摘要
随着计算机技术、网络技术、视频压缩编码技术的快速发展,特别是宽带通信和Internet的快速普及,极大地促进了流媒体传输技术的发展。然而由于Internet与生俱来的尽力而为(best effort)的特性,不能保证流媒体传输系统的服务质量,因此要求流媒体系统必须能够提供与网络带宽波动相适应的传输与播放控制机制。流媒体传输具有很强的挑战性,它涉及计算机网络拥塞控制、媒体同步、实时服务质量(QoS, Quality of Service)控制等多个领域。尽管控制理论在流媒体视频传输、播放的研究上已经取得了一些成果,但由于流媒体模型的复杂性、建模困难,具有参数时变、积分时变时滞以及非线性等特点使得传统控制控制理论很难取得满意的控制效果。因此,本文使用先进控制理论作为理论工具对流媒体传输、播放控制中遇到的关键问题进行研究,主要包括系统建模、稳定性分析、网络带宽约束下优化求解及传输时滞等问题。本文的主要工作有：
针对基于缓冲区控制的流媒体速率控制模型进行了系统辨识,辨识中使用了流媒体传输中常用的UDP协议作为传输层协议,实验中发现实验数据具有较强的相关性,使用最小二乘法辨识得到的模型不准确,动态偏差和稳态偏差都很大。为此在数据分析的基础上进行了理论推导,对基于缓冲区控制的流媒体速率控制模型进行了验证、补充和完善。
针对流媒体稳定平滑传输问题进行研究。在实时流媒体传输中,将服务质量QoS指标映射为反馈控制框架中的性能指标和约束条件,其中映射的两个要素是网络带宽限制和网络带宽波动最小化,作为控制约束条件直接参与控制器的设计,结合已辨识的实时流媒体传输数学模型,流媒体稳定平滑传输问题的求解可归结为先进控制理论中的动态矩阵控制算法(DMC,预测控制最经典的算法),实现QoS控制。
基于TFRC算法构建了一个实时流媒体传输架构,在其下针对播放端进行理论分析,得出播放速率控制模型,在播放速率约束的条件下通过经典的PI控制器调节播放速率对缓冲区占用水平进行调整。
对于网络时延对流媒体系统的作用进行研究分析。指出经典的时滞算法存在的缺陷。针对发送端发送数据和播放端播放数据共同作用于播放缓冲区的原理,在其基础上提出了一种双重控制策略,通过控制结构的改变增强系统对时滞问题的处理能力。在发送端使用内模控制(IMC)来抑制时延对控制系统的影响,在播放端使用PI控制器调节播放速率调整缓冲区占用水平,两种控制的双重作用下进一步提高了对于时滞的控制效果。
为了更好地保证流媒体系统的实时约束下的同步性能,解决流媒体缓冲区设置中同步与播放时延的矛盾,在缓冲区设置最大化的基础上,结合设定点跟踪算法,提出了一种基于最小方差控制(MVC, Minimum Variance Control)控制的实时流媒体缓冲区自适应设定值跟踪传输算法,在动态设置缓冲区占用水平的条件下,采用最小方差控制对缓冲区设定值进行跟踪,避免超调量过大对实时约束的影响。
本文使用仿真实验等手段验证了所提出算法的正确性和有效性。
The rapid development of the computer and network technology and video compression coding, especially the popularization of broadband and Internet, stimulates the development of streaming media greatly. Because of the inherent "best effort" characteristic of Internet, it cannot guarantee the Quality of Service of the streaming media.So streaming media system should provide adaptive control mechanism of transmission and playback according to the network bandwidth fluctuation. The transmission control of streaming media is full of challenge, it involves network congestion control,meidia synchronization and Quality of Service in realtime. Although there are some achievements of control theory on the research of transmission and playback, it is very difficult to achieve satisfactory result due to the complexity of streaming media model such as difficult modeling, time varying parameters, and integral delay etc. So we use advanced control theory as research tool in the transmission and playback control, including system modeling, stability analysis, optimization under the constraint of bandwidth and transmission delay.The main works and contributions of this dissertation are summarized as follows:
A practical system identification approach is adopted to identify the rate control model based on buffer control. UDP is commonly used as the transmission protocol for real time property. The accurate model can't be analysed by the least squares estimator for the experiment data relevance and dynamic and steady error is very big. We testify, supplement and improve the rate control model based on buffer control by data analysis and theoretic proof.
