高速公路入口匝道智能控制方法的研究
|
摘要
近年来，随着车辆数目的急剧增加，高速公路上交通堵塞的现象屡见不鲜，这使得如何采用合理的控制方法，预防或缓解高速公路上的拥挤状态显得尤为重要。入口匝道控制方法是控制高速公路交通状态最直接有效的方法之一。本文针对高速公路交通系统的特点，对入口匝道智能控制方法展开研究，主要的研究内容和结论如下：
针对高速公路交通模型难确定的问题以及高速公路交通状态具有一定的可重复性的特点，本文采用迭代学习控制(ILC)方法对高速公路的交通密度进行控制。在一阶开环ILC匝道控制算法的基础之上，设计了一阶开闭环ILC匝道控制算法。严格的理论分析证明了此一阶开闭环ILC匝道控制算法的收敛性，仿真分析验证了此算法的有效性与优越性。此一阶开闭环ILC匝道控制算法结构简单，且不依赖于精确交通模型，能够有效避免由交通模型难确定而带来的控制结果难精确问题。与一阶开环ILC匝道控制方法相比，此一阶开闭环方法使系统具有更快的收敛速度以及更好的暂态性能。
在一阶开闭环ILC匝道控制算法的基础上，设计了高阶开闭环ILC匝道控制算法，此算法不仅具有一阶开闭环ILC匝道控制算法的优点，而且因为其利用了更多的状态信息进行控制，系统的稳定性更好。严格的理论分析证明了此方法的收敛性，仿真分析验证了它的有效性以及优越性。与经典的ALINEA方法相比，此方法使系统拥有更精确的控制结果以及更好的稳定性。
针对高速公路交通模型难确定，控制量难自适应交通变化以及难实现多路段协调控制三个问题，本文设计了一种基于模糊神经网络方法的匝道控制算法。此算法中，模糊规则既实现了相邻路段间的协调控制，又调整了高速公路干道交通状态与匝道排队长度的大小；神经网络的自学习功能使此方法对外界交通变化具有自适应能力。该算法能够在将高速干道交通密度维持在理想密度附近的同时，保持入口匝道排队长度尽可能的短。在抑制交通密度波动和排队长度增长方面比经典的反馈控制方法ALINEA取得了更好的效果。值得提出的是此方法具有一定的协调控制能力，比单匝道控制方法对道路容量的利用更加充分，这在仿真数据中也有明确体现。
In rescent years the phenomena of traffic congestion on urban freeways is becoming more and more familiar due to the dramatic expansion of car-ownership, which makes it urgent and significant to apply efficient control approaches to keep the traffic condition in an ideal way. On-ramp metering, when properly applied, is considered as an efficient traffic management tool for freeways. A research on intelligent freeway on-ramp metering approaches was conducted according to some characteristics of the freeway traffic state. The main contents and results are as follows:
According to the repeatability of the traffic pattern for freeway, iterative learning approach (ILC, in short) method was applied to regulate the congestion state. A one order open-closed loop ILC scheme was designed to control the traffic density. By means of rigorous theoretic analysis, its validity and superiority are adequately proved. Simulation results validate its efficiency. When compared to the conventional open loop ILC approach, the strategy proposed in this thesis has faster convergence speed and better transient performance than the open loop ILC law.
Based on the one order open-closed ILC approach, a high order open-closed loop ILC scheme was further analyzed and designed. This approach does not only have the advantages as the first order one, but also makes the entire freeway system more stable as it collects more system information to design the metering rate. Simulation results confirmed its efficiency. When compared to the well-known ALINEA approach, this high order open-closed loop ILC approach has faster convergence speed and better robust performance.
