draw the wrong conclusion to some unschedulable message set, the modified Worst CaseResponse Time (WCRT) analysis proposed the FPNS model of CAN message set, modifiedthe misjudgement of the tranditional WCRT based on the FPPS model of single processor’stask set. However, the modified analysis is pessimistic and reduces the solution space ofshedulable priority sequence. This would make the misjudgement to some schedulablemessage set, cost more computation and time to find the solution. This dissertation discussedand revised the pessimistic element in the modified analysis with FPNS model. The relativeanalysis makes it is possible to find schedulable sequence in the original solution space. Onthe other hand, although the WCRT as the judging index of message schedulablity can be usedas an important basis to reflect the real-time of CAN communication system, it is only theupper bounder of the delay that message may experienced, not the quantitative index tomeasure the real-time of CAN communication system. After analysing the WCRT of CANmessage, this dissertation proposed that the probability distribution of CAN frame queuingdelay can be used as a quantitative index to measure the real-time of CAN communicationsystem. By analysing the transport activity of CAN bus with Markov chain theory, theblocking model and the direct experienced interference model of CAN frame are establishedto deduce the corresponding probability distribution, and get the probability distribution ofCAN frame queuing delay eventually.
As the most widely used network for in-vehicle Network Control System, CAN is stillplaying an irreplaceable role in automotive industry. However, since the Carrier SenseMultiple Access/Collision Resolution (CSMA/CR) mechanism of CAN, the communicationbetween NCS and Network Managment (NM) system is still incompatible in some scenariowith the CAN NM system applying OSEK/VDX direct NM. During the formation of logicalring, the nodes which participating the direct NM would have too many sending requests ofalive message and ring message in a short time. It is hard to schedule, and may corrupt theeffectiveness and the reliability of NCS in worst case. This dissertation proposed a sequentialformation mechanism of logical ring, which limits the sending requests of alive message andring message and reduces the possibility of collision with application message. The applyingof this mechanism will provide a more safable and reliable control platform for vehicle.Futhermore, This dissertation desgined and implemented a software simulation platform forthe theoretical research and application research of NM system applying OSEK/VDX direct NM mechanism, through researching the existing simulation of network protocol in NS2,analysing and abstracting CAN bus protocol.
As the next generation solution, Media Oriented System Tranport (MOST) has met thenew demands of in-vehicle infoentertainment network in bandwith, transmission mode,resource sharing, etc. But with the increasing amount of infoentertainment device distributedin in-vehcile infoentertainment system, the long time startup and the much power consumedstandby have become the critical problem of in-vehcile infoentertainment system. Moreover,with the types of audio and video resources provided by infoentertainment continuouslyenriched, dynamic selecting and sharing of various audio and video resource have become theinevitable demand of consumer. Therefore MOST need to manage the streaming connectionsand corresponding streaming channels dynamically according to the requirements of sinks tothe various resource provided by sources. However, MOST specification only describes thebuilding and releasing between single couple of source and sink, and their relative function.This dissertation proposed a sleep-wakeup strategy with Electrical Control Line (ECL)wakeup functionality, which applies to all the network implement standards defined by MOSTspecification to solve the long time startup and the much power consumed standby ofin-vehcile infoentertainment system. It also proposed a streaming connection managementscheme of MOST based on NM mechanism. This scheme supported by the infomantionmanagement mechanism and the building, releasing, switching strategy of the multiplestreaming connections, makes Connection Master (CM) manage multiple streamingconnections effectively when the requirements of sinks to the resource provided by sourceshas changed, provides more perfect service to in-vehcile infoentertainment. The researchabove guarantees the industrialization development and trade promotion of MOST.
引文
[1]周如培,李恩玲,冯同鑫.现代科学的奇迹——半导体技术[J].微电子学,2001,31(2):77-82.
[2]兴业证券.半导体行业发展研究报告[EB/OL]. http://www.xyzq.com.cn.
[3]李国杰.21世纪上半叶信息科学技术展望[J].科学发展,2010,25(1):78-86.
