城市交通信号智能控制系统设计及仿真
|
摘要
随着经济的快速发展,城市交通问题日趋严重。解决好交叉路口的交通信号控制问题是整个城市交通管理的关键所在。因此我们应该充分利用现有道路资源,通过科学合理的交通控制手段,最大限度地提高交叉口的通行能力,减少延迟时间。
本文以交通工程基本理论为基础,应用神经网络和模糊控制理论,对单交叉和多交叉路口交通信号控制问题进行分析和研究。归纳起来,本文所做的主要工作有：
(1)经典交通信号模糊控制器的模糊规则是预先人为确定的,不能随着交通流的变化而动态改变,不能保证模糊控制系统获得很好的性能,为此设计一种改进的模糊控制算法,根据路口交通流的变化,利用修正因子动态调整模糊规则,改变控制器的输出。数值实验表明,带有自修正因子的模糊控制方法在车辆平均延误时间上较之定时控制方法及经典模糊控制方法优越。
(2)针对人、车混合流现象,对四相位三车道十字交叉口的交通信号进行模糊配时。首先根据路口的交通流状况,合理地设置相位、相序,采用“迟起”、“早断”方式对行人交通信号进行配时；然后,考虑到行人控制信号对机动车通行的影响,设计模糊控制算法,对机动车信号进行配时。数值实验表明,考虑行人因素的模糊控制方法在减少车辆平均延误时间上较之传统定时控制方法大为改善。
(3)以贵阳市南明区的主要交叉路口的交通信号控制问题为背景,提出了多交叉路口的信号两级模糊控制器模型。模型分为用于控制单个交叉路口的第一级模糊控制单元和路网的第二级协调控制单元。前者以车辆平均延误时间为评价指标,综合考虑当前相、后继相的车辆排队长度,以此决定绿时分配；后者将BP神经网络与模糊控制器相结合,协调和平衡各路口之间的车流。
With the rapid development of economy, the problem on city traffic has become more and more serious. Solving the problem of the cross traffic signal control in a better way is the key for managers to manage and control the traffic of the whole city. Thus, we should make the better use the resources of existing roads, increase traffic capacity, and reduce delay time as much as possible by means of scientific and reasonable traffic control measures.
Based on the traffic engineer basic theory, we have studied the problem of traffic signal control for some single intersections and multiple ones in terms of the neural network and fuzzy control theory in this dissertation. Summarily, the main work consists of the three sections as follows:
(A) Conventional traffic signal fuzzy controllers use man-made control rules. However, the fuzzy rules can't change dynamically different traffic flows. Also, the approach does not guarantee that fuzzy control system results in the satisfactory system performance. Therefore, an improved fuzzy control algorithm is proposed to dynamically adjust the fuzzy rules. In this algorithm, the self-modifying factors adjust the fuzzy rules relying upon the conditions of traffic flows and change the output of the fuzzy controller. Experimental results show the fuzzy control method with self-modifying factors is more effective than the timed control method and conventional fuzzy control with the aspect of reducing average vehicle delay time.
(B) This section studies the problem on traffic signal fuzzy timing of a single intersection composed of four phases associated to three traffic lanes, depending upon mixed traffic flows caused by pedestrians and vehicles. Firstly, rational phases and phase orders are designed to carry out pedestrian traffic signal timing by means of different fashions of "late startup" and "early stop", relying upon the conditions of traffic flows. Secondly, a fuzzy control algorithm is applied to the vehicle traffic signal timing, while considering how pedestrian exclusive control signals affect vehicle signals. Experimental results show the fuzzy control method is more effective than the timed control method in reducing average vehicle delay time.
(C) This section investigates a two-stage fuzzy traffic controller model for the major intersections of Nanming District of Guiyang City. The system is made of two control units, one is the basic control unit for each single intersection, and the other is used for communicating the traffic network. The first control unit estimates the average vehicle delay time, in which each phase green light time is adjusted dynamically according to the current phase and the next phase waiting queues; on the other hand, in order to harmonize and balance the traffic flows of each intersection, the second control unit integrate BP neural network with the fuzzy controller.