The stable and smooth transmission mechanism is studied. QoS guarantees are mapped into the feedback control framework as index and constraints,in which two main factors bandwidth constraint and minimum network bandwidth fluctuation are added as constraint into the controller design.Combined with the model identified he stable and smooth transmission mechanism is solved by DMC in advanced control,which is a classical algorithm in the predictive control. QoS guarantees are achieved.
A framework of realtime streaming media transmission based on TFRC is constructed. We get the playback rate model by doing theory analysis on playback in the framework.. The playback rate is controlled by classic PI controller, which is applied to achieve continuous media playback by controlling the buffer level of the sink under playback rate constraints.
The effect of transmission delay on the streaming media system is studied. The inherent problem of existing control method is pointed out. Data transmission in the source and playback in the sink are applied to control the buffer length of the sink simultaneously, on which dual mode control is applied by changing the structure of controller in order to improve the performance of transmission delay. An internal mode feedback control of sending rate in the source is adopted to overcome the adverse effect caused by propagation delay, and a simple PI controller of playback rate in the sink is also adopted, which can deal with transmission delay simultaneously.
A transmission algorithm of adaptive buffer setpoint trace in real-time stream media based on MVC (Minimum Variance Control) is proposed in order to get synchronization performance ganrantee under the realtime constraint in the streaming media system and to deal with the conflict between synchronization and playback delay based on buffer maximization. The buffer level is dynamically adjusted and MVC is used to track the buffer setpoint and reduce effect of the overshoot on realtime constraint. Simulation experiments that have been conducted prove that the proposed algorithm is effective.
引文
1.李靖.流媒体服务系统中接入控制与缓存策略的研究[D].合肥：中国科学技术大学,2009
2. Huang C, Xu L H. SRC:Stable rate control for streaming media[C]//Proc. IEEE Globecom, San Francisco, USA,2003,7:4016-4021
3. Wang Y, Osermann J, Zhang YQ. Video Processing and Communication[M]. Prentice Hall,2002
4. Wu Dapeng,Hou Yiwei Thomas,Zhu Wenwu,et al.On End-to-End Architecture for Transporting MPEG-4 Video Over the Internet[J].IEEE Trans. on Circuits and Systems for Video Technology,2000,10(6):923-941
5.胡敏,刘达,余兆明.基于因特网的MPEG-4码流传输[J].电视广播与传输,2002,6：19-22
6. Lee Junghoon,Park Seungjun.Design of a DAVIC Residential Network based on EthernetfC]. Proceedings of Fourth International Workshop on Real-Time Computing Systems and Applications.October 1997,223-228
7. Talluri R,Texas Instruments.Error-resilient Video Coding in the ISO MPEG-4 Standard [J]. IEEE Communication Magazine,1998,36(6):112-119
8. IEEE EFM Task Force,IEEE 802.3ah Ethernet in the First Mile Task Force,http://grouper.ieee.org/ groups/802/3/efm/public/,May 2002
9.江滢,孟丹.基于接纳时间比控制和比例积分调节器的接纳控制机制[J].计算机研究与发展,2007,44(1)：65-70
10.林闯,单志广,任丰原.计算机网络的服务质量(QoS)[M],北京,清华大学出版社,2004
11. Nelakuditi S, Harinath RR, Kusmierek E, Zhang ZL. Providing smoother quality layered video stream[C]. In:Proc.Of the NOSSDAV.Chapel Hill,2000.1-10
12. Biersack E, Geyer W. Synchronized delivery and playout of distributed stored multimedia streams[J]. Multimedia Systems,1999,7(1):70-90
