基于浮动车的城市道路行程时间采集与预测方法研究
|
摘要
利用浮动车进行城市道路行程时间数据的采集和预测,能够在数据源上保证数据的精度和可靠性,提高数据的质量,另外,一方面,可以为用户出行提供交通诱导服务,使用户躲开交通拥挤和事故发生路段,以便合理地掌握出行时间,提高出行效率;另一方面,为交通管理部门提供数据支持,预测交通拥挤的持续时间和交通事故的事发路段及持续时间,进行交通异常状态的判别和交通事件的自动检测,缓解交通压力,提高行车速度。
论文首先确定了影响行程时间精度、可靠性、稳定性和质量的影响因素,如GPS数据的采样周期和传输周期、行程时间的分析周期、浮动车样本量和路段长度的划分,并用实测数据进行分析验证。
论文还对GPS数据进行了预处理,对错误数据进行修正、丢失数据进行补充,以期提高数据的精度。GPS数据和GIS地图进行匹配时,由于GPS误差和GIS地图也存在有误差,本文就提高地图匹配精度的方法进行了研究。
最后,运用ARMA模型、指数平滑方法、RBF神经网络和数据融合技术进行路段行程时间预测,路段平均行程速度为输入变量,输出特定路段的行程时间,并用实测数据进行分析验证。结果表明,RBF神经网络方法的预测精度要优于ARMA模型和指数平滑方法,而数据融合技术的预测精度更优于RBF神经网络,预测效果较为理想。
Lately, for industrialization quickening and the number of cars increasing rapidly, and road mileage and the vehicle’s rate is in a lack of balance, a series of transportation problems appear. Presently, ITS is a best way to solve transportation problems in cities, and it can solve transportation congestion、reform traveling safety、enhance vehicles’speed. It also can make people、cars and road harmonious. Travel time data’s collection and estimation, is an important part of ITS, can reflect transportation state’movement more truly. One way, it can provide transportation guidance service for users, make them save traveling time; On another way, provide data for transportation management department, they can identify traffic abnormal state and traffic incident automatically, to reduce transportation accidents and enhance transportation industry’s efficiency.
GPS floating cars are used to collect urban road datum, during datum’s transmit, for there is a series of reasons in existence, misdata and missing data would appear. So, to improve data’s quality and precision, guarantee travel time datum’s stability and reliability, the paper did data processing of GPS datum. The paper adopts floating cars based on GPS to collect travel time, send place, velocity and time to TIC by GPRS, match with GIS, to receive travel time and section average velocity, then use multi-kind model and arithmetic, to estimate travel time in the future. Finally, by GPRS, feed information back to users, provide selective drive route to them.
Travel time collection and estimation for urban roadway based on GPS Floating Cars can gain dynamic traffic information, provide traffic service information at present for users, and query travel manners, travel time, travel route, so they can arrive destination efficiently. It can also provide reliable datum for traffic management department, so they can differentiate traffic abnormal state and detect traffic incident, to save travel time, alleviate traffic congestion, reduce traffic accidents, and heighten traffic velocities.
GPS Floating Cars early is used in transportation outside, but the study inside drops behind. Urban highway travel time estimation also lags behind outside, whose studies focuses on freeway and arterial roadways, many scholars inside also process the travel time study. The paper discusses study status quo of Travel time collection and estimation for urban roadway based on GPS Floating Cars, review and judge them.
For all kinds of affective factors exist, the datum’s quality, precision and reliability for GPS floating cars can’t guarantee, so to improve precision and stability of the datum, the paper analyses the affective factors, such as GPS and GIS’s precision, GPS signal losing, GPS sampling interval, transmitting interval, analyzing interval, section length, the number of floating cars and dependability lack of sample, confirm the studying work of the paper.
Pretty sampling interval can enhance datum’s precision and reliability, and can provide datum for traffic management department. So the paper divides GPS sampling interval into sampling cycle、transmitting cycle and analysis cycle, and analyze via datum survey practically, at last make the best GPS’s sampling interval certain.
GPS datum collected must combine with link, travel time would be confirmed. Link partition directly influence travel time datum’s stability and reliability. So to improve travel time datum’s quality, the paper studies link partition, and analyze via datum survey practically.
To obtain road network’s true state, there must be enough floating cans to satisfy basic requirement of transportation information collection. But, the amount of floating cars achieves to a certainty, we increase the number of floating cars continually, it can’t improve datum’s precision evidently, and travel time won’t have stability. So the paper is to find the best floating car’s number, and analyze via datum survey practically.