According to the problems that are encountered in current ramp control research: uncertainty of the traffic model、difficulty in designing an adaptive ramp controller and the necessary of the coordinated algorithms, the neuro-fuzzy network method was applied to resolve the problems above and an adaptive and coordinated on-ramp strategy was designed. This approach considers both the traffic density and the on-ramp queue length together and aims at controlling both of them simultaneously. When compared to the well-known ALINEA strategy, this approach has better performance in restraining traffic flow fluctuation and queue increase. It is worthy to point out that the approach is coordinated, which can utilize the road capacity better than the single on-ramp methods.
针对高速公路交通模型难确定的问题以及高速公路交通状态具有一定的可重复性的特点，本文采用迭代学习控制(ILC)方法对高速公路的交通密度进行控制。在一阶开环ILC匝道控制算法的基础之上，设计了一阶开闭环ILC匝道控制算法。严格的理论分析证明了此一阶开闭环ILC匝道控制算法的收敛性，仿真分析验证了此算法的有效性与优越性。此一阶开闭环ILC匝道控制算法结构简单，且不依赖于精确交通模型，能够有效避免由交通模型难确定而带来的控制结果难精确问题。与一阶开环ILC匝道控制方法相比，此一阶开闭环方法使系统具有更快的收敛速度以及更好的暂态性能。
在一阶开闭环ILC匝道控制算法的基础上，设计了高阶开闭环ILC匝道控制算法，此算法不仅具有一阶开闭环ILC匝道控制算法的优点，而且因为其利用了更多的状态信息进行控制，系统的稳定性更好。严格的理论分析证明了此方法的收敛性，仿真分析验证了它的有效性以及优越性。与经典的ALINEA方法相比，此方法使系统拥有更精确的控制结果以及更好的稳定性。
针对高速公路交通模型难确定，控制量难自适应交通变化以及难实现多路段协调控制三个问题，本文设计了一种基于模糊神经网络方法的匝道控制算法。此算法中，模糊规则既实现了相邻路段间的协调控制，又调整了高速公路干道交通状态与匝道排队长度的大小；神经网络的自学习功能使此方法对外界交通变化具有自适应能力。该算法能够在将高速干道交通密度维持在理想密度附近的同时，保持入口匝道排队长度尽可能的短。在抑制交通密度波动和排队长度增长方面比经典的反馈控制方法ALINEA取得了更好的效果。值得提出的是此方法具有一定的协调控制能力，比单匝道控制方法对道路容量的利用更加充分，这在仿真数据中也有明确体现。
In rescent years the phenomena of traffic congestion on urban freeways is becoming more and more familiar due to the dramatic expansion of car-ownership, which makes it urgent and significant to apply efficient control approaches to keep the traffic condition in an ideal way. On-ramp metering, when properly applied, is considered as an efficient traffic management tool for freeways. A research on intelligent freeway on-ramp metering approaches was conducted according to some characteristics of the freeway traffic state. The main contents and results are as follows:
According to the repeatability of the traffic pattern for freeway, iterative learning approach (ILC, in short) method was applied to regulate the congestion state. A one order open-closed loop ILC scheme was designed to control the traffic density. By means of rigorous theoretic analysis, its validity and superiority are adequately proved. Simulation results validate its efficiency. When compared to the conventional open loop ILC approach, the strategy proposed in this thesis has faster convergence speed and better transient performance than the open loop ILC law.
Based on the one order open-closed ILC approach, a high order open-closed loop ILC scheme was further analyzed and designed. This approach does not only have the advantages as the first order one, but also makes the entire freeway system more stable as it collects more system information to design the metering rate. Simulation results confirmed its efficiency. When compared to the well-known ALINEA approach, this high order open-closed loop ILC approach has faster convergence speed and better robust performance.
According to the problems that are encountered in current ramp control research: uncertainty of the traffic model、difficulty in designing an adaptive ramp controller and the necessary of the coordinated algorithms, the neuro-fuzzy network method was applied to resolve the problems above and an adaptive and coordinated on-ramp strategy was designed. This approach considers both the traffic density and the on-ramp queue length together and aims at controlling both of them simultaneously. When compared to the well-known ALINEA strategy, this approach has better performance in restraining traffic flow fluctuation and queue increase. It is worthy to point out that the approach is coordinated, which can utilize the road capacity better than the single on-ramp methods.