13. Ralf Steinmetz.Human perception of jitter and media synchronization[J],IEEE Journal on Selected Areas in Communications,1996,14(1),61-72
14.许延,常义林,刘增基.多媒体同步系统的缓冲区补偿技术[J].计算机学报,2003,23(4)：484-490
15. Rothermel K, Helbig T. An Adaptive Protocol for Synchronizing Media Streams[J].Multimedia Systems,1997,5(5):324-336
16. Koehler D, Mueller H. Multimedia playout synchronization using buffer level control[C].In:Steinmetz R(ed).The 2nd InternationalWorkshop on Advanced Teleservices and High-Speed Communication Architectures, LCNS 868, Heidelberg:Springer,1994.167-180
17.许延,常义林,刘增基.一种新的媒体内同步控制算法[J].计算机研究与发展,2002,39(12)：1675-1680
18.许延,常义林,刘增基.网络环境下视频连续播放控制算法[J].电子学报,2004,32(7)：1078-1081
19. ISO/IEC 1449622. Generic coding of audio visual objects:Visual.ISO/IEC JTC1/SC29/WG11/ N3095,Dec.1999
20. Li W P. Overview of fine granularity scalability in MPEG-4 video standard. IEEE Transactions on Circuits and Systems for Video Technology,2001,11 (3):301-317
21. Zhang Q, Zhu W, Zhang Y Q. Resource allocation formultimedia st reaming over the Internet[J]. IEEE Transactions on Multimedia,2001,3 (3):339-355
22. Kim T, Ammar MH. Optimal quality adaptation for MPEG-4 fine-grained scalable video[C]. In:Proc. of the IEEE INFOCOM. San Francisco,2003.641-645
23.尹浩,林闯,张谦等.端到端MPEG_4FG视频TCP友好的平滑传输[J].软件学报,2005,16(5)：931-939
24.黄天云.一类MPEG-4 FGS流视频传输的质量定义及算法[J].计算机学报,2006,29(5)：751-759
25.周正武,董育宁.IP网络实时视频流的传输控制算法AVTC的研究[J].计算机研究与发展,2004,41(5)：812-820
26. Tarek F. Abdelzaher, John A. Stankovic, Chenyang Lu, Ronghua Zhang, Ying Lu. Feedback performance control in software services[J]. IEEE Control Systems Magazine,2003,23(3):74-90
27. J.A.Stankovic, C.Lu, S.H.Son, and GTa.The Case for Feedback Control Real-Time Scheduling[C]. EuroMicro Conference on Real-Time Systems.1999.11-20
28. C. Lu, J. A. Stankovic, G. Tao, and S. H. Son. Feedback Control Real-Time Scheduling:Framework, Modeling, and Algorithms[J]. Real-Time Systems Journal,2002,23(1/2):85-126
29. C.Lu J.A.Stankovic G.Tao and S.H.Son. Design and Evaluation for a Feedback Control EDF Scheduling Algorithm[C]. IEEE Real-Time Systems Symposium.1999.56-67
30.任丰原,林闯,刘卫东.IP网络中的拥塞控制[J].计算机学报,2003,26(9)：1025-1034
31.任丰原,林闯,任勇,山秀明.ATM网络拥塞控制中PID控制器的设计[J].计算机学报,2002,25(10)：1024-1029
32.任丰原,林闯,任勇,山秀明.大时滞网络中的拥塞控制算法[J].软件学报,2003,14(3)：503-511
33.任丰原,林闯,王福豹.ABR流量控制中的变结构控制器[J].软件学报,2003,14(3)：562-568
34.于海滨,魏立峰,陆勇林.实时多媒体流控中的内模控制器设计[J].通信学报,2004,25(11)：90-97
35.魏立峰,于海斌,祁红岩.实时多媒体流控中的预测最优均匀控制器设计[J].控制与决策,2005,20(2)：214-216
36.任丰原,林闯,王福豹.RED算法的稳定性：基于非线性控制理论的分析.计算机学报,2002,25(12)：1302-1307
37. Hong Y. and Yang O.W.W. Design of an adaptive PI rate controller for streaming media traffic based on gain and phase margins[J]. IEE Proc.Commun.,2006,153(1):5-14
38. Yang Hong, Oliver W.W. Yang.Design of Adaptive PI Rate Controller for Best-Effort Traffic in the Internet Based on Phase Margin[J]. IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS,2007,18(4):550-561.
39. Ruijun Zhu,Cuiping Zhen.Fuzzy PI Rate Controller for Streaming Media Traffic[J]. Lecture Series on Computer and Computational Sciences,2005.
40. Ruijun Zhu,Cuiping Zhen. Robust Output Feedback Controller of TCP/AQM Router Based on LMI[J]. Lecture Series on Computer and Computational Sciences,2005
41. Yuang Maria C, Liang Shih T, Chen Yu G. Dynamic vi-deo playout smoothing method for multimedia applications[J].Multimedia Tools and Applications,1998,6 (1):47-60.