[4]龚进峰,曹健,袁大宏.浅谈我国汽车电子产业现状及发展建议[J].汽车工程,2004,26(3):363-366.
[5] Navet N, Simonot-Lion F. Automotive embedded systems handbook[M]. Boca Raton:The Chemical Rubber Company Press,2008.
[6] Navet N, Song Y, Simonot-Lion F, Wilwert C. Trends in automotive communicationsystems[C]. In Proceedings of the IEEE2005,93(6):1204-1224.
[7] Leen G, Heffernan D. Expanding automotive electronic systems[J]. Computer,2002,35(1):88-93.
[8] Albert A. Comparison of event-triggered and time-triggered concepts with regard todistributed control systems[C]. In Proceedings of Embedded World2004. Nuremberg:2004,235-252.
[9]徐高潮,胡亮,付晓东.计算机网络[M].长春:吉林大学版社,2002:1-7.
[10]白涛.网络化控制系统性能分析与调度优化[D].上海交通大学,2005.
[11] Cena G, Valenzano A, Vitturi S. Advances in automotive digital communications.Computer Standard and Interface[J],2005,27(5):665-678.
[12] Navet N, Song Y Q, Simonot F. Worst-case deadline failure probalility in real-timeapplications distributed over controller area network[J]. Journal of Systems Architecture,2000,46(7):607-617.
[13] Bosch. CAN specification version2.0[EB/OL].[2014-04-01] http://www.intron-tec-h.com.cn/downfile%5CCANcriterion%5CBOSCH%20CAN%20V2.0_L.pdf
[14] Davis R I, Burns A, Bril R J, Lukkien J J. Controller Area Network (CAN)schedulability analysis: refuted, revisited and revised[J]. Real-Time System,2007,35(3):239-272.
[15] LIN Consortium. LIN specification package revision2.1[EB/OL].
[2014-04-01]http://www.intron-tech.com.cn/downfile%5CCANcriterion%5CLIN%20v2.0_L.pdf
[16] Kimberley, William. LIN networks in the automotive industry[J]. AutomotiveEngineer (London),2004,29(8):50-54.
[17] FlexRay Consortium. FlexRay Communications Syetem protocol specificationVersion2.1Revision A[EB/OL].[2014-04-01] http://wenku.baidu.com/view/577c4ed428ea8-1c758f578ff.html
[18] Park I, Sunwoo M. FlexRay network parameter optimization method for automotiveapplication[J]. IEEE transactions on industrial electronics,2011,58(4):1449-1459.
[19]陈智琦,罗峰,陈觉晓,孙泽昌. FlexRay特性分析及应用实现[J].2007中国汽车工程学会年会论文集.北京:机械工业出版社,2007.
[20] MOST Cooperation. MOST Specification Revsion3.0.1[R]. Germany: MOSTCooperation.
[21]周户星.车联网环境下交通信息采集与处理方法研究[D].长春:吉林大学交通学院,2013.
[22]汪峥.车联网蠕虫传播与防治研究[D].南京:南京理工大学,2013.
[23] Tindell K, Burns A. Guaranteeing message latencies on controller area network[C].In Proceedings of the1stInternational CAN conference. Mainz:[出版者不详],1994.
[24] Tindell K W, Burns A, Wellings A J. Calculating Controller Area Network(CAN)message response times[J]. Control Engineering Practice,1995,3(8):1163-1169.
[25] Tindell K W, Hansson H, Wellings A J. Analysing Real-Time Communications:Controller Area Network (CAN)[C]. In Proceedings of the15th Real-Time SystemsSymposium1994(RTS’94). USA: IEEE,1994.
[26] Nolte T, Hansson H, Norstom C. Probabilistic worst-case response time analysis forthe controller area network[C]. In Proceedings of the9th Real-Time and EmbeddedTechnology and Application Symposium2003(RTAS’03). USA: IEEE,2003.