本文以交通工程基本理论为基础,应用神经网络和模糊控制理论,对单交叉和多交叉路口交通信号控制问题进行分析和研究。归纳起来,本文所做的主要工作有：
(1)经典交通信号模糊控制器的模糊规则是预先人为确定的,不能随着交通流的变化而动态改变,不能保证模糊控制系统获得很好的性能,为此设计一种改进的模糊控制算法,根据路口交通流的变化,利用修正因子动态调整模糊规则,改变控制器的输出。数值实验表明,带有自修正因子的模糊控制方法在车辆平均延误时间上较之定时控制方法及经典模糊控制方法优越。
(2)针对人、车混合流现象,对四相位三车道十字交叉口的交通信号进行模糊配时。首先根据路口的交通流状况,合理地设置相位、相序,采用“迟起”、“早断”方式对行人交通信号进行配时；然后,考虑到行人控制信号对机动车通行的影响,设计模糊控制算法,对机动车信号进行配时。数值实验表明,考虑行人因素的模糊控制方法在减少车辆平均延误时间上较之传统定时控制方法大为改善。
(3)以贵阳市南明区的主要交叉路口的交通信号控制问题为背景,提出了多交叉路口的信号两级模糊控制器模型。模型分为用于控制单个交叉路口的第一级模糊控制单元和路网的第二级协调控制单元。前者以车辆平均延误时间为评价指标,综合考虑当前相、后继相的车辆排队长度,以此决定绿时分配；后者将BP神经网络与模糊控制器相结合,协调和平衡各路口之间的车流。
With the rapid development of economy, the problem on city traffic has become more and more serious. Solving the problem of the cross traffic signal control in a better way is the key for managers to manage and control the traffic of the whole city. Thus, we should make the better use the resources of existing roads, increase traffic capacity, and reduce delay time as much as possible by means of scientific and reasonable traffic control measures.
Based on the traffic engineer basic theory, we have studied the problem of traffic signal control for some single intersections and multiple ones in terms of the neural network and fuzzy control theory in this dissertation. Summarily, the main work consists of the three sections as follows:
(A) Conventional traffic signal fuzzy controllers use man-made control rules. However, the fuzzy rules can't change dynamically different traffic flows. Also, the approach does not guarantee that fuzzy control system results in the satisfactory system performance. Therefore, an improved fuzzy control algorithm is proposed to dynamically adjust the fuzzy rules. In this algorithm, the self-modifying factors adjust the fuzzy rules relying upon the conditions of traffic flows and change the output of the fuzzy controller. Experimental results show the fuzzy control method with self-modifying factors is more effective than the timed control method and conventional fuzzy control with the aspect of reducing average vehicle delay time.
(B) This section studies the problem on traffic signal fuzzy timing of a single intersection composed of four phases associated to three traffic lanes, depending upon mixed traffic flows caused by pedestrians and vehicles. Firstly, rational phases and phase orders are designed to carry out pedestrian traffic signal timing by means of different fashions of "late startup" and "early stop", relying upon the conditions of traffic flows. Secondly, a fuzzy control algorithm is applied to the vehicle traffic signal timing, while considering how pedestrian exclusive control signals affect vehicle signals. Experimental results show the fuzzy control method is more effective than the timed control method in reducing average vehicle delay time.
(C) This section investigates a two-stage fuzzy traffic controller model for the major intersections of Nanming District of Guiyang City. The system is made of two control units, one is the basic control unit for each single intersection, and the other is used for communicating the traffic network. The first control unit estimates the average vehicle delay time, in which each phase green light time is adjusted dynamically according to the current phase and the next phase waiting queues; on the other hand, in order to harmonize and balance the traffic flows of each intersection, the second control unit integrate BP neural network with the fuzzy controller.
引文
[1]于德新,杨兆升,陈林.城市主干道的多路口模糊协调控制,吉林大学学报(工学版),2006,(36)：23-28
[2]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,2002：173-180
[3]王品,李玉兰,蔡自兴,等基于遗传算法的模糊系统优化设计方法,控制理论与应用.1999，16(5)：699-704.
