基于PSO-SVM高速公路交通事件检测算法的分析与研究
|
摘要
近些年,国家大力发展交通运输事业,高速公路领域发展十分迅猛,到2012年底,我国高速公路通车里程达到了96000公里,超越了美国成为了世界最大规模的高速公路系统,但同时也因为高速路车流量的增大,交通拥挤、事故发生的几率增大,极大的影响了高速公路的通行能力,降低了运行效率,甚至造成人员伤亡财产损失等严重后果。如何快速准确检测出交通事件,尽可能做到有效的改善高速公路通行能力,是目前普遍关注的一个问题。其中高速公路交通事件检测算法对交通事件的检测效率有着直接影响,因此研究高速公路交通事件检测算法有着十分重要的意义。
本文说明了高速路交通流特性以及交通事件的检测原理,通过分析支持向量机应用到高速路交通事件检测中的可行性,设计了基于SVM-AID模型,针对SVM参数选择上过于依赖人为经验等问题,采用PSO优化算法对支持向量机参数进行选优,并且对PSO算法进行了改进,通过调整惯性权重、动态选择加速常数等改进策略,取得了不错的检测效果。其次,本论文在第五章中将混沌与粒子群优化算法相结合,利用混沌具有的遍历性、随机性等特点,能够更好的跳出局部最小,达到全局最优；构建出了基于CPSO-SVM的高速公路交通事件检测模型,利用该模型进行交通事件检测。
本文的实验数据来自于1-880交通流数据库,分别构建了训练数据集和测试数据集作为论文的实验数据。仿真软件采用MATLAB7.0。通过对实验结果的分析得知,采用混沌优化的粒子群优化支持向量机算法在具有较高的分类精度、检测率,并且误检率较低,能够在高速路交通事件检测中取得很好的检测效果。
In recent years, the state has vigorously the development of the transportation industry, the highway is very rapid developments in the field, by the end of2012, Chinese highway mileage has reached96,000kilometers, beyond the United States to become the world's largest highway system. Highway development has brought many economic benefits for the community, but also because of highway traffic volume increases, traffic congestion, increased the chance of accidents, great highway capacity, reduce operating efficiency, or even cause casualties and property damage and other serious consequences. How to quickly and accurately detect the traffic incident as much as possible to improve highway capacity is the current issue of universal concern. Quality of highway traffic incident detection algorithm directly affect the efficiency of the traffic incident detection, and therefore of great significance to study the highway traffic incident detection algorithm.
Highway traffic flow characteristics as well as the basic principles of the traffic incident detection analysis, by highway traffic incident detection in adaptive analysis applied to the support vector machine, to build a model based on SVM-AID, simulation experiment for SVM based on experience, the choice of the model parameters using PSO optimization algorithm and PSO algorithm has been improved by adjusting the inertia weight dynamically select the acceleration constant improvement strategies, and achieved good detection effect. Secondly, in this paper, in the fifth chapter chaotic combination with particle swarm optimization algorithm, using the chaos's traversal property, randomness and other characteristics, and better able to escape from the local minimum, the globally optimal; build out based on CPSO-SVM model of freeway traffic incident detection, traffic incident detection using the model.
The experimental data from1-880traffic flow database build a training data set and test data set of experimental data as a thesis. Simulation software uses MATLAB7.0. Learned through the analysis of the experimental results, the particle swarm optimization chaos optimization support vector machine algorithm has higher classification accuracy in detection rate and low false detection rate, good highway traffic incident detection results.
本文说明了高速路交通流特性以及交通事件的检测原理,通过分析支持向量机应用到高速路交通事件检测中的可行性,设计了基于SVM-AID模型,针对SVM参数选择上过于依赖人为经验等问题,采用PSO优化算法对支持向量机参数进行选优,并且对PSO算法进行了改进,通过调整惯性权重、动态选择加速常数等改进策略,取得了不错的检测效果。其次,本论文在第五章中将混沌与粒子群优化算法相结合,利用混沌具有的遍历性、随机性等特点,能够更好的跳出局部最小,达到全局最优；构建出了基于CPSO-SVM的高速公路交通事件检测模型,利用该模型进行交通事件检测。
本文的实验数据来自于1-880交通流数据库,分别构建了训练数据集和测试数据集作为论文的实验数据。仿真软件采用MATLAB7.0。通过对实验结果的分析得知,采用混沌优化的粒子群优化支持向量机算法在具有较高的分类精度、检测率,并且误检率较低,能够在高速路交通事件检测中取得很好的检测效果。
In recent years, the state has vigorously the development of the transportation industry, the highway is very rapid developments in the field, by the end of2012, Chinese highway mileage has reached96,000kilometers, beyond the United States to become the world's largest highway system. Highway development has brought many economic benefits for the community, but also because of highway traffic volume increases, traffic congestion, increased the chance of accidents, great highway capacity, reduce operating efficiency, or even cause casualties and property damage and other serious consequences. How to quickly and accurately detect the traffic incident as much as possible to improve highway capacity is the current issue of universal concern. Quality of highway traffic incident detection algorithm directly affect the efficiency of the traffic incident detection, and therefore of great significance to study the highway traffic incident detection algorithm.
