CN102800197A - Preprocessing method of road section dynamic traffic stream essential data of urban road - Google Patents

Preprocessing method of road section dynamic traffic stream essential data of urban road Download PDF

Info

Publication number
CN102800197A
CN102800197A CN2012100454645A CN201210045464A CN102800197A CN 102800197 A CN102800197 A CN 102800197A CN 2012100454645 A CN2012100454645 A CN 2012100454645A CN 201210045464 A CN201210045464 A CN 201210045464A CN 102800197 A CN102800197 A CN 102800197A
Authority
CN
China
Prior art keywords
data
traffic
step
represents
lane
Prior art date
Application number
CN2012100454645A
Other languages
Chinese (zh)
Other versions
CN102800197B (en
Inventor
夏井新
黄卫
陆振波
张韦华
安成川
聂庆慧
Original Assignee
东南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 东南大学 filed Critical 东南大学
Priority to CN201210045464.5A priority Critical patent/CN102800197B/en
Publication of CN102800197A publication Critical patent/CN102800197A/en
Application granted granted Critical
Publication of CN102800197B publication Critical patent/CN102800197B/en

Links

Abstract

The invention discloses a preprocessing method of road section dynamic traffic stream essential data of an urban road. The method comprises the following steps of: acquiring the traffic stream essential data dynamically for time warping; detecting the effectiveness of divided lane traffic stream essential data; collecting the time of divided lane traffic stream essential data; collecting the fracture surface space at individual vehicle driving direction of the divided lane traffic stream essential data; and estimating the dynamic traffic stream loss data at individual vehicle driving direction. According to the invention, based on the characteristics of collected data of a fixed-type vehicle detector, the essential idea of the traffic stream theory is combined by adopting a threshold method, a set of data effectiveness detecting method is provided; aiming at different data loss patterns, history data is utilized reasonably, and the continuity and integrity of data are ensured furthest; the algorithm related to data preprocessing considers the timeliness and accuracy requirements at the same time, and the data analysis processing capability is high; and the preprocessing method of road section dynamic traffic stream essential data of the urban road plays positive significance to urban intelligent transportation system construction, road traffic informationization level improvement, and road operation management level improvement.

Description

一种城市道路路段动态交通流基础数据的预处理方法 Pretreatment method for dynamic traffic flow in urban road sections basic data

技术领域 FIELD

[0001] 本发明涉及智能交通应用领域,特别是一种城市道路路段动态交通流基础数据的预处理方法。 [0001] The present invention relates to the field of intelligent transportation applications, especially in urban road sections pretreatment method of dynamic traffic data base.

背景技术 Background technique

[0002] 基于固定式车辆检测器采集的动态交通流基础数据是重要的智能运输系统(ITS)数据源。 [0002] It is important Intelligent Transport System (ITS) data source based on dynamic traffic data of a fixed basic vehicle detector acquisition. 由于固定式车辆检测器受自身工作状态、网络传输、道路交通状况及周围环境等不确定性因素影响,采集的数据存在错误、丢失、时间点漂移、噪声过大等问题。 Since the stationary vehicle detector by its own working conditions, uncertainties affecting network traffic, road conditions and the surrounding environment, there is an error in data acquisition, loss, time drift, excessive noise and other issues. 如果对原始数据不加以预处理直接应用,则会影响上层智能运输应用系统对城市道路交通状况及性能评价指标估计与预测的准确性和可靠性,造成系统运行人为干预度大、应用系统可持续性不强、应用范围受限等问题。 If not pre-applied directly to the raw data, it will affect the accuracy and reliability of the upper intelligent transport system applications for urban road traffic conditions and performance evaluation estimates and projections, and cause the system to run a large degree of human intervention, sustainable applications is not strong, the limited scope of application and other issues. 可见数据预处理的主要目的是控制道路采集交通流数据的质量, 降低问题数据对整体数据精确度的影响,保证道路交通流数据的精确处理和安全应用,也为改善城市道路动态交通流数据采集系统提供反馈信息和理论依据。 The main purpose of the visible data preprocessing is to control the quality of road traffic flow data collection, reducing the impact on the overall accuracy of the data problem data to ensure accurate handling and security applications traffic flow data, but also to improve the urban road dynamic traffic data collection The system provides feedback and theoretical basis.

[0003] 动态交通流基础数据的预处理最关键的两方面内容为数据有效性检验和缺失数据估计。 [0003] The two most critical aspects of pretreatment dynamic traffic flow data base for data validation and missing data estimation. 数据有效性检验方法经历了从简单的阈值检验法到阈值检验法结合交通流理论的综合检验法的发展过程,典型的算法有华盛顿算法(Washington Algorithm)、日常统计算法(Daily Statistics Algorithm)等;关于缺失数据的估计,早期的ITS系统对缺失数据并没有进行估算处理,而是采用忽略的方法,直接去除某些可疑或错误数据。 Data validation methods through the development process of a comprehensive test method combined with traffic flow theory from a simple threshold test method to the threshold test, typical algorithms Washington algorithm (Washington Algorithm), daily statistical algorithm (Daily Statistics Algorithm) and so on; About estimates of missing data, early ITS system for missing data did not estimate process, instead of using the method ignores the direct removal of some suspicious or wrong data. 显然忽略缺失数据不仅降低了交通流数据的样本量,也很有可能损失部分有价值的信息。 Apparently ignored missing data not only reduces the amount of sample traffic flow data are also likely to lose some valuable information. 基于此,国内外研究和工程人员认识到对缺失数据进行估计的理论与现实意义,并对缺失估计算法进行了大量的深入研究。 Based on this, research and engineering personnel at home and abroad recognize the theoretical and practical significance estimate of the missing data, and the estimation algorithm missing a lot of in-depth research. 目前交通流缺失数据估计方法大致可以分为基于经验的估计方法、基于统计的估计方法、人工智能估计方法以及组合式估计方法四类。 Currently missing data traffic flow estimation method can be divided into empirical estimation method based on statistical estimation method based on artificial intelligence and a combined estimation method estimation method four categories. 其中基于经验的估计方法是交通专家运用自己掌握的交通知识和经验总结出来的方法,如历史平均法,基于车道分布的交通缺失数据估计方法等;基于统计的估计方法主要包括时间序列分析法、ΕΜ/DA模型、线性回归模型、卡尔曼滤波模型等;人工智能估计方法主要包括各类神经网络算法及遗传算法等;组合式估计方法是对多种方法的融合应用,如基于历史平均与时间序列分析发的组合算法、配对模型、遗传算法与神经网络或回归模型的组合算法等。 Which is estimated based on the experience of traffic experts use their grasp of transportation knowledge and experience summed up methods, such as the historical average, Transportation missing data estimation methods such as lane distribution based; estimation method based on statistics including time series analysis, ΕΜ / DA model, linear regression model, the Kalman filter models; AI estimation method includes various types of artificial neural network and genetic algorithms; combined estimation method applied is the integration of a variety of methods, such as time based on the historical average combining algorithm sequence analysis of hair, matching model, a combination of genetic algorithm and the neural network algorithm or a regression model.