[4]王爱勤,曾蔚.两种基于遗传算法的交通量预测方法,公路工程与运输,No.129：
[5]史忠科,黄辉先,曲仕茹,陈小峰.交通控制系统导论.科学出版社,2003：1,3-5
[6]冯兴乐,丁岳伟.多交叉路口车流模拟软件的设计,上海理工大学学报,22(4),2000
[7]许伦辉,衷路生,徐建闽基于神经网络实现交叉口多相位模糊控制。中南公路工程,2004,29(2)：9-12.
[8]朱中.杨兆升.实时交通流量人工神经网络顶测模型,中国公路学报,1998,11(4)：89-92.
[9]刘智勇,朱劲,李秀平,尹征琦.单交叉口的多相位模糊控制.信息与控制.1999,28(6)：453-458
[10]刘智勇等.基于模糊控制的多功能交通控制系统。98中国智能自动化学术会议论文集.1998：381-386
[11]李建斌,高成修.城市道路网络多交叉路口交通信号实时优化控制模型与算法,系统工程,22(10),2004,
[12]陈森发,陈洪,徐吉谦.城市单路口交通的两级模糊控制及其仿真.系统仿真学报。1998,10(2)：35-40
[13]张良智,姜华平.基于遗传算法的交通流量组合预测.山东交通学院学报.2005(6)。
[14]陈建平,房海蓉,胡准庆.城市交通的智能控制系统综述,北京汽车.2001,4：10-13.
[15]陈洪,陈森发.单路口交通实时模糊控制的一种方法.信息与控制。1997,26(3)：229-233
[16]陈丕刚.多交叉口信号模糊线控制模型研究,交通运输系统工程与信,息,3(2),2003
[17]张国伍.智能交通系统工程导论.电子工业出版社,2003,7-9
[18]杨佩昆.智能交通运输系统体系结构.上海：同济大学出版社,2001
[19]杨兆升,谷远利.实时动态交通流预测模型研究.公路交通科技,1998,15(3)：4-7.
[20]杨晓光,马万经,林瑜.“两相位信号控制交叉口行人专用相位设置条件研究”.公路交通科技,1002-0268(2005)01-0127-05
[21]杨立才,贾磊,王红.双交叉口两级模糊协调控制算法的研究,系统工程理论与实践,第6期,2005.
[22]金春霞,王慧。基于路网的城域多路口交通模糊控制,公路交通科技,19(3),2002.
[23]杨兆升,谷远利.实时动态交通流预测模型研究.公路交通科技,1998,15(3)：4-7.
[24]尚宁等。基于BP神经网络的路口短时交通流预测方法,计算机应用与软件,1992,23(2)：23-27。
[25]姚亚夫,曹锋.多路口模糊控制及其仿真研究,机械工程与自动化,No.3,2006
[26]徐良杰,王炜.“信号交叉口行人过街时间模型”.交通运输工程学报,167121637(2005)0120111205.
[27]徐建闽,舒宁,尹宏宾.新型模糊控制算法在交叉口信号控制中的应用.华南理工大学学报.2000,28(6)：1-5
[28]徐良杰,王炜.“信号交叉口行人过街时间模型”.交通运输工程学报,167121637(2005)0120111205.
[29]徐良杰等,“信号交叉口左转非机动车影响分析”.中国公路学报,2006(01)。
[30]徐冬玲,方建安,邵世煌.交通系统的模糊控制及其神经网络实现.信息与控制.1992,21(2)：7 4.78
[31]黄卫,陈里得.智能运输系统(ITS)概论.北京：人民交通出版社,2001
[32]廉小亲.模糊控制技术,中国电力出版社,2003.
[33]Abdulhai, B., Porwal, H., Recker, W.,1999."Short-term freeway traffic flow prediction using genetically-optimized time-delay-based neural networks". Transportation Research Board, 1999.