As a result of SA policy, and there is error in GPS satellite、transmit pathway and user equipment, and all make GPS datum and GIS map matching’s precision fall. So to improve vehicle’s orientation precision、travel time’s stability and reliability, the paper studies how to improve map matching precision.
Based on GPS datum collection, only datum collected is of high quality, it can go to travel time estimation. The paper analyzes travel time’s predictability so as to improve datum’s precision and reduce economic cost.
For ARMA model, Exponential Smoothing Technique, RBF Neural Network and MMFA model have commonness, in the same way, they can be used to estimate travel time estimation of highways. The paper sets up travel time estimation model, and analyses steps of them. The paper adopts SPSS software, matlab network toolbox based on history datum to estimate travel time.
Based on precision, reliability and stability of datum collected, the paper adopts ARMA model, Exponential Smoothing Technique, RBF Neural Network and MMFA to estimate travel time, link average travel speed is input variable, output certain link’s travel time. Judging estimation results via estimation guideline, the conclusion is that MMFA’s estimation precision precedes to other models.
论文首先确定了影响行程时间精度、可靠性、稳定性和质量的影响因素,如GPS数据的采样周期和传输周期、行程时间的分析周期、浮动车样本量和路段长度的划分,并用实测数据进行分析验证。
论文还对GPS数据进行了预处理,对错误数据进行修正、丢失数据进行补充,以期提高数据的精度。GPS数据和GIS地图进行匹配时,由于GPS误差和GIS地图也存在有误差,本文就提高地图匹配精度的方法进行了研究。
最后,运用ARMA模型、指数平滑方法、RBF神经网络和数据融合技术进行路段行程时间预测,路段平均行程速度为输入变量,输出特定路段的行程时间,并用实测数据进行分析验证。结果表明,RBF神经网络方法的预测精度要优于ARMA模型和指数平滑方法,而数据融合技术的预测精度更优于RBF神经网络,预测效果较为理想。
Lately, for industrialization quickening and the number of cars increasing rapidly, and road mileage and the vehicle’s rate is in a lack of balance, a series of transportation problems appear. Presently, ITS is a best way to solve transportation problems in cities, and it can solve transportation congestion、reform traveling safety、enhance vehicles’speed. It also can make people、cars and road harmonious. Travel time data’s collection and estimation, is an important part of ITS, can reflect transportation state’movement more truly. One way, it can provide transportation guidance service for users, make them save traveling time; On another way, provide data for transportation management department, they can identify traffic abnormal state and traffic incident automatically, to reduce transportation accidents and enhance transportation industry’s efficiency.
GPS floating cars are used to collect urban road datum, during datum’s transmit, for there is a series of reasons in existence, misdata and missing data would appear. So, to improve data’s quality and precision, guarantee travel time datum’s stability and reliability, the paper did data processing of GPS datum. The paper adopts floating cars based on GPS to collect travel time, send place, velocity and time to TIC by GPRS, match with GIS, to receive travel time and section average velocity, then use multi-kind model and arithmetic, to estimate travel time in the future. Finally, by GPRS, feed information back to users, provide selective drive route to them.
Travel time collection and estimation for urban roadway based on GPS Floating Cars can gain dynamic traffic information, provide traffic service information at present for users, and query travel manners, travel time, travel route, so they can arrive destination efficiently. It can also provide reliable datum for traffic management department, so they can differentiate traffic abnormal state and detect traffic incident, to save travel time, alleviate traffic congestion, reduce traffic accidents, and heighten traffic velocities.
GPS Floating Cars early is used in transportation outside, but the study inside drops behind. Urban highway travel time estimation also lags behind outside, whose studies focuses on freeway and arterial roadways, many scholars inside also process the travel time study. The paper discusses study status quo of Travel time collection and estimation for urban roadway based on GPS Floating Cars, review and judge them.
For all kinds of affective factors exist, the datum’s quality, precision and reliability for GPS floating cars can’t guarantee, so to improve precision and stability of the datum, the paper analyses the affective factors, such as GPS and GIS’s precision, GPS signal losing, GPS sampling interval, transmitting interval, analyzing interval, section length, the number of floating cars and dependability lack of sample, confirm the studying work of the paper.
Pretty sampling interval can enhance datum’s precision and reliability, and can provide datum for traffic management department. So the paper divides GPS sampling interval into sampling cycle、transmitting cycle and analysis cycle, and analyze via datum survey practically, at last make the best GPS’s sampling interval certain.