[0004] 总结目前国内外在动态交通流基础数据预处理方法的研究和实践,还存在着如下几个问题: [0004] summarizes the current domestic and international research and practice in basic dynamic traffic flow data preprocessing methods, there are still several problems as follows:

[0005] I、现有数据预处理涉及的算法无法同时兼顾ITS系统应用的实时性和准确性要求,缺乏交通流理论的支撑,数据分析处理能力较弱、挖掘层次较低; [0005] I, the existing data preprocessing algorithm can not be both at the same time involved in real-time and accuracy requirements of the application of ITS systems, lack of support traffic flow theory, data analysis and processing capability is weak, lower mining levels;

[0006] 2、数据有效性检验未形成一套全面、综合的规则体系,检验规则本身缺乏一定的层次性; [0006] 2, data validation is not formed a comprehensive, integrated system of rules and inspection rules themselves a certain lack of hierarchy;

[0007] 3、缺失数据估计缺乏对实际数据缺失模式的识别和历史数据的利用,包含经验化的成分较多,而且采用单一算法的应用效果不理想。 [0007] 3, the missing data is estimated using the lack of actual data missing pattern recognition and historical data, it contains the experience of many ingredients, and the use of a single application of the algorithm is not ideal. 发明内容 SUMMARY

[0008] 发明目的 [0008] Object of the Invention

[0009] 本发明的目的在于,基于固定式车辆检测器采集数据特点,采用阈值法结合交通流理论的基本思想,形成一套全面、综合、具有层次性的数据有效性检验方法;针对不同的数据缺失模式,合理利用历史数据,采用相对最优的方法估计缺失数据,最大限度地保证数据的连续性和完整性;在此基础上,综合考虑数据预处理算法的实时性和准确性要求。 [0009] The object of the present invention is based on the characteristics of the data collection vehicle stationary detector, the basic idea of ​​using a threshold method combined traffic flow theory, the formation of a comprehensive, integrated, having a hierarchy of data validation methods; for different data deletion mode, the rational use of historical data, a method using a relatively optimal estimate the missing data, and ensure maximum continuity of data integrity; on this basis, considering the accuracy requirements of real-time data and preprocessing algorithm.

[0010] 技术方案 [0010] Technical Solution

[0011] 本发明的目的是通过如下步骤实现的: [0011] The object of the present invention is achieved by the following steps:

[0012] 一种城市道路路段动态交通流基础数据的预处理方法,包括如下步骤: [0012] A pretreatment method for dynamic traffic urban road sections basic data stream, comprising the steps of:

[0013] I)以一定时间间隔获取固定式车辆检测器采集的分车道交通流基础数据,规整数据的时间戳属性为最近相邻的整分钟时刻;· [0013] I) at predetermined time intervals, lane-based traffic data collection vehicle detector fixed, the time stamp attribute data structured nearest neighbor full minute time point; -

[0014] 2)分车道交通流基础数据有效性检验,依次对各车道的交通流基础数据进行有效性检验: [0014] 2) sub-lane traffic flow basic data validation, followed by basic traffic flow data for each lane validity test:

[0015] 2-1)数据非空检验,若数据为空,则标记为无效数据,转入步骤2-7); [0015] 2-1) a non-null test data, if the data is empty, the data marked invalid, go to step 2-7);

[0016] 2-2)日期时间戳有效性检验,若为日期时间戳错误数据,则标记为无效数据,转入步骤2-7); [0016] 2-2) validity check date stamp, the date stamp if the erroneous data, the data is marked as invalid, go to step 2-7);

[0017] 2-3)非重复数据检验,若为重复数据,则标记为无效数据,转入步骤2-7); [0017] 2-3) of unique data checking, if the repetitive data, the data is marked as invalid, go to step 2-7);

[0018] 2-4)车辆存在检验,若交通流基础数据属性变量值均为零,则标记为无效数据,转入步骤2-7); Testing the presence of [0018] 2-4) of the vehicle, if the traffic attribute data based variable values ​​are zero, the data marked invalid, go to step 2-7);

[0019] 2-5)单车道交通流基础数据阈值法检验; [0019] 2-5) traffic lane data based thresholding test;

[0020] 2-6)单车道交通流基础数据交通流理论法检验; [0020] 2-6) lane traffic flow theory underlying data is examined;

[0021] 2-7)若所有车道交通流基础数据均已完成有效性检验,则转入步骤3),否则进行下一车道交通流基础数据有效性检验; [0021] 2-7) If all lanes of traffic flow data base have been completed validation, then go to step 3), otherwise the validity test the next lane traffic flow data base;

[0022] 3)判断原始数据累计采集时间间隔是否等于汇集时间间隔,若满足,则将原始数据累计采集时间间隔归零,并转入步骤4); [0022] 3) Analyzing the original data acquisition time interval is equal to the total integration time interval, if satisfied, the original data acquisition time interval cumulative zero, and proceeds to step 4);

[0023] 4)分车道交通流基础数据时间汇集,即得到各车道汇集时间间隔的有效交通流基础数据; [0023] 4) basic traffic lane-time data collection, i.e., the effective data of each traffic lane based collection time interval;