引文
[1]Papageorgiou M, Kotsialos A. Freeway ramp metering: an overview. IEEE Transactions on Intelligent Transportation Systems, Dearborn, USA, December, 2002:271-280.
[2]Kotsialos A, Papageorgiou M. Overview of motorway network traffic control strategies. Dynamic Systems and Simulation Laboratory, Technical University of Crete, Greece:381-386.
[3]Karimi A, Hegyi A, Schutter D et al. Integration of dynamic route guidance and freeway ramp metering using model predictive control. Proceedings of the 2004 American Control Conference(ACC 2004). Boston, Massachusetts, June-July, 2004:5533-5538.
[4]陈德望，李灵犀，刘小明等．城市高速公路交通控制方法研究的回顾及展望．信息与控制，2002年8月，31(4)：341—345．
[5]王亦兵，韩曾晋，贺国光．城市高速公路交通控制综述．自动化学报，1998年7月，24(4)：484—496．
[6]Wang F Y, Mirchandani B. An Integrated Ramp Metering Control System for freeway surface street management. PARCS report, System and Engineering Department, University of Arizona, Tucson, Arizona, 1999.
[7]Hellinga B, Aerde M. Examining the Potential of Using Ramp Metering as A Component of An ATMS. Transportation Research Record, 1997, 1494:75-83.
[8]Wattleworth, J A, Berry. D S. Peak Period Control of a Freeway System-Some Theoretical investigation. HRB Rec, 1965, 89:1-25.
[9]佐佐木纲，明神证．都市高速道路网入口匝道流入控制理论．交通工学，1968，3：8—16．
[10]Papageorgiou M, Habib H S and Blosseville J M. ALINEA: A local feedback control law for on-ramp metering. Transportation research record, 1991, 3201:58-64.
[11]Kachroo P, Krishen K. System dynamics and feedback control design problem formulations for real time ramp metering. Society for Design and Process Science, Printed in the United States of America, 2000.
[12]姜紫峰，荆便顺，韩锡令．高速公路入口匝道控制的仿真研究．中国公路学报，1997年6月10(2)：83—89．
[13]Zhang H M, Jayakrishnant R, Reckerl W W. An optimization algorithm for freeway traffic control. 2001 IEEE Intelligent Transportation Systems Conference Proceedings. Oakland, USA, August 25-29, 2001:100-106.
[14]Zhang L, Levinson D. Optimal freeway ramp control without origin destination. Department of Civil Engineering, University of Minnesota. 500 Pillsbury Drive SE, Minneapolis, MN, 55455.
[15]刘智勇，万百五．高公路智能交通控制系统的建模及多层描述．公路交通科技，1998，15(1)：39—44．
[16]罗方述，潘文敏．西安-临潼高速公路递阶控制方法研究．自动化学报，1991，17(3)：363—367．
[17]Pappis C P, Mamdani E H. A fuzzy logic controller for a traffic junction. IEEE Trans on SMC. 1977, 7(I0):707-717.
[18]张立明．人工神经网络的模型及其应用．上海：复旦大学出版社，1993．
[19]刘智勇等．高公路动态系统的神经网络预测控制．信息与控制，1998，27(4)：260-266．
[20]Nguyen D H. Neural network for self learning control systems. IEEE Control System Magazine, 1990, 10(1):18-23.
[21]Parisini R, Zoppoli T. Neural network for feedback feed forward and non-linear control system. IEEE Trans on NN, 1994, 5(3):436-499.
[22]Zhang H. A local neural network controller for freeway ramp metering. The 7th IFAC IFORS Symp on Transp Syst: Theory and Applications of Advanced Technology. 1994:704-707.