GPS datum collected must combine with link, travel time would be confirmed. Link partition directly influence travel time datum’s stability and reliability. So to improve travel time datum’s quality, the paper studies link partition, and analyze via datum survey practically.
To obtain road network’s true state, there must be enough floating cans to satisfy basic requirement of transportation information collection. But, the amount of floating cars achieves to a certainty, we increase the number of floating cars continually, it can’t improve datum’s precision evidently, and travel time won’t have stability. So the paper is to find the best floating car’s number, and analyze via datum survey practically.
As a result of SA policy, and there is error in GPS satellite、transmit pathway and user equipment, and all make GPS datum and GIS map matching’s precision fall. So to improve vehicle’s orientation precision、travel time’s stability and reliability, the paper studies how to improve map matching precision.
Based on GPS datum collection, only datum collected is of high quality, it can go to travel time estimation. The paper analyzes travel time’s predictability so as to improve datum’s precision and reduce economic cost.
For ARMA model, Exponential Smoothing Technique, RBF Neural Network and MMFA model have commonness, in the same way, they can be used to estimate travel time estimation of highways. The paper sets up travel time estimation model, and analyses steps of them. The paper adopts SPSS software, matlab network toolbox based on history datum to estimate travel time.
Based on precision, reliability and stability of datum collected, the paper adopts ARMA model, Exponential Smoothing Technique, RBF Neural Network and MMFA to estimate travel time, link average travel speed is input variable, output certain link’s travel time. Judging estimation results via estimation guideline, the conclusion is that MMFA’s estimation precision precedes to other models.
引文
[1] D.J. Dailey, The use of transit vehicles as speed probes for traffic management and traveler information in addition to performance monitoring, TransNow, Univ. Washington, 2001
[2] D.J. Dailey. The use of transit vehicles as speed probes for traffic management and traveler information in addition to performance monitoring[R]. TransNow, Univ. Washington, 2001
[3] Robert L.Bertini,Sutti Tantiyanugulchai. Transit Buses as Traffic Probes: Empirical Evaluation Using Geo-Location Data[C]. Transportation Research Board 83rd Annual Meeting. USA, 2004
[4] 姜桂艳. 道路交通状态判别技术与应用[M]. 北京:人民交通出版社,2004
[5] 姜桂艳, 蔡志理, 冮龙晖, 唐金芝. 基于3GS技术的道路交通智能监控系统研究[C]. 第二届中日亚洲城市发展与智能运输系统研讨会. 中国,2006
[6] 董敬欣, 吴建平. 使用浮动车检测 OD 矩阵的算法及可靠性分析[J].北京交通大学学报(自然科学版). 2005 (3):73-76
[7] 盛志杰,吴卉,喻泉,刘允才.采用 GPS 探测车的城市交通流分析[J]. 微型电脑应用, 2006(1) :3-6,53
[8] 孙晓峰, 吴建平.基于浮动车数据采集技术的城市交通网络功能评价方法研究[J]. 现代交通技术, 2006(6):55-58
[9] 储浩,吴志周,杨晓光.基于城市公交探测车数据的分析和处理方法[J]. 系统工程,2006(8):113-118
[10] 张存保,杨晓光,严新平.基于浮动车的交通信息采集系统研究,交通与计算机.2006 Vol.24 No.5: 31-34
[11] 涂智、李昊、姚琛、袁理,基于最小浮动车样本数量的道路覆盖率与交通信息更新周期研究 ,中国铁道科学, 2006 Vol.27 No.5 :127-131
[12] 姚琛.基于路段覆盖率的浮动车样本数量研究.山东理工大学学报(自然科学版),2006 Vol.20 No.3:96-98
[13] 张晓春, 吕北岳等. 基于车载GPS技术的交通浮动车检测系统设计研究[J].国际智能交通,2005(02)
[15] 杨兆升.关于智能运输系统的关键理论-综合路段行程时间预测的研究[J].交通运输工程学报,2001,3.
[16] 翁剑成,荣健,任福田.基于浮动车采集技术的动态车载导航系统体系构架研究[A],第一届中国智能交通年会论文集,同济大学出版社.
[17] 杨兆升.城市交通流诱导系统理论与模型[M],人民交通出版社 2000,14~29.
[18] 孙建平,温慧敏,郭继孚,浮动车交通信息采集系统建设框架研究[A],第一届中国智能交通年会论文集,同济大学出版社.
[19] 王力,王川久,张海,范跃祖.基于浮动车的城市动态交通信息采集处理方法研究[A],第一届中国智能交通年会论文集,同济大学出版社.