[27] Broster I, Burns A, Rodríguez-Navas G. Probabilistic analysis of CAN withfaults[C]. In Proceedings of the23rd Real-Time Systems Syposium (RTSS’02). USA: IEEE,2002.
[28] Broster I, Burns A, Rodríguez-Navas G. Timing analysis of real-timecommunication under electromagetic interference[J]. Real-Time Systems,2005,30(1-2):55-81.
[29]陈曦. CAN总线实时性和可靠性若干问题的研究[D].天津:天津大学电气与自动化工程学院,2010.
[30]杨云江.计算机网络管理技术[M].北京:清华大学出版社,2005:41-43.
[31]肖德宝,徐慧.网络管理理论与技术[M].武汉:华中科技大学出版社,2009:1-9.
[32] OSEK/VDX Committee. OSEK/VDX Network Management Concept andApplication Programming Interface Version2.5.3[EB/OL].[2014-04-01]. portal.osek-vdx.or-rg/files/pdf/specs/nm253.pdf
[33] Kim J H, Seo S H, Moon T Y, Hwang S H, Jeon J W. A Mehthod of ImplementingNetwork Mangement without OSEK/VDX OS[C]. In Proceedings of the7thInternationalConference on Control, Automation and Systems2007(ICROS’07). USA: IEEE,2007.
[34]陈筠翰.基于OSEK/VDX规范的CAN网络管理[D].长春:吉林大学计算机科学与技术学院,2010.
[35] Son C W, Seo S H, Kim J H, Kwon K H, Jeon J W, Hwang S H. Comparsion of twotypes of OSEK network management for CAN[C]. In Proceedings of2009IEEE InternationalSymposium on Industrial Electronics2009(IEEE ISIE’09).USA: IEEE,2009.
[36] Liu Z, Huang Z, Wang M, Hu K. A research on the network management for vehiclecomfort control system based on OSEK/VDX NM[J]. Automotive Engineering,2012,35(9):836-840.
[37] Hao B, Liu Y H, Qu L D, Wei D. Research and implementation ofgrouping-merging strategy for CAN network[J]. Chinese Journal of Scientific Instrument,2012,33(9):2137-2143.
[38] Suwatthikul J, McMurran R, Jones R P. Adaptive OSEK network management forin-vehicle network fault detection[C]. In Proceedings of2007IEEE International Conferenceon Vehicular Electronics and Safety (ICVES’07). USA: IEEE,2007.
[39] Wei C J, Yao M, Lu P, Hu Q, Zheng N G. OSEK/VDX-based dynamic networkmanagement on automotive network[C]. In the Proceedings of2009International Conferenceon Embedded Software and Systems (ICESS’09). USA: IEEE,2009.
[40] Luis M P, Francisco V. Reliable Real-Time Communication in CAN Network[J].IEEE Trans. on Computer,2003,52(12):1594-1607.
[41] Salvatore C. Meeting Real-time Constraints in CAN[J]. Industrial Informatics,2005,9(2):124-130.
[42]秦贵和,黄永平,等. MOST汽车多媒体网络[M].北京:北京理工大学出版社,2009:60-77.
[43] Liu C L, Layland J W. Scheduling algorithms for multiprogramming in a hardreal-time enviroment[J]. Journal of the ACM,1973,20(1):46-61.
[44] Lehoczky J P. Fixed priority scheduling of periodic task sets with arbitrarydeadline[C]. In Proceedings of the11th Real-Time Systems Symposium1990(RTSS’90).USA: IEEE,1990.
[45] Harbour M G, Klein M H, Lehoczky J P. Fixed priority scheduling of periodic taskswith varying execution priority[C]. In Proceedings of the12th Real-Time SystemsSymposium1991(RTSS’91). USA: IEEE,1991.
[46] Sha L, Rajkumar R, Lehoczky J P. Priority inheritance prtocols: an approach toreal-time synchronisation[J]. IEEE Transactions on Computer,1990,39(9):1175-1185.