[0024] 5)分车道交通流基础数据单个车辆行驶方向断面空间汇集,若空间汇集结果为有效数据,即得到单个车辆行驶断面汇集时间间隔的有效交通流基础数据,若空间汇集结果为无效数据,则将其视为缺失数据,转入步骤6); [0024] 5) lane-sectional spatial traffic flow based data from a single vehicle traveling direction pooled effective if the data space together result, i.e., an effective traffic flow based data from a single vehicle traveling section integration time interval, if the space collection result is invalid data , it is treated as missing data, the process proceeds to step 6);

[0025] 6)单个车辆行驶方向断面交通流缺失数据估计,即得到单个车辆行驶断面汇集时间间隔的估计交通流基础数据: [0025] 6) a single vehicle traveling direction estimation section missing data traffic flow, i.e., to obtain an estimated traffic flow data of a single base section together with the vehicle interval:

[0026] 6-1)判断连续缺失时间间隔,若时间间隔小于15分钟,则转入步骤6-2),否则转入步骤6-3); [0026] 6-1) is determined continuously deletion interval, if the time interval is less than 15 minutes, the process proceeds to step 6-2), otherwise it proceeds to step 6-3);

[0027] 6-2)采用随机漫步法估计缺失数据; [0027] 6-2) using a random walk method to estimate the missing data;

[0028] 6-3)若连续缺失时间间隔大于等于15分钟,小于等于30分钟,则转入步骤6_4),否则转入步骤6-5); [0028] 6-3) If time intervals greater than or equal consecutive deletions 15 minutes, 30 minutes or less, the process proceeds to step 6_4), otherwise proceeds to step 6-5);

[0029] 6-4)采用历史平均结合实时调整法估计缺失数据; [0029] 6-4) using the historical average estimate missing data in real-time adjustment in conjunction with law;

[0030] 6-5)采用历史平均法估计缺失数据。 [0030] 6-5) using the historical average method to estimate missing data. [0031] 上述连续缺失时间间隔,选择15分钟、30分钟的时间节点是因为: [0031] The lack of continuous time intervals, selecting 15 minutes, 30 minutes because node:

[0032] 这是由缺失数据估计方法的效果决定的,当连续缺失时间间隔小于15分钟时,随机漫步法为最优估计方法;当连续缺失时间间隔大于等于15分钟小于等于30分钟,历史平均结合实时调整法为最优估计方法;当连续缺失时间间隔大于30分钟时,历史平均法为最优估计方法。 [0032] This is determined by the effect of the missing data estimation method, when the continuous time interval is less than 15 minutes deletions, random walk method for the optimal estimation method; deleted when the continuous time interval less than or equal to 15 minutes 30 minutes, the historical average real-time adjustment in conjunction with the optimal estimation method; missing when the continuous time intervals greater than 30 minutes, most of the historical average estimation method.

[0033] 步骤I)中,一定时间间隔为固定式车辆检测器的原始数据采集时间间隔;动态交通流基础数据是交通流量、车辆行驶平均速度、平均时间占有率三个属性变量数据的总称; [0033] Step I), a time interval for the vehicle stationary detector raw data acquisition time interval; dynamic traffic based data traffic, the average speed of the vehicle is traveling, the average time occupancy three generic name attribute of the data variable;

[0034] 步骤2-2)中,日期时间戳错误数据定义为数据日期时间戳不等于有效值的数据; In [0034] Step 2-2), error data is defined as the date stamp date stamp data is not equal to the effective value of the transactions;

[0035] 步骤2-3)中,重复数据定义为与上一原始数据采集时间间隔的交通流基础数据属性变量值相同的数据; In [0035] Step 2-3), and repeats the same data defining the value of data traffic based data acquisition variable properties over a time interval with the original data;

[0036] 所述有效数据和无效数据指有效交通流基础数据和无效交通流基础数据,交通流基础数据属性变量具有一致的有效性。 The [0036] valid data and invalid data traffic flow refers to the effective basis of basic data and invalid data traffic flow, traffic data base attribute variables have the same effectiveness.

[0037] 步骤2-5)中,单车道交通流基础数据阈值法检验的方法如下: [0037] Step 2-5), the traffic lane based method of data threshold is examined as follows:

[0038] a)交通流量阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); [0038] a) traffic threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7);

[0039] b)车辆行驶平均速度阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); [0039] b) with the average vehicle speed threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7);

[0040] c)平均时间占有率阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); [0040] c) the average time occupancy threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7);

[0041] d)相邻速度变化率阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7),其中相邻时间间隔(t-1,t)速度变化率τ t_ljt的计算公式如下: [0041] d) adjacent to the rate of change of the threshold test, if the threshold value exceeds the range, the data is marked as invalid, proceeds to step 2-7), wherein adjacent time interval (t-1, t) of the rate of change of τ t_ljt Calculated as follows:

[0042] Vu =Minji5^I [0042] Vu = Minji5 ^ I

[0043] 式中,Xt—表示车辆行驶平均速度序列,t = 1,2,3···;相邻速度变化率阈值检验的对象为快速路交通流基础数据; [0043] wherein, Xt- sequence representing the vehicle running average speed, t = 1,2,3 ···; adjacent target speed change rate threshold test for Traffic Flow based transactions;

[0044] 上述步骤a)〜d)中,各阈值范围需要基于历史交通流基础数据的统计来进行标定。 [0044] The above-described steps a) ~d), each threshold range needs to be calibrated based on historical traffic statistics data base.