[20] 杨兆升,朱中.基于 BP 神经网络的路径行程时间实时预测模型[J],系统工程理论与实践,1999.8.
[21] 杭明升,杨晓光,彭国雄.基于卡尔曼滤波的高速道路行程时间动态预测[J],同济大学学报,2002,9.
[22] 杨兆升,孙喜梅.实时动态路段行程时间预测的一种实用方法[J],公路交通科技,2001.
[23] Karl F.Petty, Peter Bickel, Jiming Jiang,etc. Accurate estimation of travel times from single-loop detectors.[EB/OL].http://www.elsevier.com
[24] 孙晓峰,吴建平.基于浮动车数据采集技术的城市交通网络功能评价方法研究[J],现代交通技术.2005,6.
[25] Vaneet Sethi,Nikhil Bhandari,et al. Arterial incident detection using fixed detector and probe vehicle data,Transpn Research,3,1559.
[26] 杨兆升,保丽霞,朱国华.基于Fuzzy回归的快速路行程时间预测模型研究[J],公路交通科技,2004,3.
[27] Edward Chung, Majid Sarvi,Yasunori Murakami, etc al.Cleansing of probe car data to determine Trip OD, http://www.elsevier.com
[28] Cesar A.Quiroga, Darcy Bullock. Travel time studies with global positioning and geographic information systems: an integrated methodology[J]. Transportation Research Part C. 1998.
[29] 徐春荣,欧阳为民,勾海波,吴师鹏.智能交通数据分析系统的设计与实现[J].计算机工程与应用.2005,28
[30] 季常煦,杨楠,高歌.面向ATMS共用信息平台的数据预处理技术的研究[J].交通运输系统工程与信息.2005,6
[31] 苏洁,周东方,岳春生.GPS 车辆导航中的实时地图匹配算法[J].测绘学报.2001,8
[32] 钟海丽,童瑞华,李军,陈宏盛.GPS 定位与地图匹配方法研究[J].小型微型计算机系统.2003,1
[33] 张巨,陈荦,景宁,刘雨,陈宏盛.GPS/GIS 车辆路径定位方法研究[J].遥测遥控.1999,9
[34] 吴卉,盛志杰,喻泉,刘允才.GIS/GPS 城市交通流监测系统中的地图匹配算法[J].计算机工程.2006,4
[35] 刘基余,李征航,王跃虎,桑吉章.全球定位系统原理及其应用[M].中国地图出版社,1999.pp126~136
[36] 张晓东.动态交通流信息采集系统若干问题研究[D].吉林大学,2004.
[37] 詹舒波,张其善.GPS/电子地图的坐标转换算法和实现[J].北京航空航天大学学报.1996,10
[38] 曾波,江资斌.GPS 车载导航系统的地图匹配算法[J].测绘工程.2004,9
[39] 宋凝芳.差分 GPS 技术在车辆监控系统中的应用[J].中国惯性技术学报.1998,6
[40] 杨兆升,初连禹.差分 GPS 在城市交通流诱导中的作用[J].吉林工业大学学报.1998,1
[41] 路锋,崔伟宏.车辆导航与监控中 GPS/GIS 实时定位配准误差分析[J].遥感学报.1999,11
[42] 杨先平.城市道路行程时间预测方法研究[D].吉林大学,2005.
[43] 陈玉祥,张汉亚.预测计算方法[M].机械工业出版社,1985,8.pp:31-48
[44] 江之源.经济预测方法与模型[M].西南财经大学出版社,成都.1999,9.
[45] 白竹.城市主干路交通异常状态自动判别算法研究[D].吉林大学,2006.
[46] 杭明升,杨晓光,彭国雄.基于卡尔曼滤波的高速道路行程时间动态预测[J].同济学报.2002,9
[47] 龚剑,朱亮.MATLAB 入门与提高[M].清华大学出版社,北京.2000.
[48] 郭景峰,侯爽,王金慧.一种动态路段行程时间的预测模型[J].计算机工程与科学.2005,27
[49] 杨晓光,蔡润林,庄斌.基于车牌自动识别系统的城市道路行程时间预测算法[J].交通与计算机.2005,3
[50] Mei chen, StevenI.Chien. Dynamic Freeway Travel Time Prediction Using Probe Vehicle Data: Link-based vs. path-based. http://www.dtic.mil/dticasd/sbir/sbir043/n231b.pdf
[51] 孙棣华.基于径向基函数(RBF)神经网络的路段行程时间预测研究[D].重庆大学.2004
[52] 张赫,杨兆升,李贻武.无检测器交叉口交通流量预测方法综合研究[J].公路交通科技,2002.