42.严峻,戴琼海.基于模糊控制的自适应播放算法[J].清华大学学报(自然科学版),2008,48(1)：157-160
43. Wanqing Tu, Weijia Jia.Adaptive playback buffer for wireless streaming media[c]//Proceedings on 12th IEEE International Conference Networks,2004.1,191-195
44. Jean-Chrysostome Bolot. End-to-end packet delay and loss behavior in the Internet[C].In:Proc of SIGCOMM 1993. San Francisco,CA:ACM,1993.289-298
45.邵奇可,俞立,张贵军.网络时延在线估计技术与控制器的协同设计[J].自动化学报.2007,33(7)：781-784
46. Astrom K J, Wittenmark B, Adaptive control (2nd Ed.)[M], New York:Addison-Wesley,1995
47. Lu C, Abdelzaher T, Stankovic J, and Son S.A feedback control approach for guaranteeing relative delays in Web servers[C]//Proc. IEEE Real-Time Technology and Appli-cations Symp., Taipei, China:IEEE,2001,51-62.
48. S.K.Park and K.W.Miller, Random number generators:good ones are hard to find, Communications of ACM,1988,21(10)
49. Jean-Chrysostome Bolot.End-to-end packet delay and loss behavior in the Internet[C]//Proc of SIGCOMM 1993. San Francisco,CA:ACM,1993,289-298
50.耿福泉,方敏,赵林亮等.基于RTCP反馈的TCP友好的实时流媒体拥塞控制机制[J].东北大学学报,2006,27(11)：1204-1207.
51. Nagle J. Congestion Control in IP/TCP Internetworks. RFC896,1984
52. van Jacobson. Congestion avoidance and control. ACM Computer Communication Review,1988,18 (4):314-329
53. Floyd S,Fall K. Promoting the use of end-to-end congestion control in Internet[J].IEEE/ACM Transactions on Networking,1999,7 (4):458-472
54. Caserri C, Meo M. A new approach to model the stationary behavior of TCP connections[C].In: Proceedings of IEEE INFOCOM2000, Tel Aviv, Israel, CA,2000.367-375
55.汪小帆,孙金生,王执铨.控制理论在Internet拥塞控制中的应用[J].控制与决策,2002,17(2)：129-134
56. Stevens W. TCP slow start, congestion avoidance, fast retransmit, and fast eecovery algorithms. RFC 2001,1997
57. Floyd S, Henderson T. The NewReno modification to TCP's fast recovery algorithm. RFC 2582,1999
58. Mathis Matthew, Mahdavi Jamshid, Floyd Sally, Romanow A. TCP selective acknowledgement options. RFC 2018,1996
59. Brakmo L S, Peterson L L. TCP vegas:End-to-end congestion avoidance on a global Internet[J]. IEEE Journal on Selected Areas in Communications,1995,13 (8):1465-1480
60. Mahdavi J, Floyd S. TCP-friendly unicast rate-based flow control.http://www.psc.edu/networking/ tcp_friendly.html,1997
61. Padhy J, Firoiu V,Towsley D, Kourose J.Modeling TCP throughput:A simple model and its empirical validation[C]. In:Proceedings of ACM SIGCOMM Symposium on Communications Architecture and Protocol, Vancourver, Canada,1998.303-314
62. Floyd Sally, Handley Mark, Padhye Jitendra, Widmer Joerg. Equation-based congestion control for unicast applications[C].In:Proceedings of ACM SIGCOMM Symposium on Communications Architecture and Protocol,2000.43-56
63. Bansal Deepak, Balakrishnan Hari. Binomial congestion control algorithms[C]. In:Proceedeings of IEEE INFOCOM2001,Anchorage,AK,2001.631-640
64. Yang Y, Lam S. General AIMD congestion control. University of Texas at Austin:Technical Report TR22000209,2000. http://www.cs.utexas.edu/users/lam/NRL/
65. Rhee I, Ozdemir M, Yi Y. TEAR:TCP emulation at reciver-flow control for multimedia streaming. NCSU:Technical Report,2000. http://www.csc.ncsu.edu/
66. G. Franklin, J. Powell, M. Workman.Digital Control of Dynamic Systems[M],3rd ed., Addison-Wesley Longman, Inc.1997.
67.郭方路,岳红,高东杰.一种新的最优均匀液位控制策略[J].化工仪表与自动化,2002,26(5)：528.
68. McDonal K A, M cAvoy T J, Tits A. Optimal averaging level control[J]. AlChE J,1986,32 (1): 75-86
69. Campo Peter J, Manfred Morari. Model predictive optimal averaging level control[J]. AlChE J,1989, 35(4):579-591
70.俞金寿,蒋慰孙.过程控制工程(第3版)[M].北京：电子工业出版社,2007..