[34]Brian L. Smith, Michael J. Demetsky. "Traffic Flow Forecasting:Comparition of Modeling Approachies."Journal of Transportation Engineering.1997, (7):261-266.
[35]C.P.Pappis, E.H.Mamdani. "A fuzzy logic controller for a traffic junction". IEEE Trans on System, Man, and Cybernetics, SMC-7,1977, (10):707-717.
[36]C. Ledoux. "An ubran traffic flow model integrating neural networks,"Transpn Res.-C, 5(5):286-300,1997.
[37]Chih-Hsun Chou.Jen-ChaoTeng. "A Fuzzy Logic Controller for Traffic Junction Signals". Information Sciences,2002, (143):73-97.
[38]D.C.Chin,R.H.Smith. A neural network application in signal timing control. Proceedings IEEE International Conference on Neural Networks,1996,4:2101-2106
[39]Dougherty, M., Cobbett,M. "Short-term inter-urban traffic forecasts using neural networks." International Journal of Forecasting,1997, (13):21-31.
[40]E. I. Vlahogianni, M. G. Karlaftis, J. C. Golias. "Optimized and meta-optimized neural networks for short-term traffic flow prediction:A genetic approach,"Transportation Research Part C 13(2005)211-234.
[41]H.Hellendoorn, R.Baudrexl. Fuzzy neural traffic control and forecasting. Proceedings of IEEE International Conference on Fuzzy Systems.1995,4:2187-2194
[42]J Favilla, AMachion, Fgomide. FuzzyTrafic Control:Adaptive Strategies. Second IEEE International conference on Fuzzy Systems Ⅱ (1993):506-511.
[43]J Haijun Gao, Lingxi Li, Rui Liu. "Changeable phases signal control of an isolated intersection". IEEE SMC[C],2002
[44]J Licai Yang, Lie Jiai, Hong Wang. "Network with genetic algorithm and its applications for traffic flow forecasting[A]". Pro of the 5th World Congress on Intelligent Control Automation[C]. Hongzhou China,2004, (6):5330-5333.
[45]L.Fortuna, L.occhipinti,C.Vinci, M.G.Xibilia. A neuro-fuzzy model of urban traffic.Proceedings of the 37th Midwest Symposium on Circuits and Systems,1994,1:603-606
[46]Marzuki Khalid,See Chin Liang and Rubiyah Yusof."Control of a Complex Traffic Junction using Fuzzy Inference."Centre for Artificial Intelligence and Robotics.
[47]Nakatsuyama M. H Nagahashi, M Nishizura. Fuzzy Logic Phase Controller for Traffic Junctions in the One-Way Arterial Road. In:9IFAC-World Congress. Preprints. Budapest: 1984:13-18
[48]N. Messai, P. Thomasa, D. Lefebvre, etal. "Optimal neural networks architectures for the flow-density relationships of traffic models," Mathematics and Computers in Simulation 60 (2002) 401-409.
[49]Panlo Cortez, Miguel Rocha. A Neural Network Based Time Series Forecasting System. IEEE International Conference on Neural Networks,1995,5:2689-2693.
[50]P. Akritas, P.G. Akishin, I. Antoniou, etal. "Nonlinear analysis of network traffic", Chaos, Solitons and Fractals 14 (2002) 595-606.
[51]R.E.KALMAN "A New Approach to Linear Filtering and Prediction Problems Transactions of the ASME Joumal of Basic Engineering,"82(Series D); 35-45.1960 by ASME
[52]Tan Kok Khiang, Marzuki Khalid, Rubiyah Yusof,"Intelligent Traffic Lights Control By Fuzzy Logic",Malaysian Journal of Computer Science,1995,(11).
[53]Tim Hill, Marcus Ocormor, Willam Remus. Neural Network Model for Time Series Forecasts. Management Science,1996,142(7):1082-1092.
[54]Weiss.N eural networks and evolutionary computation.Ⅰ. Hybrid approaches in artificial in telligence.Evolutionary Computation,1994.Proceedings of the First IEEE Conference on Computational Intelligence,1994,1:268-272
[2]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,2002：173-180
[3]王品,李玉兰,蔡自兴,等基于遗传算法的模糊系统优化设计方法,控制理论与应用.1999，16(5)：699-704.