[53] 苏洁,周东方,岳春生.GPS 车辆导航中的实时地图匹配算法[J].测绘学报.2001,8
[54] 余艳春,邵春福,郭钰愫,熊志华.基于实时数据的路网行程时间可靠度模型研究[J].现代交通技术.2006,2
[55] 刘英伟.交通网络行程时间可靠度的初步研究[D].同济大学.2001.
[56] 熊志华,姚智胜,邵春福.基于路段相关的行程时间路网可靠性[J].中国安全科学学报.2004,10
[57] 刘海旭,卜雷,蒲云.随机路网的行程时间可靠性[J].土木工程学报.2004,8
[2] D.J. Dailey. The use of transit vehicles as speed probes for traffic management and traveler information in addition to performance monitoring[R]. TransNow, Univ. Washington, 2001
[3] Robert L.Bertini,Sutti Tantiyanugulchai. Transit Buses as Traffic Probes: Empirical Evaluation Using Geo-Location Data[C]. Transportation Research Board 83rd Annual Meeting. USA, 2004
[4] 姜桂艳. 道路交通状态判别技术与应用[M]. 北京:人民交通出版社,2004
[5] 姜桂艳, 蔡志理, 冮龙晖, 唐金芝. 基于3GS技术的道路交通智能监控系统研究[C]. 第二届中日亚洲城市发展与智能运输系统研讨会. 中国,2006
[6] 董敬欣, 吴建平. 使用浮动车检测 OD 矩阵的算法及可靠性分析[J].北京交通大学学报(自然科学版). 2005 (3):73-76
[7] 盛志杰,吴卉,喻泉,刘允才.采用 GPS 探测车的城市交通流分析[J]. 微型电脑应用, 2006(1) :3-6,53
[8] 孙晓峰, 吴建平.基于浮动车数据采集技术的城市交通网络功能评价方法研究[J]. 现代交通技术, 2006(6):55-58
[9] 储浩,吴志周,杨晓光.基于城市公交探测车数据的分析和处理方法[J]. 系统工程,2006(8):113-118
[10] 张存保,杨晓光,严新平.基于浮动车的交通信息采集系统研究,交通与计算机.2006 Vol.24 No.5: 31-34
[11] 涂智、李昊、姚琛、袁理,基于最小浮动车样本数量的道路覆盖率与交通信息更新周期研究 ,中国铁道科学, 2006 Vol.27 No.5 :127-131
[12] 姚琛.基于路段覆盖率的浮动车样本数量研究.山东理工大学学报(自然科学版),2006 Vol.20 No.3:96-98
[13] 张晓春, 吕北岳等. 基于车载GPS技术的交通浮动车检测系统设计研究[J].国际智能交通,2005(02)
[15] 杨兆升.关于智能运输系统的关键理论-综合路段行程时间预测的研究[J].交通运输工程学报,2001,3.
[16] 翁剑成,荣健,任福田.基于浮动车采集技术的动态车载导航系统体系构架研究[A],第一届中国智能交通年会论文集,同济大学出版社.
[17] 杨兆升.城市交通流诱导系统理论与模型[M],人民交通出版社 2000,14~29.
[18] 孙建平,温慧敏,郭继孚,浮动车交通信息采集系统建设框架研究[A],第一届中国智能交通年会论文集,同济大学出版社.
[19] 王力,王川久,张海,范跃祖.基于浮动车的城市动态交通信息采集处理方法研究[A],第一届中国智能交通年会论文集,同济大学出版社.
[20] 杨兆升,朱中.基于 BP 神经网络的路径行程时间实时预测模型[J],系统工程理论与实践,1999.8.
[21] 杭明升,杨晓光,彭国雄.基于卡尔曼滤波的高速道路行程时间动态预测[J],同济大学学报,2002,9.
[22] 杨兆升,孙喜梅.实时动态路段行程时间预测的一种实用方法[J],公路交通科技,2001.
[23] Karl F.Petty, Peter Bickel, Jiming Jiang,etc. Accurate estimation of travel times from single-loop detectors.[EB/OL].http://www.elsevier.com
[24] 孙晓峰,吴建平.基于浮动车数据采集技术的城市交通网络功能评价方法研究[J],现代交通技术.2005,6.