[47] Baker T P. Stack-based scheduling of realtime processes[J]. Real-Time System,1991,3(1):67-99.
[48] George L, Rivierre N, Spuri M. Pre-emptive and non-pre-emptive real-timeuni-processor scheduling, Technical Report2966[R]. France: Institut National de Rechercheet Informatique et en Automatique (INRIA),1996.
[49] Bril R J, Lukkien J J, Davis R I, Burns A. Message response time analysis for idealController Area Network (CAN) refuted, CS-Report06-09[R]. The Netherlands: TechnischeUniversiteit Eindhoven (TU/e),2006.
[50]周炯磐,庞沁华,续大我,吴伟陵.通信原理(下)[M].北京:北京邮电大学出版社,2002:124-125.
[51] JANG K, PARK I, HAN J, LEE K. Design framework for FlexRay networkparameter optimization[J]. International Journal of Autonotive Technology,2011,12(4):589-597.
[52]饶运涛,邹继军,郑勇芸.现场总线CAN原理与应用技术[M].北京:北京航空航天大学出版社,2003:20-37.
[53] Broster I, Burns A, Rodriguez-Navas G. Timing analysis of real-timecommunication under electromagnetic interference[J]. Real-Time System,2005,30(1-2):55-81
[54] Notle T, Hansson H, Norstr m C. Probabilistic Worst-Case Response-Time Analysisfor the Controller Area Network[C]. In Proceedings of the9th IEEE Real-Time andEmbedded Technology and Applications Symposium (RTAS’03), USA: IEEE,2003.
[55] Audsley N, Burns A, Richardson M, Tindell K, Wellings A. Applying newscheduling theory to static priority pre-emptive scheduling[J]. Software Engineering Journal,1993,8(5):285-292.
[56] Punnekkat S, Hansson H, Norstr m C. Response time analysis under errors forCAN[C]. In Proceedings of the6th Real-Time Technology and Applications Symposium(RTAS’00). USA: IEEE,2000.
[57] Joseph M, Pandya P. Finding response times in a real-time system[J]. ComputerJournal,1986,29(5):390-395.
[58] SAE. Class C Application Requirement Considerations, Technical ReportJ2056/1[R]. USA: Vehicle Architecture For Data Communications Standards,1993.
[59]陈曦.基于排队论的CAN总线消息响应时间建模与分析[J].天津大学学报,2012,54(3):228-235.
[60]姜丹,钱玉美.信息论与编码基础[M].北京:电子工业出版社,2013:209-329.
[61]盛骤,谢式千,潘承毅.概率论与数理统计[M].北京:高等教育出版社,1989:18-28.
[62] Notle T, Hansson H, Norstr m C. Minimizing CAN response-time jitter by messagemanipulation[C]. In Proceedings of the8th IEEE Real-Time and Embedded Technology andApplications Symposium (RTAS’02), USA: IEEE,2001.
[63] Peter V I. An essay on free will[M]. USA: Oxford University Press,1986.
[64]伊曼纽尔·帕尔逊.随机过程[M].上海:高等教育出版社,1987:200-206.
[65]张卓奎,陈慧婵.随机过程及其应用[M].西安:西安电子科技大学出版社,2012:151-159.
[66]吴祈耀.随机过程[M].北京:国防工业出版社,1984:230-240.
[67] Christos G. Cassandras, St é phane Lafortune. Introduction to Discrete EventSystems[M]. Boston:Kluwer Academic,1994.
[68]曹万科. CAN协议车载网络若干关键理论研究[D].沈阳:东北大学车辆工程研究所,2007.
[69] Joseph Lemieux. OSEK/VDX汽车电子嵌入式软件编程技术[M].北京:北京航空航天大学出版社,2004.
[70] OSEK/VDX Committee. OSEK/VDX Operating System Version2.2.3[EB/OL].[2014-04-01]. portal.osek-vdx.org/files/pdf/specs/os223.pdf.