[0045] 步骤2-6)中,单车道交通流基础数据交通流理论法检验的方法如下: [0045] Step 2-6), the underlying data traffic lane traffic flow theory method is examined as follows:

[0046] a)判断是否满足车辆行驶平均速度值等于零而交通流量值不等于零,若满足,则标记为无效数据,转入步骤2-7); [0046] a) judges whether the vehicle running speed is equal to zero and the average value is not zero traffic flow, if yes, the data is marked as invalid, go to step 2-7);

[0047] b)判断是否满足交通流量值等于零而车辆行驶平均速度值不等于零,若满足,则标记为无效数据,转入步骤2-7); [0047] b) determining whether traffic flow satisfies is equal to zero and the vehicle speed with an average value not equal to zero, if yes, the data is marked as invalid, go to step 2-7);

[0048] c)判断是否满足交通流量值和车辆行驶平均速度值都等于零而平均时间占有率值不等于零,若满足,则标记为无效数据,转入步骤2-7); [0048] c) determining whether traffic flow satisfies value and vehicle speed value equal to zero, with the average time of occupancy and the average value is not equal to zero, if yes, the data is marked as invalid, go to step 2-7);

[0049] d)判断是否满足平均时间占有率值为零而交通流量值大于最大可能交通量值,若满足,则标记为无效数据,转入步骤2-7),其中最大可能交通量qmax的计算公式如下: [0049] d) determining whether the average value of the occupancy time zero and satisfy the traffic flow of traffic is greater than the maximum possible value, if yes, then the data is marked as invalid, proceeds to step 2-7), where the maximum possible traffic qmax Calculated as follows:

ΙΟχνχΔ ΙΟχνχΔ

[0050] qmax=^~0- [0050] qmax = ^ ~ 0-

[0051] 式中,qMX——表示原始数据采集时间间隔内平均时间占有率为零时的最大可能交通量(辆/小时),V—表示车辆行驶平均速度(千米/小时),Δ。 [0051] wherein, qMX-- represents the maximum possible amount of traffic (vehicles / hour) during the average time interval within the original data acquisition time share zero, V- representing the vehicle running average speed (km / h), Δ. 一表示平均时间占有率精度(% ),I—表示行驶车辆平均有效车长(米); A represents the average time occupancy accuracy (%), I- average effective running of the vehicle represents the vehicle length (m);

[0052] e)判断车道交通流密度计算值是否超过最大交通流密度,若满足,则标记为无效数据,转入步骤2-7),其中车道交通流密度k的计算公式如下: [0052] e) judging whether a lane of traffic density calculated value exceeds the maximum traffic density, if yes, the data is marked as invalid, proceeds to step 2-7), wherein the calculated lane of traffic density k as follows:

[0053] k = q/v [0053] k = q / v

[0054] 式中,k—表示车道交通流密度(辆/车道/千米),q—表示车道交通流量(辆/小时),V——表示车辆行驶平均速度(千米/小时);最大交通流密度需要基于历史交通流基础数据的统计来进行标定; [0054] wherein, k- represents a lane of traffic density (units / lane / one thousand meters), q- represents a traffic lane (veh / h), V-- representing the vehicle running average speed (km / h); Maximum traffic density traffic flow needs based on historical statistical data base to be calibrated;

[0055] f)若平均时间占有率不为零,判断平均有效车长计算值是否超过一定的阈值范围,若满足,则标记该数据记录为无效,转入步骤2-7),其中平均有效车长I的计算公式如下: [0055] f) If the average time occupancy is not zero, it is determined whether the average effective length Calcd vehicle exceeds a certain threshold range, if yes, the data recording is marked invalid, go to step 2-7), wherein the average effective long car formula I as follows:

Figure CN102800197AD00091

[0057] 式中,I—表示平均有效车长(米),O—表示平均时间占有率(% ),V—表示车辆行驶平均速度(千米/小时),Q——表示车道交通流量(辆/小时);平均有效车长阈值范围需要基于历史交通流基础数据的统计和交通流构成进行标定。 [0057] In the formula, I- car represents the average effective length (m), O- represents the average time occupancy (%), V- representing the vehicle running average speed (km / h), Q - represents a lane traffic ( vehicles / hour); the average effective vehicle length required threshold range based on statistical and historical traffic flow data constitute the basis of traffic flow calibration.

[0058] 步骤3)中,汇集时间间隔可根据需要定义为任意原始数据采集时间间隔的整数倍。 In [0058] Step 3), a collection interval may be an integral multiple of the acquisition time interval is defined according to any of the raw data required.

[0059] 步骤4)中,时间汇集指的是在有效性检验的基础上,将原始数据采集时间间隔的分车道交通流基础数据汇集成较长时间间隔的分车道交通流基础数据,具体方法为:依次对各车道的交通流基础数据进行时间汇集,若汇集时间间隔内不存在有效数据,则标记该车道时间汇集结果为无效数据,若存在有效数据,则单车道交通流基础数据时间汇集计算公式如下: [0059] Step 4), the time of collection refers to a validity check on the basis of the lane-dividing the raw traffic data acquisition interval data base together into a lane-based traffic data for a long time interval, a specific method is: sequentially traffic basic data of each lane time together, if the collection of valid data does not exist in the time interval, the mark of the lane time for the aggregation result is invalid data, if valid data is present, then the lane traffic basic data time together Calculated as follows:

Figure CN102800197AD00092

[0063] 式中,q-表不较长时间间隔内车道交通流量(辆/小时),-表不较长时 Table no longer - [0063] In the formula, Q- longer time interval table without traffic lanes (units / hour),

间间隔内第i个原始数据采集时间间隔的有效交通流量(辆/小时),n——表示较长时间间隔内原始数据采集时间间隔的有效数据记录数,N——表示较长时间间隔内期望的原始数据采集时间间隔数据记录数,V—一表示较长时间间隔内车道车辆行驶平均速度(千米/小时),Vi——表示较长时间间隔内第i个原始数据采集时间间隔的有效车辆行驶平均速度(千米/小时),ο—表示较长时间间隔内车道平均时间占有率),Oi—表示较长时间间隔内第i个原始数据采集时间间隔的有效时间占有率)。 Interval between the i-th effective traffic (vehicles / hour) of the raw data acquisition time interval, n-- represents valid data over longer time intervals the raw data acquisition time interval the number of records, N-- represents a longer time interval desired original data acquisition time interval of data records, V- represents a longer time interval vehicle travel lane average speed (km / h), Vi-- longer time interval represents the i-th original data acquisition time interval the effective average speed of the vehicle running (km / h), ο- lane represents the time interval longer average time occupancy), Oi- longer represents the effective occupancy time interval i-th original data acquisition time interval).