Highway traffic flow characteristics as well as the basic principles of the traffic incident detection analysis, by highway traffic incident detection in adaptive analysis applied to the support vector machine, to build a model based on SVM-AID, simulation experiment for SVM based on experience, the choice of the model parameters using PSO optimization algorithm and PSO algorithm has been improved by adjusting the inertia weight dynamically select the acceleration constant improvement strategies, and achieved good detection effect. Secondly, in this paper, in the fifth chapter chaotic combination with particle swarm optimization algorithm, using the chaos's traversal property, randomness and other characteristics, and better able to escape from the local minimum, the globally optimal; build out based on CPSO-SVM model of freeway traffic incident detection, traffic incident detection using the model.
The experimental data from1-880traffic flow database build a training data set and test data set of experimental data as a thesis. Simulation software uses MATLAB7.0. Learned through the analysis of the experimental results, the particle swarm optimization chaos optimization support vector machine algorithm has higher classification accuracy in detection rate and low false detection rate, good highway traffic incident detection results.
引文
[1]翟润平,周彤梅.道路交通控制原理及应用[M].中国人民公安大学出版社,2002.
[2]张凯.高速公路交通事件检测与仿真研究[D].长安大学硕士论文,2009.
[3]陈斌.高速公路交通意外事件管理关键技术[M].西南交通大学出版社,2007.
[4]蔡志理.高速公路交通事件检测及交通疏导技术研究[D].吉林大学博士论文,2007。
[5]龚炯.基于SVM的高速公路交通事件检测研究[D].西南交通大学硕士论文,2010.
[6]陈祥梅.GA-SVM在高速公路交通事件检测中的应用研究.西南交通大学硕士论文,2009.
[7]Persaud B.N, Hall F.L.Catastrophe theory and patterns in 30 second freeway traffic data implications for incident detection. Transportation research. Part A: general,1989,23(2):103-113.
[8]Martin P.T., Perrin H.J., Hansen B.G. Incident detection algorithm evaluation[J]. University of Utah, March,2001:1-29.
[9]张敬磊,王晓原.交通事件检测算法研究进展[J].武汉理工大学学报：交通科学与工程版.2005,29(002)：215-218.
[10]Levin M., Krause G.M. Incident detection:A Bayesian approach^J].1978.
[11]Ahmed M, Cook A. Analysis of Freeway Traffic Time Series Data by Using Box-Jenkins Techniques." [J].Transportation Research Board,1979,722:1-9.
[12]Chew R.L.Ritchie5.Automatic Detection of Lane-blocking Freeway Incidents Using Artificial Neural Networks[J].Transportation Research Part C,1995,3(6): 371-388.
[13]Hsiao C.H. Lin C. Application of Fuzzy Logical and Neural Networks to Automatically Detect Freeway Traffic Incident[J]. Journal of Transportation Engineering,1993,125(5).
[14]Chang E. C, Wang S. H. Improved Freeway Incident Detection using Fuzzy Set Theory. Transportation Research Board.1994,21 (7):75-82
[15]Chew Stephen G Ritchie S. Neural Network Models for Automated Detection of Non-Recurring Congestion [J].1991:10-16.
[16]Li Y., McDonald M. Motorway incident detection using Probe vehicles[J].Transport,2005,158(1):11-15.
[17]Yuan F, Cheu R.L. Incident detection using support vector machines [J].Transporta tion Research Part C,2003,11(3-4):309-328.
[18]李文江,荆便顺,杨光王.基于小波分析的事件检测算法[J].西安公路交通大学学报.1997,17(2B)：134-138
[19]姜紫峰,刘小坤.基于神经网络的交通事件检测算法[J].西安公路交通大学学报.2001
[20]姜桂艳,温慧敏,杨兆升.高速公路交通事件自动检测系统与算法设计[J].交通运输Ⅰ程学报,2001,1(001)：77-81.
[21]温娟,贺国光.基于粗集理论预处理数据的神经网络交通事件自动检测算法《交通运输系统工程与信息》,2004,4：54-59.