[71] OSEK/VDX Committee. OSEK/VDX Communication Version3.0.3[EB/OL].[2014-04-01]. portal.osek-vdx.org/files/pdf/speces/osekcom303.pdf.
[72] OSEK/VDX Committee. OSEK/VDX Implementation Language SpecificationVersion2.5[EB/OL].[2014-04-01]. portal.osek-vdx.org/files/pdf/speces/oil25.pdf.
[73] Road vehicles-Interchange of digital information-Controller Area Network(CAN)-Part1: Data link layer and physical signaling[R], ISO11898-1,2003.
[74] Leohold J. Automotive system architecture. In the Proceedings of the summerschool on architectural paradigms for dependable embedded systems[C]. Austria: ViennaUniversity of Technology,2005.
[75] Notle T. Share-driven scheduling of embedded networks[D]. Sweden: MalardalenUniversity,2006.
[76] Frischkron H-G. Automotive architecture requirements. In the Proceedings of thesummer school on architectural paradigms for dependable embedded systems[C]. Austria:Vienna University of Technology,2005.
[77]瞿亮,王亚,凌志刚等.计算机仿真技术[M].北京:电子工业出版社,2013:1-6.
[78]肖田元,范文慧.系统仿真导论[M].北京:清华大学出版社,2010:1-22.
[79]王精业,谭亚新,孙明等.仿真科学与技术原理[M].北京:电子工业出版社,2012:1-64.
[80]王建平,李怡菲,杜玉红等.计算机网络仿真技术[M].北京:清华大学出版社,2013:1-2.
[81]周时莹. CAN网络控制系统的智能调度研究[D].长春:吉林大学计算机科学与技术学院,2009.
[82]赵耀. NS2对CAN协议模拟的扩展[D].长春:吉林大学计算机科学与技术学院,2010.
[83] The Network Simulator-ns-2[EB/OL].[2014-04-01]. www.isi.edu/nsnam/ns,2004.
[84] Daimler Introduces MOST150in the New Mercedes-Benz S-class[R]. Germany:MOST Cooperation,2014.
[85] SMSC. MOST50Intelligent Network Interface Controller OS81092[R]. Germany:SMSC,2011.
[86] MOST Cooperation. MOST High Protocol Service Version3.0.x UserManual/Specification[R]. Germany: MOST Cooperation,2012.
[87] MOST Cooperation. MOST NetServices LayerII Wrapper for INIC Version3.0.xUser Manual/Specification[R]. Germany: MOST Cooperation,2012.
[88] MOST Cooperation. MOST NetServices LayerI Wrapper for INIC Version3.0.xUser Manual/Specification[R]. Germany: MOST Cooperation,2012.
[89] SMSC. MOST INIC Hardware Concepts[R]. Germany: SMSC,2011.
[90] MOST Cooperation. MOST FunctionBlock NetBlock Version3.0.2[R]. Germany:MOST Cooperation,2011.
[91] SMSC. OS81092INIC API Preliminary User’s Manual[R]. Germany: SMSC,2010.
[92] MOST Cooperation. MOST Electrical Control Line Specification Vervsion1.1.1[R].Germany: MOST Cooperation,2011.
[93] MOST Cooperation. MOST Dynamic Specification Version3.0.1[R]. Germany:MOST Cooperation,2010.
[94] MOST Cooperation. Optolyzer OL3050e Version1.2.x User Manual[R]. Germany:MOST Cooperation,2010.
[95]秦贵和,南洋,陈筠翰等.面向媒体的系统传输网络连接管理策略[J].吉林大学学报(工学版).2012,42(4):963-970.
[96] Andreas Grzemba. The Automotive Multimedia Network from MOST25toMOST150[M]. Germany: Franzis Publishing,2011:78-92.
[97]南洋. MOST网络管理与多信道连接管理的设计与实现[D].长春:吉林大学计算机科学与技术学院,2012.
[98] MOST Cooperation. MOST FunctionBlock ConnectionMaster Version3.0.2[R].Germany: MOST Cooperation,2011.