[0064] 步骤5)中,空间汇集指的是在时间汇集基础上,将分车道交通流基础数据汇集成单个车辆行驶方向断面的交通流基础数据,具体方法如下: [0064] Step 5), the space refers to the collection of collection time on the basis of the lane-dividing the basic data traffic flow together into a single section of the vehicle traveling direction based data traffic, as follows:

[0065] a)若进行空间汇集的分车道交通流基础数据均为无效数据,则标记该单个车辆行驶方向断面空间汇集结果为无效数据,若存在有效数据,则转入b); [0065] a) If the traffic lane dividing space collection of basic data are invalid data, the single mark is a cross-sectional direction of vehicle travel as the result invalid data collection space, if there is valid data, the process proceeds to B);

[0066] b)基于有效数据的分车道交通流基础数据空间汇集计算公式如下: [0066] b) together is calculated as follows based on traffic flow data lane-dividing spatial valid data:

Figure CN102800197AD00101

[0070] 式中,q—表示单个车辆行驶方向断面交通流量(辆/小时),qi—表示第i个车道的有效交通流量(辆/小时),n——表示车道有效数据记录数,N——表示单个车辆行驶方向断面总的车道数,V—表示单个车辆行驶方向断面车辆行驶平均速度(千米/小时),Vi—表示第i个车道的有效车辆行驶平均速度(千米/小时),ο—表示单个车辆行驶方向断面平均时间占有率(%),0i—表示第i个车道的有效时间占有率(%)。 [0070] wherein, q- represents a single cross-sectional direction of travel of the vehicle traffic (vehicles / hour), qi- represents the i-th effective traffic lanes (units / hour), n-- represents the number of valid data records lanes, N - represents the total number of lanes in a single direction of travel of the vehicle section, V- represents a single cross-sectional direction of the vehicle with the vehicle running average speed (km / h), Vi- effectively represents the i-th vehicle traveling lane average speed (km / h ), ο- represents a single vehicle traveling direction sectional average time occupancy (%), 0i- represents the effective time of the i th lane occupancy (%).

[0071] 步骤6-2)中,随机漫步法估计缺失数据的表达式为: In [0071] Step 6-2), the random walk method for the estimation expression missing data:

_2] _2]

[0073] 式中,Xi——表示t时刻交通流基础数据的估计值,Xt^1——表示t-Ι时刻交通流基础数据的真实值; [0073] wherein, Xi-- time value t represents an estimate of the traffic data stream based, Xt ^ 1-- represent the true value of the t-Ι time traffic data base;

[0074] 步骤6-4)中,历史平均结合实时调整法估计缺失数据的表达式为: In [0074] Step 6-4), in conjunction with the historical average real-time adjustment method to estimate the missing data for the expression:

Figure CN102800197AD00102

[0078] 式中,;——表示第Cli天t时刻交通流基础数据的估计值,——表示第Cli天t-Ι时刻交通流基础数据的真实值,1)4和7^7^4—分别表示为第Cli天t-Ι和t时刻交通流基础数据的历史平均值,和及^——分别表示第d(g天t-ι和t时刻交通流基础数据的真实值,N——表示固定时间窗口长度; [0078] wherein; - represents Cli day estimate based data traffic flow at time t, - represent the true value of the t-Ι Cli day time traffic data base, 1) and 4 ^ 7 ^ 7 4 - respectively for the first Cli historical average traffic data base and t-Ι days at time t, and, and ^ - respectively denote d (g t-ι days and the true value at time t based data traffic, N- - represents a fixed time window length;

[0079] 步骤6-5)中,历史平均法估计缺失数据的表达式为: In [0079] Step 6-5), the historical average method to estimate missing data expression is:

[0080] XtA =Vf4X tA』N [0080] XtA = Vf4X tA "N

M M

[0081] 式中,尤,4——表示第Cli天t时刻交通流基础数据的估计值,丨“——表示第Cl(^j)天的t时刻交通流基础数据的真实值,N—表示固定时间窗口长度。[0082] 技术效果: [0081] wherein, in particular, represents the estimated value of 4-- t Cli day time base data traffic, | 'or - represent the true value at time t of Cl (^ j) days basic data traffic flow, N- It represents a fixed time window length [0082] The technical effects:

[0083] 本发明提供了一种城市道路路段动态交通流基础数据的预处理方法,针对城市道路路段固定式车辆检测器采集数据及不同类型道路交通流运行特点,确定了具有统计意义的阈值指标及交通流运行三参数(流量、速度、占有率)之间的约束关系,在此基础上,提供了一套全面的数据有效性检验规则;从时间和空间两个维度对数据进行汇集处理,生成了多层次类型的交通流基础数据,为不同应用需求的数据服务奠定了基础;通过对不同程度的数据缺失时间间隔的识别,采用相对最优的方法估计缺失数据,最大限度地保证了数据的完整性和可靠性。 [0083] The present invention provides a method of pre-urban road sections based dynamic traffic flow data, for Urban road vehicle stationary detector data acquisition and operating characteristics of different types of traffic flow, the threshold is determined statistically significant indicators and traffic flow constraints between three parameters (flow, speed, occupancy), on this basis, provide a comprehensive set of validation rules data; the data collection process from the two dimensions of time and space, generating a multi-level type of traffic flow data base, it laid the foundation for data service needs of different applications; identified by different levels of data deletion time interval, using the best methods to estimate the relative lack of data, to ensure maximum data the integrity and reliability. 本发明方法为城市道路路段交通流数据预处理提供了一套全面、综合、具有层次性的解决方案,可以满足交通运营管理、交通规划决策、科研研究、公共服务等领域对于交通流基础数据的需求,同时也为构建城市道路交通信息采集系统和交通综合信息平台提供了技术支持。 The method of the invention Urban road traffic flow data preprocessing provides a comprehensive, integrated, with a hierarchy of solutions to meet the transportation field operations management, transportation planning decisions, scientific research, public service and other basic data for traffic flow of demand, but also for the construction of urban road traffic information collection systems and traffic information platform to provide comprehensive technical support.