[23]Hou Z S, Zhong H W, Xu J X. An iterative learning approach for local ramp metering. Proceedings of 2004 IEEE International Conference on Control Applications, Taipei, Taiwan, September 2-4, 2004:989-994.
[24]Xu J X, Xing Y F. A learning approach for freeway traffic control. 2005 International Conference on Control and Automation (ICCA2005), Budapest, Hungary, June 27-29, 2005: 654-659.
[25]Hou Z S, Xu J X. Freeway traffic density control using iterative learning control approach. 0-7803-8125-4/03/S17.00, 20031EEE:1081-1086.
[26]Papageorgiou M. Handbook of Transportation Science. chapter 8:Taffic control.
[27]陈德望，王飞跃，陈龙．基于模糊神经网络的城市高速公路入口匝道控制算法．交通运输工程学报，2003年6月，3(2)：100-105．
[28]Bogenberger K, Adolf D. Advanced coordinated traffic responsive ramp strategies. California PATH Working Paper. UCB-ITS-PWP-99-19, November, 1999:1-63.
[29]刘智勇．智能交通理论及其应用．北京：科学出版社，2003．
[30]Bien Z, Huh K. High order Iterative Learning Control Algorithm. IEEE Proc. of Control Theory and Applications. 1989, 32:105-112.
[31]Papageorgiou M, Hadjsalem H, Middleham F. ALINEAIocal ramp metering: summary of the field results. Transp. Res. Record. 1603. TRB. National Research Council, Washington D.C, 1997:90-98.
[32]丁建梅，王可崇．高速公路入口匝道交通模糊控制研究．哈尔滨工业大学学报，2002年10月．34(5)：671—674．
[33]Sasaki T. Fuzzy on-ramp control model on urban expressway and its extension. Intern Symp on Transportation and Traffic Theory, 1968, 7:377-395.
[34]Chen L L. Freeway ramp control using fuzzy set theory for inexact reasoning. Transpn, Res-B, 1990,24(1):220-226
[35]梁新荣，刘智勇，徐建闽等．高速公路多路段的模糊逻辑协调控制．华南理工大学学报，2005年9月，33(9)：35-40．
[36]李士勇．模糊控制，神经控制和智能控制．哈尔滨：哈尔滨工业大学出版社，1999．
[37]Teodorovic D, Vukadinovic K. Traffic Control and Transport Planning: A Fuzzy Sets and Neural Networks Approach (International Series in Intelligent Technologies). Kluwer Academic Publishers, 1998.
[38]Bogenberger K, Hartmut K, Stephen G. Adaptive fuzzy systems for traffic responsive and coordinated ramp metering. UCI-ITS-WP-01-08, November 2001:1-19.
[39]易继锴，侯媛彬．智能控制技术．北京：北京工业大学出版社，1999．
[40]Hussein D. Neural Network Models for Freeway Incident Detection. The 16th Conference of Australian Institutes of Transport Research. Australia, DecemberT-9, 2004:145-152.
[41]周志坚，毛宗源．一种基于遗传算法的模糊神经网络最优控制．控制理论与应用，2000年10月，17(5)：784—788．
[42]张其光，王执铨．基于遗传算法的模糊神经网络控制器设计及其稳定性分析．控制理论与应用．1999，16(5)：767-769．
[43]Blanco A, Delgado M, Pegalajar M C. A real-coded genetic algorithm for training recurrent neural networks. Neural Networks, 2001, 14:93-105.
[44]Alieva R A, Fazlollahib B J. Genetic algorithm based learning of fuzzy neural networks. Part 1: feed-forward fuzzy neural networks, Fuzzy Sets and Systems. 2001, 118:351-358.