[22]姚智胜,邵春福.基于V—支持向量分类机的交通事件检测方法研究[J].ITS通讯,2005,7(004)：38-41.
[23]梁新荣,刘智勇.基于支持向量机的高速公路事件检测[J].计算机工程与应用,2006(16)：213-218.
[24]蔡志理,姜桂艳.多SVM分类器融合技术在高速公路AID中的应用[C].第27届中国控制会议,中国昆明,2008(EL ISTP收录)
[25]朱红斌.基于Bagging算法和遗传神经网络的交通事件检测[J].计算机应用与软件.2010,1(27)：234-236
[26]周林英.基于支持向量机的高速公路事件检测算法[D].长安大学硕士论文,2009.
[27]姜卉.基于FA-SVM的高速公路交通事件自动检测方法研究.吉林大学硕士论文,2011.
[28]刘伟铭.高速公路系统控制方法[M].人民交通出版社,1998.
[29]覃频频.基于信息融合的高速公路事件检测建模与仿真[D].西南交通大学博士论文,2007.
[30]朱焱.基于AdaBoost-LMBP的高速公路交通事件检测算法研究[D].西南交大硕士论文,2012.
[31]张敬磊,王晓原.交通事件检测算法研究进展[J].武汉理工大学学报：交通科学与工程版,2005,29(002)：215-218.
[32]温慧敏,杨兆升.交通事件检测技术的进展研究[J].交通运输系统工程与信息.2005,5(001)：25-28
[33]Chen S, Wang W, Van Zuylen H. Construct support vector machine ensemble to detect traffic incident. Expert Systems with Applications,2009
[34]Vapnik V. N.,张学工.统计学习理论的本质[J].The nature of Statistical learning theory. NY:Springer Verlag,1995.
[35]Vapnik VN. The nature of statistical learning theory[M]. Springer Verlag,2000.
[36]边肇棋.模式识别(第二版)[M].清华大学出版社,2000.
[37]白鹏,张喜斌,张斌.支持向量机理论及工程应用实例[M].西安电子科技大学出版社,2007.
[38]]Boser B.E., Guyon I.M.,Vapnik V.N. A training algorithm for Optimal margin Classifiers,1992[C].ACM New York, NY, USA.
[39]祁亨年.支持向量机及其应用研究综述[J].计算机工程,2004,30(010)：6-9.
[40]王琪.基于神经网络和支持向量机的高速公路交通事件检测[D].西南交通大学硕士论文,2006.
[41]Hsu C.W.,Chang C.C., Lin C.J. A practical guide to support vector classification[Z]. Citeseer,2003.
[42]Karl Petty, The Analysis Software for the FSP Project [OL], http://ipa. eecs. berkeley. edu/-pettyk/FSP/.2005-01-01
[43]Petty K.F., Noeimi H.Sanwal K.The freeway service patrol evaluation project: Database Support Programs, and Accessibility[J].Transportation Research Part C: Emerging Technologies,1996,4(2):71-85.
[44]肖汉光,蔡从中.特征向量的归一化比较性研究.计算机工程与应用[J],2009,45(022)：117-119.
[45]Ajith Abraham,He Guo,Hongbo Liu. Swarm intelligence:foundations, perspectives and applications [J]. Swarm Intelligence Systems,Studies in Computational Intelligence,2006.
[46]李佳.基于粒子群优化支持向量机的异常入侵检测研究[D].中南林业科技大学,2009.
[47]KENNEDY J,EBERHART R C. Particle Swarm Optimization[A]. Proceedings of IEEE International Conference on Neural Networks,1995,1942-1948.
[48]唐贤伦.混沌粒子群优化算法理论及应用研究[D].重庆大学博士论文,2007.
[49]王胥鹏,胡劲松.一种改进的微粒群算法[J].计算机应用研究.2009.10(10)3642-3648.
[50]Shi Y, Eberhart R. Empirical study of particle swarm optimization [C]. Proceeding of the 1999 Congress on Evolutionary Computation,1999:1945-1950.
[51]Trelea I. The particle swarm optimization algorithm:Convergence analysis and parameter selection [J]. Information Processing Letters,2003,85(6):317-325.
[52]贾永斌.基于SVM的交通流短时预测方法研究[D].西南交通大学硕士论 文,2012.
[53]魏秀业,潘宏侠.粒子群优化及智能故障诊断[M].国防工业大学,2010.
[54]Huiyuan Fan. A modification to particle swarm optimization algorithm. Engineering Computations [J].2002,19(8):970-989.
[55]黄席樾.现代智能算法理论及应用[M].北京：科学出版社,2009.