附图说明 BRIEF DESCRIPTION

[0084] 图I是一种城市道路路段动态交通流基础数据的预处理方法流程图; [0084] FIG. I is a method for pre-urban road sections based dynamic traffic flow data flowchart;

[0085] 图2是采集数据断面位置图; [0085] FIG. 2 is a cross-sectional position of FIG data collection;

[0086] 图3是2035号断面某时间段流量-速度时间序列图; [0086] FIG. 3 is a flow period of a number of section 2035 - FIG velocity time series;

[0087] 图4是北京市西二环2046断面2009年9月4日中午流量-速度时间序列图; [0087] Figure 4 is a cross-section 2046 West Second Ring Road in Beijing in September 2009 at noon on the 4th traffic - speed time series chart;

[0088] 图5是北京市西二环2041断面2009年9月11_12日速度-平均有效车长关系图; [0088] Figure 5 is a cross-section 2041 West Second Ring Road in Beijing in September 2009 11 _ the 12th rate - the average effective vehicle length relationship diagram;

[0089] 图6是车辆行驶平均速度MAE图; [0089] FIG. 6 is a view of a vehicle with MAE average speed;

[0090] 图7是速度MAPE图; [0090] FIG. 7 is a speed MAPE;

[0091] 图8是速度RMSE图。 [0091] FIG. 8 is a speed RMSE FIG.

具体实施方式 Detailed ways

[0092] 下面结合附图和具体实施例对本发明作进一步说明。 [0092] conjunction with the accompanying drawings and the following specific embodiments of the present invention will be further described.

[0093] 本发明采用北京市道路交通数据采集系统微波车辆检测器采集的快速路与主干道交通流基础数据进行一种城市道路路段动态交通流基础数据的预处理方法的实例验证。 [0093] The present invention employs Beijing trunk road and expressway traffic data stream based microwave vehicle detector collected traffic data collection system for example to verify a pretreatment Urban road traffic dynamic data base. 交通流基础数据是交通流量、车辆行驶平均速度、平均时间占有率三个属性变量数据的总称,交通流量为单位时间间隔内的交通量值,车辆行驶平均速度为单位时间间隔内车辆平均地点瞬时速度,平均时间占有率为有车通过的时间占车辆检测器采集时间间隔的百分t匕。 Traffic flow data is the basis of traffic flow, vehicle travel average speed, average time to share three attributes a general term variable data, traffic is traffic value per unit time interval, an average speed of vehicles traveling average vehicle location unit time interval instantaneous speed, the average time of occupancy time accounts by car vehicle detector percent acquisition time interval t dagger. 北京市道路交通数据采集系统设置的原始数据采集时间间隔为两分钟,快速路数据采集系统对平均时间占有率小于I %的数据全部记录为I %,而主干道数据采集系统对平均时间占有率小于I%的数据则全部记录为零。 Beijing's road traffic data in the original data set of the data acquisition system acquisition time interval of two minutes, fast channel data acquisition system for the average time of less than I% share of all recorded as I%, while the main road data acquisition system for the average time share the data is less than I% of all the records is zero.

[0094] 如附图I所示,具体实施步骤如下: [0094] As shown in FIG. I, the specific implementation steps are as follows:

[0095] 步骤I.以两分钟原始数据采集时间间隔获取城市快速路和主干道的分车道动态交通流基础数据,并进行时间规整,将原始数据集中错乱的时间戳属性规整为最近相邻的整分钟时刻,生成标准的两分钟数据。 [0095] Step I. to two minutes to obtain the original data acquisition time interval dynamic lane divided traffic flow data base of urban expressways and main roads, and time warping, the original data set is structured disorder timestamp attribute nearest neighbors time full minute, two minutes to generate the standard data.

[0096] 步骤2.分车道交通流基础数据有效性检验,依次对各车道的交通流基础数据进行有效性检验,其检验结果包含两类数据,分别为有效交通流基础数据和无效交通流基础数据,简称为有效数据和无效数据,且交通流基础数据属性变量具有一致的有效性。 [0096] Step 2. sub-lane traffic flow basic data validation, followed by basic traffic flow data for each lane of the validity test, the test results contain two types of data, are the basis of effective traffic flow data and invalid traffic flow basis data, referred to as valid and invalid data, attribute data base and traffic flow variables with the same effectiveness. 本实施例通过对数据记录增加一个数据有效性属性来进行区分,默认为有效数据,属性值为1,若检验结果为无效数据,则属性值变为O,具体步骤如下: In this embodiment, increases the effectiveness of the data record attribute data to distinguish a default valid data, the attribute value 1, if the test result is invalid data, the attribute value becomes O, the following steps:

[0097] 2-1)数据非空检验,若数据为空,则标记为无效数据,转入步骤2-7); [0097] 2-1) a non-null test data, if the data is empty, the data marked invalid, go to step 2-7);

[0098] 2-2)日期时间戳有效性检验,若为日期时间戳错误数据,则标记为无效数据,转入步骤2-7),其中日期时间戳错误数据定义为数据日期时间戳不等于有效值的数据,数据时间戳的有效值,即基准时间值,是数据采集系统预先设定的数据上传时刻序列; [0098] 2-2) validity check date stamp, the date stamp if the erroneous data, the data is marked as invalid, proceeds to step 2-7), where the date stamp error data is defined as data is not equal to the date stamp RMS data, the effective value of the time stamp data, i.e., reference time value, the time series data is a data acquisition system to upload a preset;

[0099] 2-3)非重复数据检验,若为重复数据,则标记为无效数据,转入步骤2-7),其中重复数据定义为与上一原始数据采集时间间隔的交通流基础数据属性变量值相同的数据; [0099] 2-3) of unique data checking, if the repetitive data, the data is marked as invalid, proceeds to step 2-7), wherein the data repetition based data traffic is defined as a collection of raw data and the time interval attribute the value of the same data variable;

[0100] 2-4)车辆存在检验,若交通流基础数据属性变量值均为零,则标记为无效数据,转入步骤2-7); Testing the presence of [0100] 2-4) of the vehicle, if the traffic attribute data based variable values ​​are zero, the data marked invalid, go to step 2-7);

[0101] 2-5)单车道交通流基础数据阈值法检验。 Test [0101] 2-5) based data traffic lane threshold method. 阀值法检验的核心思想为基于实际采集的交通流数据,依据数学统计规律合理制定固定式车辆检测器采集交通流基础数据属性变量的极大和极小值,包括如下检验规则: The core idea of ​​the threshold test method is based on the actual traffic flow data collection, according to the laws of mathematical statistics and reasonable development of stationary vehicle detectors collect the maximum and minimum values ​​of the underlying data traffic attribute variables, including inspection rules as follows:

[0102] a)交通流量阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); [0102] a) traffic threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7);

[0103] b)车辆行驶平均速度阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); [0103] b) with the average vehicle speed threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7);

[0104] c)平均时间占有率阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); [0104] c) the average time occupancy threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7);

[0105] d)相邻速度变化率阈值检验,若超过阈值范围,则标记为无效数据,转入步骤 [0105] d) adjacent to the rate of change of the threshold test, if the threshold value exceeds the range, the data is marked as invalid, proceeds to step

2-7),其中相邻时间间隔(t-1,t)速度变化率τ t_ljt的计算公式如下: 2-7), in which adjacent time intervals is calculated (t-1, t) rate of change of τ t_ljt follows:

[0106] rt_ht =Minji5^I [0106] rt_ht = Minji5 ^ I

[0107] 式中,Xt——表示车辆行驶平均速度序列,t = 1,2,3···; [0107] wherein, Xt-- sequence representing the vehicle running average speed, t = 1,2,3 ···;

[0108] 本实施例基于北京市城市快速路与主干道微波车辆检测器采集的历史交通流数据的统计对各阈值范围进行了标定,快速路单车道动态交通流基础数据的最大与最小交通量(辆)、车辆行驶平均速度(千米/小时)、平均时间占有率检验(%)、相邻速度变化率(无量纲)阀值范围分别为[0,80]、[0,120], [0,100], [0.4,1];主干道单车道动态交通流基础数据的最大与最小交通量、车辆行驶平均速度、平均时间占有率阀值范围分别为[0,70]、[0,120]、[0,100]。 [0108] The present embodiment is based on historical traffic statistics data and trunk Beijing Urban Expressway microwave vehicle detector values ​​acquired for each threshold calibration range, the maximum and minimum traffic lane freeway dynamic traffic flow data base (vehicles), the vehicle running average speed (km / h), the average time occupancy test (%), rate of change of adjacent (dimensionless) are threshold range [0,80], [0, 120], [0,100 ], [0.4,1]; the maximum and minimum dynamic traffic lane roads based data traffic, the average speed of the vehicle is traveling, the average time occupancy threshold range, respectively [0,70], [0, 120], [0,100 ]. 由于城市快速路的交通流属于连续性交通流范畴,具有相邻较短时间间隔内,交通流状态变化较小的特点,因此设定了相邻速度变化率检验规则,而城市主干道的交通流属于间断性交通流范畴,交通流状态波动性较大,不适用该检验规则; Because of urban freeway traffic flow is a continuous flow of traffic areas, with a short time interval within an adjacent, small changes in the state of traffic flow characteristics, thus setting the rate of change of adjacent speed inspection rules, and the city's main road traffic traffic belonging to the intermittent stream visible volatile state traffic flow, the test is not applicable rule;

[0109] 2-6)单车道交通流基础数据交通流理论法检验。 Test [0109] 2-6) single-lane traffic flow traffic flow theory underlying data method. 基于交通流理论的动态交通流数据有效性检验的基本思想是利用交通流运行参数之间的一致性关系来构建动态交通数据有效性检验规则,固定式车辆检测器实际采集的交通流基础数据除了某一个属性变量为零,另外的属性变量不能为零的情况外,还存在三个属性变量值均不为零,但是属性变量之间不一致的情形。 Based on the basic idea of ​​dynamic traffic data on traffic flow theory of validation is to build a dynamic traffic data validation rules using the consistency of the relationship between the operating parameters of traffic flow, vehicle stationary detector actual collection of basic data in addition to traffic flow one variable attribute is zero, where the outer variable additional attributes can not be zero, there are also three attributes variable values ​​are not zero, but the case of inconsistencies between the properties of the variable. 具体检验环节的方法如下: Specific aspects of the test method is as follows:

[0110] a)判断是否满足车辆行驶平均速度值等于零而交通流量值不等于零,若满足,则标记为无效数据,转入步骤2-7);[0111] b)判断是否满足交通流量值等于零而车辆行驶平均速度值不等于零,若满足,则标记为无效数据,转入步骤2-7); [0110] a) judges whether the vehicle running speed is equal to zero and the average value is not zero traffic flow, if yes, the data is marked as invalid, go to step 2-7); [0111] b) determining whether traffic flow satisfies equal to zero and vehicle travel average speed value not zero, if yes, the data is marked as invalid, go to step 2-7);

[0112] c)判断是否满足交通流量值和车辆行驶平均速度值都等于零而平均时间占有率值不等于零,若满足,则标记为无效数据,转入步骤2-7); [0112] c) determining whether traffic flow satisfies value and vehicle speed value equal to zero, with the average time of occupancy and the average value is not equal to zero, if yes, the data is marked as invalid, go to step 2-7);

[0113] d)判断是否满足平均时间占有率值为零而交通流量值大于最大可能交通量值,若满足,则标记为无效数据,转入步骤2-7),该规则是针对固定式车辆检测器占有率采集精度不足产生的交通流量大于零而平均时间占有率为零的情况设置,如北京市主干道数据采集系统对平均时间占有率小于I %的数据则全部记录为零。 [0113] d) the average time occupancy is judged whether the value is greater than zero and the maximum possible value of the traffic flow of traffic value, if yes, then the data is marked as invalid, proceeds to step 2-7), the rule is for a fixed type vehicle occupancy detector insufficient traffic generated acquisition accuracy greater than zero to zero where the average time occupancy setting, such as a data acquisition system trunk Beijing average time occupancy is less than I% of all recorded data is zero. 其中最大可能交通量qmax的计算公式如下: Which is calculated the maximum possible amount of traffic qmax as follows:

Figure CN102800197AD00131

[0115] 式中,qmax——表示原始数据采集时间间隔内平均时间占有率为零时的最大可能交通量(辆/小时),V—表示车辆行驶平均速度(千米/小时),Λ。 [0115] wherein, qmax-- represents the maximum possible amount of traffic (vehicles / hour) during the average time interval within the original data acquisition time share zero, V- representing the vehicle running average speed (km / h), Λ. 一表示平均时间占有率精度(% ),I—表示行驶车辆平均有效车长(米); A represents the average time occupancy accuracy (%), I- average effective running of the vehicle represents the vehicle length (m);