[45]宫晓燕，汤淑明．基于非参数回归的单点短时交通流预测算法．中国公路学报，2003，16(1)：82—87．
[46]Chen D W, Gong X Y, Zhang C Q. A wireless traffic information collection system in Beijing urban freeways.Proc.of the 5th IEEE International Conference on ITS.Singapore,Sept,2002:190-195．
[47] Kotsialos A,Papageorgiou M.Efficiency and equity properties of freeway network wide ramp metering with AMOC.Transportation Research C.2004,12(6) :401-420．
[2]Kotsialos A, Papageorgiou M. Overview of motorway network traffic control strategies. Dynamic Systems and Simulation Laboratory, Technical University of Crete, Greece:381-386.
[3]Karimi A, Hegyi A, Schutter D et al. Integration of dynamic route guidance and freeway ramp metering using model predictive control. Proceedings of the 2004 American Control Conference(ACC 2004). Boston, Massachusetts, June-July, 2004:5533-5538.
[4]陈德望，李灵犀，刘小明等．城市高速公路交通控制方法研究的回顾及展望．信息与控制，2002年8月，31(4)：341—345．
[5]王亦兵，韩曾晋，贺国光．城市高速公路交通控制综述．自动化学报，1998年7月，24(4)：484—496．
[6]Wang F Y, Mirchandani B. An Integrated Ramp Metering Control System for freeway surface street management. PARCS report, System and Engineering Department, University of Arizona, Tucson, Arizona, 1999.
[7]Hellinga B, Aerde M. Examining the Potential of Using Ramp Metering as A Component of An ATMS. Transportation Research Record, 1997, 1494:75-83.
[8]Wattleworth, J A, Berry. D S. Peak Period Control of a Freeway System-Some Theoretical investigation. HRB Rec, 1965, 89:1-25.
[9]佐佐木纲，明神证．都市高速道路网入口匝道流入控制理论．交通工学，1968，3：8—16．
[10]Papageorgiou M, Habib H S and Blosseville J M. ALINEA: A local feedback control law for on-ramp metering. Transportation research record, 1991, 3201:58-64.
[11]Kachroo P, Krishen K. System dynamics and feedback control design problem formulations for real time ramp metering. Society for Design and Process Science, Printed in the United States of America, 2000.
[12]姜紫峰，荆便顺，韩锡令．高速公路入口匝道控制的仿真研究．中国公路学报，1997年6月10(2)：83—89．
[13]Zhang H M, Jayakrishnant R, Reckerl W W. An optimization algorithm for freeway traffic control. 2001 IEEE Intelligent Transportation Systems Conference Proceedings. Oakland, USA, August 25-29, 2001:100-106.
[14]Zhang L, Levinson D. Optimal freeway ramp control without origin destination. Department of Civil Engineering, University of Minnesota. 500 Pillsbury Drive SE, Minneapolis, MN, 55455.
[15]刘智勇，万百五．高公路智能交通控制系统的建模及多层描述．公路交通科技，1998，15(1)：39—44．
[16]罗方述，潘文敏．西安-临潼高速公路递阶控制方法研究．自动化学报，1991，17(3)：363—367．
[17]Pappis C P, Mamdani E H. A fuzzy logic controller for a traffic junction. IEEE Trans on SMC. 1977, 7(I0):707-717.
[18]张立明．人工神经网络的模型及其应用．上海：复旦大学出版社，1993．
[19]刘智勇等．高公路动态系统的神经网络预测控制．信息与控制，1998，27(4)：260-266．
[20]Nguyen D H. Neural network for self learning control systems. IEEE Control System Magazine, 1990, 10(1):18-23.
[21]Parisini R, Zoppoli T. Neural network for feedback feed forward and non-linear control system. IEEE Trans on NN, 1994, 5(3):436-499.
[22]Zhang H. A local neural network controller for freeway ramp metering. The 7th IFAC IFORS Symp on Transp Syst: Theory and Applications of Advanced Technology. 1994:704-707.
[23]Hou Z S, Zhong H W, Xu J X. An iterative learning approach for local ramp metering. Proceedings of 2004 IEEE International Conference on Control Applications, Taipei, Taiwan, September 2-4, 2004:989-994.