71. Wai-tian Tan, Avideh Zakhor.Real-time Internet video using error resilient scalable compression and TCP-friendly transport protocol[J]. IEEE Trans on Multimedia,1999,1 (2):172-186
72. Reza Rejaie, Mark Handley, Deborah Estrin.RAP:An end-to-end rate-based congestion control mechanism for realtime streams in the Internet[C]//Proceeding of IEEE INFOCOM. Piscataway:IEEE Press,1999.1337-1345
73.席裕庚.预测控制[M].北京：国防工业出版社,1993
74.邹涛等.模型预测控制工程应用导论[M].北京,化学工业出版社,2010
75.邹涛,刘红波,李少远.锅炉汽包水位非自衡系统的预测控制[J].控制理论与应用,2004,21(3)：386-390
76. Pannocchia R, Rawlings J B,Wright S J. Fast, large-scale model predictive control by partial enumeration [J]. Automatica,2007,5(43):852-860
77. Karmarkar N. A new polynomial-time algorithm for linear programming[J].Combinatorica, 1984,4(4),373-395
78. Camacho E F,Bordons C. Predictive Control with Constraints[M].London:Springer,1999
79. Wang Liuping. Model Predictive Control System Design and Implementation Using MATLAB[M]. London:Springer,2009
80. Vanderbei V. an Interior Point code for Quadratic Programming[J].Technical Report SOR 94-95,Princeton Universty,1999,1(11):451-484
81. Nocedal J.Wright S J. Numerical Optimization[M]. New York:Springer,1999
82. Kouvaritakis B, Rossiter J A,Schuurmans J. Efficient robust predictive control [J]. IEEE Transactions on Automatic Control,2000,8(45):1545-1549
83. Johansen T, Grancharova A. Approximate explicit constrained linear model predictive control via orthogonal search tree [J]. IEEE Transactions on Automatic Control,2003,5(48):810-815
84. Schmid C, Biegler L T. Quadratic programming methods for reduced hessian SQP[J].Comp. Chem. Eng,1994,9(18):817-832
85. Qin S Joe, Badgwell A. A survey of industrial model predictive control technology[J].Control Engineering Practice,2003,7(11):733-764
86. KALMAN M, STEINBACH E, GIROD B. Adaptive media playout for low-delay video streaming over error-prone channel [J]. IEEE Trans on Circuits and Systems for Video TechSpecial Issue on Wireless Video,2004,14 (6):841-851
87.谢希仁.计算机网络(第三版)[M].大连：大连理工大学出版社,2000
88.王树青.工业过程控制工程[M].化学工业出版社,2003
89. Garcia C E, Mo rari M. Internal Model Control-l.A Unifying Review and Some New ResultsfJ]. I&EC Process Des.Dev.,1982,21 (2):308-323
90. MORARI M, zAF IoI Js E. Robust Process Control[M]. New Jeresy:Prentice Hall,1989
91. ECONOMOU C G. Internal model control-5:extension to nonlinear systems [J].Industrial Engineering Chemistry Process Design and Development,1986,25(3):403-411
92. HUNT K J, SBURBURO D.Neural networks for nonlinear internal model control[J].IEE Proc-D,1991,13(8):431-438
93. NAHAS E P. Nonlinear internal model control strategy for neural network modcls[J].Computers Chemical Engineering,1992,16 (12):1039-1057
94. HENSON M A,SEBORG D E. An internal model control strategy for nonlinear systems [JJ.America Institute of Chemical Engineering Journal,1991,37 (7):1065-1081
95. ALVAREZ J, ZAZUETA S. An internal-model controller for a class of single-input single-output nonlinear systems:stability and robustness [J]. Dynamics and Control,1998,8 (2):123-144
96. HU Q,SAHA P,RANGAIAH G P. Adaptive internal model controlof nonlinear processes[J]. Chemical Engineering Science,1999,54(9):1205-1220
97.周涌,陈庆伟,胡维礼.内模控制研究的新发展[J].控制理论与应用,2004,21(3)：475-482
98.邵惠鹤.工业过程高级控制(第二版)[M].上海：上海交通大学出版社,2003
99.刘士荣,林卫星,俞金寿等.非线性动态系统神经模糊建模与内模/PID双重控制系统设计[J].控制理论与应用,2004,21(4):553-560
100. FRANKLIN G,POWELL J,EMAMI-NAEINI A.Feedback Control of Dynamic Systems(3rd ed.) [M],Addison-Wesley Publishing Company,1993