[56]MATLAB中文论坛Matlab神经网络30个案例分析[Z].北京航空航大学出版社,2010.
[57]纪震,廖惠连,吴青华.粒子群算法及应用[M].北京：科学出版社,2009.
[58]王兴玲,李占斌.基于网格搜索的支持向量机核函数参数的确定[J].中国海洋大学学报,2005,9.5(35)：859-862.
[59]罗启彬.基于混沌理论的遥测信息安全技术研究[D].中国工程物理研究院博士论文,2008.
[60]李冬萍.基于混沌粒子群优化的SVM分类器研究[J].计算机仿真,27(4)：185-191.
[61]徐文星.混沌粒子群优化算法及应用研究[D].北京化工大学博士论文,2012.
[2]张凯.高速公路交通事件检测与仿真研究[D].长安大学硕士论文,2009.
[3]陈斌.高速公路交通意外事件管理关键技术[M].西南交通大学出版社,2007.
[4]蔡志理.高速公路交通事件检测及交通疏导技术研究[D].吉林大学博士论文,2007。
[5]龚炯.基于SVM的高速公路交通事件检测研究[D].西南交通大学硕士论文,2010.
[6]陈祥梅.GA-SVM在高速公路交通事件检测中的应用研究.西南交通大学硕士论文,2009.
[7]Persaud B.N, Hall F.L.Catastrophe theory and patterns in 30 second freeway traffic data implications for incident detection. Transportation research. Part A: general,1989,23(2):103-113.
[8]Martin P.T., Perrin H.J., Hansen B.G. Incident detection algorithm evaluation[J]. University of Utah, March,2001:1-29.
[9]张敬磊,王晓原.交通事件检测算法研究进展[J].武汉理工大学学报：交通科学与工程版.2005,29(002)：215-218.
[10]Levin M., Krause G.M. Incident detection:A Bayesian approach^J].1978.
[11]Ahmed M, Cook A. Analysis of Freeway Traffic Time Series Data by Using Box-Jenkins Techniques." [J].Transportation Research Board,1979,722:1-9.
[12]Chew R.L.Ritchie5.Automatic Detection of Lane-blocking Freeway Incidents Using Artificial Neural Networks[J].Transportation Research Part C,1995,3(6): 371-388.
[13]Hsiao C.H. Lin C. Application of Fuzzy Logical and Neural Networks to Automatically Detect Freeway Traffic Incident[J]. Journal of Transportation Engineering,1993,125(5).
[14]Chang E. C, Wang S. H. Improved Freeway Incident Detection using Fuzzy Set Theory. Transportation Research Board.1994,21 (7):75-82
[15]Chew Stephen G Ritchie S. Neural Network Models for Automated Detection of Non-Recurring Congestion [J].1991:10-16.
[16]Li Y., McDonald M. Motorway incident detection using Probe vehicles[J].Transport,2005,158(1):11-15.
[17]Yuan F, Cheu R.L. Incident detection using support vector machines [J].Transporta tion Research Part C,2003,11(3-4):309-328.
[18]李文江,荆便顺,杨光王.基于小波分析的事件检测算法[J].西安公路交通大学学报.1997,17(2B)：134-138
[19]姜紫峰,刘小坤.基于神经网络的交通事件检测算法[J].西安公路交通大学学报.2001
[20]姜桂艳,温慧敏,杨兆升.高速公路交通事件自动检测系统与算法设计[J].交通运输Ⅰ程学报,2001,1(001)：77-81.
[21]温娟,贺国光.基于粗集理论预处理数据的神经网络交通事件自动检测算法《交通运输系统工程与信息》,2004,4：54-59.
[22]姚智胜,邵春福.基于V—支持向量分类机的交通事件检测方法研究[J].ITS通讯,2005,7(004)：38-41.
[23]梁新荣,刘智勇.基于支持向量机的高速公路事件检测[J].计算机工程与应用,2006(16)：213-218.
[24]蔡志理,姜桂艳.多SVM分类器融合技术在高速公路AID中的应用[C].第27届中国控制会议,中国昆明,2008(EL ISTP收录)
[25]朱红斌.基于Bagging算法和遗传神经网络的交通事件检测[J].计算机应用与软件.2010,1(27)：234-236
[26]周林英.基于支持向量机的高速公路事件检测算法[D].长安大学硕士论文,2009.
[27]姜卉.基于FA-SVM的高速公路交通事件自动检测方法研究.吉林大学硕士论文,2011.
[28]刘伟铭.高速公路系统控制方法[M].人民交通出版社,1998.
[29]覃频频.基于信息融合的高速公路事件检测建模与仿真[D].西南交通大学博士论文,2007.