[24]Xu J X, Xing Y F. A learning approach for freeway traffic control. 2005 International Conference on Control and Automation (ICCA2005), Budapest, Hungary, June 27-29, 2005: 654-659.
[25]Hou Z S, Xu J X. Freeway traffic density control using iterative learning control approach. 0-7803-8125-4/03/S17.00, 20031EEE:1081-1086.
[26]Papageorgiou M. Handbook of Transportation Science. chapter 8:Taffic control.
[27]陈德望，王飞跃，陈龙．基于模糊神经网络的城市高速公路入口匝道控制算法．交通运输工程学报，2003年6月，3(2)：100-105．
[28]Bogenberger K, Adolf D. Advanced coordinated traffic responsive ramp strategies. California PATH Working Paper. UCB-ITS-PWP-99-19, November, 1999:1-63.
[29]刘智勇．智能交通理论及其应用．北京：科学出版社，2003．
[30]Bien Z, Huh K. High order Iterative Learning Control Algorithm. IEEE Proc. of Control Theory and Applications. 1989, 32:105-112.
[31]Papageorgiou M, Hadjsalem H, Middleham F. ALINEAIocal ramp metering: summary of the field results. Transp. Res. Record. 1603. TRB. National Research Council, Washington D.C, 1997:90-98.
[32]丁建梅，王可崇．高速公路入口匝道交通模糊控制研究．哈尔滨工业大学学报，2002年10月．34(5)：671—674．
[33]Sasaki T. Fuzzy on-ramp control model on urban expressway and its extension. Intern Symp on Transportation and Traffic Theory, 1968, 7:377-395.
[34]Chen L L. Freeway ramp control using fuzzy set theory for inexact reasoning. Transpn, Res-B, 1990,24(1):220-226
[35]梁新荣，刘智勇，徐建闽等．高速公路多路段的模糊逻辑协调控制．华南理工大学学报，2005年9月，33(9)：35-40．
[36]李士勇．模糊控制，神经控制和智能控制．哈尔滨：哈尔滨工业大学出版社，1999．
[37]Teodorovic D, Vukadinovic K. Traffic Control and Transport Planning: A Fuzzy Sets and Neural Networks Approach (International Series in Intelligent Technologies). Kluwer Academic Publishers, 1998.
[38]Bogenberger K, Hartmut K, Stephen G. Adaptive fuzzy systems for traffic responsive and coordinated ramp metering. UCI-ITS-WP-01-08, November 2001:1-19.
[39]易继锴，侯媛彬．智能控制技术．北京：北京工业大学出版社，1999．
[40]Hussein D. Neural Network Models for Freeway Incident Detection. The 16th Conference of Australian Institutes of Transport Research. Australia, DecemberT-9, 2004:145-152.
[41]周志坚，毛宗源．一种基于遗传算法的模糊神经网络最优控制．控制理论与应用，2000年10月，17(5)：784—788．
[42]张其光，王执铨．基于遗传算法的模糊神经网络控制器设计及其稳定性分析．控制理论与应用．1999，16(5)：767-769．
[43]Blanco A, Delgado M, Pegalajar M C. A real-coded genetic algorithm for training recurrent neural networks. Neural Networks, 2001, 14:93-105.
[44]Alieva R A, Fazlollahib B J. Genetic algorithm based learning of fuzzy neural networks. Part 1: feed-forward fuzzy neural networks, Fuzzy Sets and Systems. 2001, 118:351-358.
[45]宫晓燕，汤淑明．基于非参数回归的单点短时交通流预测算法．中国公路学报，2003，16(1)：82—87．
[46]Chen D W, Gong X Y, Zhang C Q. A wireless traffic information collection system in Beijing urban freeways.Proc.of the 5th IEEE International Conference on ITS.Singapore,Sept,2002:190-195．
[47] Kotsialos A,Papageorgiou M.Efficiency and equity properties of freeway network wide ramp metering with AMOC.Transportation Research C.2004,12(6) :401-420．