[25] Vaneet Sethi,Nikhil Bhandari,et al. Arterial incident detection using fixed detector and probe vehicle data,Transpn Research,3,1559.
[26] 杨兆升,保丽霞,朱国华.基于Fuzzy回归的快速路行程时间预测模型研究[J],公路交通科技,2004,3.
[27] Edward Chung, Majid Sarvi,Yasunori Murakami, etc al.Cleansing of probe car data to determine Trip OD, http://www.elsevier.com
[28] Cesar A.Quiroga, Darcy Bullock. Travel time studies with global positioning and geographic information systems: an integrated methodology[J]. Transportation Research Part C. 1998.
[29] 徐春荣,欧阳为民,勾海波,吴师鹏.智能交通数据分析系统的设计与实现[J].计算机工程与应用.2005,28
[30] 季常煦,杨楠,高歌.面向ATMS共用信息平台的数据预处理技术的研究[J].交通运输系统工程与信息.2005,6
[31] 苏洁,周东方,岳春生.GPS 车辆导航中的实时地图匹配算法[J].测绘学报.2001,8
[32] 钟海丽,童瑞华,李军,陈宏盛.GPS 定位与地图匹配方法研究[J].小型微型计算机系统.2003,1
[33] 张巨,陈荦,景宁,刘雨,陈宏盛.GPS/GIS 车辆路径定位方法研究[J].遥测遥控.1999,9
[34] 吴卉,盛志杰,喻泉,刘允才.GIS/GPS 城市交通流监测系统中的地图匹配算法[J].计算机工程.2006,4
[35] 刘基余,李征航,王跃虎,桑吉章.全球定位系统原理及其应用[M].中国地图出版社,1999.pp126~136
[36] 张晓东.动态交通流信息采集系统若干问题研究[D].吉林大学,2004.
[37] 詹舒波,张其善.GPS/电子地图的坐标转换算法和实现[J].北京航空航天大学学报.1996,10
[38] 曾波,江资斌.GPS 车载导航系统的地图匹配算法[J].测绘工程.2004,9
[39] 宋凝芳.差分 GPS 技术在车辆监控系统中的应用[J].中国惯性技术学报.1998,6
[40] 杨兆升,初连禹.差分 GPS 在城市交通流诱导中的作用[J].吉林工业大学学报.1998,1
[41] 路锋,崔伟宏.车辆导航与监控中 GPS/GIS 实时定位配准误差分析[J].遥感学报.1999,11
[42] 杨先平.城市道路行程时间预测方法研究[D].吉林大学,2005.
[43] 陈玉祥,张汉亚.预测计算方法[M].机械工业出版社,1985,8.pp:31-48
[44] 江之源.经济预测方法与模型[M].西南财经大学出版社,成都.1999,9.
[45] 白竹.城市主干路交通异常状态自动判别算法研究[D].吉林大学,2006.
[46] 杭明升,杨晓光,彭国雄.基于卡尔曼滤波的高速道路行程时间动态预测[J].同济学报.2002,9
[47] 龚剑,朱亮.MATLAB 入门与提高[M].清华大学出版社,北京.2000.
[48] 郭景峰,侯爽,王金慧.一种动态路段行程时间的预测模型[J].计算机工程与科学.2005,27
[49] 杨晓光,蔡润林,庄斌.基于车牌自动识别系统的城市道路行程时间预测算法[J].交通与计算机.2005,3
[50] Mei chen, StevenI.Chien. Dynamic Freeway Travel Time Prediction Using Probe Vehicle Data: Link-based vs. path-based. http://www.dtic.mil/dticasd/sbir/sbir043/n231b.pdf
[51] 孙棣华.基于径向基函数(RBF)神经网络的路段行程时间预测研究[D].重庆大学.2004
[52] 张赫,杨兆升,李贻武.无检测器交叉口交通流量预测方法综合研究[J].公路交通科技,2002.
[53] 苏洁,周东方,岳春生.GPS 车辆导航中的实时地图匹配算法[J].测绘学报.2001,8
[54] 余艳春,邵春福,郭钰愫,熊志华.基于实时数据的路网行程时间可靠度模型研究[J].现代交通技术.2006,2
[55] 刘英伟.交通网络行程时间可靠度的初步研究[D].同济大学.2001.
[56] 熊志华,姚智胜,邵春福.基于路段相关的行程时间路网可靠性[J].中国安全科学学报.2004,10
[57] 刘海旭,卜雷,蒲云.随机路网的行程时间可靠性[J].土木工程学报.2004,8