[4]王爱勤,曾蔚.两种基于遗传算法的交通量预测方法,公路工程与运输,No.129：
[5]史忠科,黄辉先,曲仕茹,陈小峰.交通控制系统导论.科学出版社,2003：1,3-5
[6]冯兴乐,丁岳伟.多交叉路口车流模拟软件的设计,上海理工大学学报,22(4),2000
[7]许伦辉,衷路生,徐建闽基于神经网络实现交叉口多相位模糊控制。中南公路工程,2004,29(2)：9-12.
[8]朱中.杨兆升.实时交通流量人工神经网络顶测模型,中国公路学报,1998,11(4)：89-92.
[9]刘智勇,朱劲,李秀平,尹征琦.单交叉口的多相位模糊控制.信息与控制.1999,28(6)：453-458
[10]刘智勇等.基于模糊控制的多功能交通控制系统。98中国智能自动化学术会议论文集.1998：381-386
[11]李建斌,高成修.城市道路网络多交叉路口交通信号实时优化控制模型与算法,系统工程,22(10),2004,
[12]陈森发,陈洪,徐吉谦.城市单路口交通的两级模糊控制及其仿真.系统仿真学报。1998,10(2)：35-40
[13]张良智,姜华平.基于遗传算法的交通流量组合预测.山东交通学院学报.2005(6)。
[14]陈建平,房海蓉,胡准庆.城市交通的智能控制系统综述,北京汽车.2001,4：10-13.
[15]陈洪,陈森发.单路口交通实时模糊控制的一种方法.信息与控制。1997,26(3)：229-233
[16]陈丕刚.多交叉口信号模糊线控制模型研究,交通运输系统工程与信,息,3(2),2003
[17]张国伍.智能交通系统工程导论.电子工业出版社,2003,7-9
[18]杨佩昆.智能交通运输系统体系结构.上海：同济大学出版社,2001
[19]杨兆升,谷远利.实时动态交通流预测模型研究.公路交通科技,1998,15(3)：4-7.
[20]杨晓光,马万经,林瑜.“两相位信号控制交叉口行人专用相位设置条件研究”.公路交通科技,1002-0268(2005)01-0127-05
[21]杨立才,贾磊,王红.双交叉口两级模糊协调控制算法的研究,系统工程理论与实践,第6期,2005.
[22]金春霞,王慧。基于路网的城域多路口交通模糊控制,公路交通科技,19(3),2002.
[23]杨兆升,谷远利.实时动态交通流预测模型研究.公路交通科技,1998,15(3)：4-7.
[24]尚宁等。基于BP神经网络的路口短时交通流预测方法,计算机应用与软件,1992,23(2)：23-27。
[25]姚亚夫,曹锋.多路口模糊控制及其仿真研究,机械工程与自动化,No.3,2006
[26]徐良杰,王炜.“信号交叉口行人过街时间模型”.交通运输工程学报,167121637(2005)0120111205.
[27]徐建闽,舒宁,尹宏宾.新型模糊控制算法在交叉口信号控制中的应用.华南理工大学学报.2000,28(6)：1-5
[28]徐良杰,王炜.“信号交叉口行人过街时间模型”.交通运输工程学报,167121637(2005)0120111205.
[29]徐良杰等,“信号交叉口左转非机动车影响分析”.中国公路学报,2006(01)。
[30]徐冬玲,方建安,邵世煌.交通系统的模糊控制及其神经网络实现.信息与控制.1992,21(2)：7 4.78
[31]黄卫,陈里得.智能运输系统(ITS)概论.北京：人民交通出版社,2001
[32]廉小亲.模糊控制技术,中国电力出版社,2003.
[33]Abdulhai, B., Porwal, H., Recker, W.,1999."Short-term freeway traffic flow prediction using genetically-optimized time-delay-based neural networks". Transportation Research Board, 1999.