[30]朱焱.基于AdaBoost-LMBP的高速公路交通事件检测算法研究[D].西南交大硕士论文,2012.
[31]张敬磊,王晓原.交通事件检测算法研究进展[J].武汉理工大学学报：交通科学与工程版,2005,29(002)：215-218.
[32]温慧敏,杨兆升.交通事件检测技术的进展研究[J].交通运输系统工程与信息.2005,5(001)：25-28
[33]Chen S, Wang W, Van Zuylen H. Construct support vector machine ensemble to detect traffic incident. Expert Systems with Applications,2009
[34]Vapnik V. N.,张学工.统计学习理论的本质[J].The nature of Statistical learning theory. NY:Springer Verlag,1995.
[35]Vapnik VN. The nature of statistical learning theory[M]. Springer Verlag,2000.
[36]边肇棋.模式识别(第二版)[M].清华大学出版社,2000.
[37]白鹏,张喜斌,张斌.支持向量机理论及工程应用实例[M].西安电子科技大学出版社,2007.
[38]]Boser B.E., Guyon I.M.,Vapnik V.N. A training algorithm for Optimal margin Classifiers,1992[C].ACM New York, NY, USA.
[39]祁亨年.支持向量机及其应用研究综述[J].计算机工程,2004,30(010)：6-9.
[40]王琪.基于神经网络和支持向量机的高速公路交通事件检测[D].西南交通大学硕士论文,2006.
[41]Hsu C.W.,Chang C.C., Lin C.J. A practical guide to support vector classification[Z]. Citeseer,2003.
[42]Karl Petty, The Analysis Software for the FSP Project [OL], http://ipa. eecs. berkeley. edu/-pettyk/FSP/.2005-01-01
[43]Petty K.F., Noeimi H.Sanwal K.The freeway service patrol evaluation project: Database Support Programs, and Accessibility[J].Transportation Research Part C: Emerging Technologies,1996,4(2):71-85.
[44]肖汉光,蔡从中.特征向量的归一化比较性研究.计算机工程与应用[J],2009,45(022)：117-119.
[45]Ajith Abraham,He Guo,Hongbo Liu. Swarm intelligence:foundations, perspectives and applications [J]. Swarm Intelligence Systems,Studies in Computational Intelligence,2006.
[46]李佳.基于粒子群优化支持向量机的异常入侵检测研究[D].中南林业科技大学,2009.
[47]KENNEDY J,EBERHART R C. Particle Swarm Optimization[A]. Proceedings of IEEE International Conference on Neural Networks,1995,1942-1948.
[48]唐贤伦.混沌粒子群优化算法理论及应用研究[D].重庆大学博士论文,2007.
[49]王胥鹏,胡劲松.一种改进的微粒群算法[J].计算机应用研究.2009.10(10)3642-3648.
[50]Shi Y, Eberhart R. Empirical study of particle swarm optimization [C]. Proceeding of the 1999 Congress on Evolutionary Computation,1999:1945-1950.
[51]Trelea I. The particle swarm optimization algorithm:Convergence analysis and parameter selection [J]. Information Processing Letters,2003,85(6):317-325.
[52]贾永斌.基于SVM的交通流短时预测方法研究[D].西南交通大学硕士论 文,2012.
[53]魏秀业,潘宏侠.粒子群优化及智能故障诊断[M].国防工业大学,2010.
[54]Huiyuan Fan. A modification to particle swarm optimization algorithm. Engineering Computations [J].2002,19(8):970-989.
[55]黄席樾.现代智能算法理论及应用[M].北京：科学出版社,2009.
[56]MATLAB中文论坛Matlab神经网络30个案例分析[Z].北京航空航大学出版社,2010.
[57]纪震,廖惠连,吴青华.粒子群算法及应用[M].北京：科学出版社,2009.
[58]王兴玲,李占斌.基于网格搜索的支持向量机核函数参数的确定[J].中国海洋大学学报,2005,9.5(35)：859-862.
[59]罗启彬.基于混沌理论的遥测信息安全技术研究[D].中国工程物理研究院博士论文,2008.
[60]李冬萍.基于混沌粒子群优化的SVM分类器研究[J].计算机仿真,27(4)：185-191.
[61]徐文星.混沌粒子群优化算法及应用研究[D].北京化工大学博士论文,2012.
© 2004-2018 中国地质图书馆版权所有 京ICP备05064691号 京公网安备11010802017129号 地址：北京市海淀区学院路29号 邮编：100083 电话：办公室：(+86 10)66554848；文献借阅、咨询服务、科技查新：66554700 |