[34]Brian L. Smith, Michael J. Demetsky. "Traffic Flow Forecasting:Comparition of Modeling Approachies."Journal of Transportation Engineering.1997, (7):261-266.
[35]C.P.Pappis, E.H.Mamdani. "A fuzzy logic controller for a traffic junction". IEEE Trans on System, Man, and Cybernetics, SMC-7,1977, (10):707-717.
[36]C. Ledoux. "An ubran traffic flow model integrating neural networks,"Transpn Res.-C, 5(5):286-300,1997.
[37]Chih-Hsun Chou.Jen-ChaoTeng. "A Fuzzy Logic Controller for Traffic Junction Signals". Information Sciences,2002, (143):73-97.
[38]D.C.Chin,R.H.Smith. A neural network application in signal timing control. Proceedings IEEE International Conference on Neural Networks,1996,4:2101-2106
[39]Dougherty, M., Cobbett,M. "Short-term inter-urban traffic forecasts using neural networks." International Journal of Forecasting,1997, (13):21-31.
[40]E. I. Vlahogianni, M. G. Karlaftis, J. C. Golias. "Optimized and meta-optimized neural networks for short-term traffic flow prediction:A genetic approach,"Transportation Research Part C 13(2005)211-234.
[41]H.Hellendoorn, R.Baudrexl. Fuzzy neural traffic control and forecasting. Proceedings of IEEE International Conference on Fuzzy Systems.1995,4:2187-2194
[42]J Favilla, AMachion, Fgomide. FuzzyTrafic Control:Adaptive Strategies. Second IEEE International conference on Fuzzy Systems Ⅱ (1993):506-511.
[43]J Haijun Gao, Lingxi Li, Rui Liu. "Changeable phases signal control of an isolated intersection". IEEE SMC[C],2002
[44]J Licai Yang, Lie Jiai, Hong Wang. "Network with genetic algorithm and its applications for traffic flow forecasting[A]". Pro of the 5th World Congress on Intelligent Control Automation[C]. Hongzhou China,2004, (6):5330-5333.
[45]L.Fortuna, L.occhipinti,C.Vinci, M.G.Xibilia. A neuro-fuzzy model of urban traffic.Proceedings of the 37th Midwest Symposium on Circuits and Systems,1994,1:603-606
[46]Marzuki Khalid,See Chin Liang and Rubiyah Yusof."Control of a Complex Traffic Junction using Fuzzy Inference."Centre for Artificial Intelligence and Robotics.
[47]Nakatsuyama M. H Nagahashi, M Nishizura. Fuzzy Logic Phase Controller for Traffic Junctions in the One-Way Arterial Road. In:9IFAC-World Congress. Preprints. Budapest: 1984:13-18
[48]N. Messai, P. Thomasa, D. Lefebvre, etal. "Optimal neural networks architectures for the flow-density relationships of traffic models," Mathematics and Computers in Simulation 60 (2002) 401-409.
[49]Panlo Cortez, Miguel Rocha. A Neural Network Based Time Series Forecasting System. IEEE International Conference on Neural Networks,1995,5:2689-2693.
[50]P. Akritas, P.G. Akishin, I. Antoniou, etal. "Nonlinear analysis of network traffic", Chaos, Solitons and Fractals 14 (2002) 595-606.
[51]R.E.KALMAN "A New Approach to Linear Filtering and Prediction Problems Transactions of the ASME Joumal of Basic Engineering,"82(Series D); 35-45.1960 by ASME
[52]Tan Kok Khiang, Marzuki Khalid, Rubiyah Yusof,"Intelligent Traffic Lights Control By Fuzzy Logic",Malaysian Journal of Computer Science,1995,(11).
[53]Tim Hill, Marcus Ocormor, Willam Remus. Neural Network Model for Time Series Forecasts. Management Science,1996,142(7):1082-1092.
[54]Weiss.N eural networks and evolutionary computation.Ⅰ. Hybrid approaches in artificial in telligence.Evolutionary Computation,1994.Proceedings of the First IEEE Conference on Computational Intelligence,1994,1:268-272