[0116] e)判断车道交通流密度计算值是否超过最大交通流密度,若满足,则标记为无效数据,转入步骤2-7),其中车道交通流密度k的计算公式如下: [0116] e) judging whether a lane of traffic density calculated value exceeds the maximum traffic density, if yes, the data is marked as invalid, proceeds to step 2-7), wherein the calculated lane of traffic density k as follows:

[0117] k = q/v [0117] k = q / v

[0118] 式中,k—表示车道交通流密度(辆/车道/千米),q—表示车道交通流量(辆/小时),V—表示车辆行驶平均速度(千米/小时); [0118] wherein, k- represents a lane of traffic density (units / lane / one thousand meters), q- represents a traffic lane (veh / h), V- representing the vehicle running average speed (km / h);

[0119] f)若平均时间占有率不为零,判断平均有效车长计算值是否超过一定的阈值范围,若满足,则标记为无效数据,转入步骤2-7),其中平均有效车长I的计算公式如下: [0120] [0119] f) If the average time occupancy is not zero, it is determined whether the average effective length Calcd vehicle exceeds a certain threshold range, if yes, the data is marked as invalid, proceeds to step 2-7), wherein the effective average vehicle length I is calculated as follows: [0120]

Figure CN102800197AD00132

[0121] 式中,I—表示平均有效车长(米),O—表示平均时间占有率(% ),V—表示车辆行驶平均速度(千米/小时),Q——表示车道交通流量(辆/小时); [0121] In the formula, I- car represents the average effective length (m), O- represents the average time occupancy (%), V- representing the vehicle running average speed (km / h), Q - represents a lane traffic ( vehicles / hour);

[0122] 上述步骤d)中,参数I标定为北京市主干道行驶的小汽车平均有效车长2. 4米,参数Δ。 [0122] The step d), the calibration parameter I is the main road with the vehicle car Beijing average effective length 2.4 m, the parameter Δ. 标定为I ; Calibration is I;

[0123] 上述步骤e)中,基于北京市城市快速路与主干道微波车辆检测器采集的历史交通流数据的统计,快速路最大交通流密度值标定为100辆/车道/千米,主干道最大交通流密度值标定为70辆/车道/千米; [0123] The step e), based on statistical historical traffic flow data of Beijing city expressways and main roads microwave vehicle detectors collected expressway traffic flow density maximum value scaled to 100 / lane / km main road The maximum density of traffic flow calibration value for the 70 / lane / km;

[0124] 上述步骤f)中,基于北京市城市快速路与主干道微波车辆检测器采集的历史交通流数据和北京市快速路和主干道的车辆构成分析,快速路和主干道平均有效车长阀值范围标定为[2. 4,18]; [0124] The step f), composition analysis based on historical traffic data and trunk Beijing Urban Expressway microwave vehicle detector and acquired Beijing expressways and main roads vehicles, roads and expressways effective average vehicle length calibration of threshold range [4,18 2];

[0125] 2-7)若所有车道交通流基础数据均已完成有效性检验,则转入步骤3,否则进行下一车道交通流基础数据有效性检验。 [0125] 2-7) If all lanes of traffic flow data base have been completed validation, then go to step 3, otherwise the validity test the next lane traffic flow data base.

[0126] 步骤3.判断原始数据累计采集时间间隔是否等于汇集时间间隔,若满足,则将原始数据累计采集时间间隔归零,并转入步骤4,其中汇集时间间隔可根据需要定义为任意原始数据采集时间间隔的整数倍。 [0126] Step 3. Analyzing the original data acquisition time interval is equal to the total integration time interval, if satisfied, the original data acquisition time interval cumulative zero, and proceeds to step 4, wherein a collection interval may be defined as needed in any of the original an integer multiple of the data acquisition interval.

[0127] 步骤4.分车道交通流基础数据时间汇集,即得到各车道汇集时间间隔的有效交通流基础数据,交通流数据时间汇集指的是对固定式车辆检测器采集的较短时间间隔(如I分钟)的交通流量、车辆行驶平均速度、时间占有率数据汇集为较长时间间隔(如2分钟或5分钟)的交通流量、车辆行驶平均速度数据与时间占有率数据。 [0127] Step 4. traffic lane-dividing the basic data collection time, i.e., the effective data of each traffic lane based aggregation time interval, traffic flow data collection time means a vehicle detector fixed short time interval collected ( the I min) of the traffic flow, the average traveling speed of the vehicle, data gathering traffic occupancy time interval is longer (e.g. 2 minutes or 5 minutes), the vehicle traveling speed data and the average time occupancy data. 将原始数据采集时间间隔的分车道交通流基础数据汇集成较长时间间隔的分车道交通流基础数据,具体方法为:依次对各车道的交通流基础数据进行时间汇集,若汇集时间间隔内不存在有效数据,则标记该车道时间汇集结果为无效数据,若存在有效数据,则单车道交通流基础数据时间汇集计算公式如下: The lane-based traffic flow data of the original data acquisition time interval together into a traffic lane-dividing the basic data of a long time interval, the specific method is: the basic data sequentially traffic lane of each collection time, if the integration time interval without valid data, then the lane mark time for the aggregation result is invalid data, if valid data is present, then the base data traffic lane collection time is calculated as follows:

Figure CN102800197AD00141

[0131] 式中,q-表不较长时间间隔内车道交通流量(辆/小时),-表不较长时 Table no longer - [0131] In the formula, Q- longer time interval table without traffic lanes (units / hour),

间间隔内第i个原始数据采集时间间隔的有效交通流量(辆/小时),n——表示较长时间间隔内原始数据采集时间间隔的有效数据记录数,N——表示较长时间间隔内期望的原始数据采集时间间隔数据记录数,V—一表示较长时间间隔内车道车辆行驶平均速度(千米/小时),Vi——表示较长时间间隔内第i个原始数据采集时间间隔的有效车辆行驶平均速度(千米/小时),ο—表示较长时间间隔内车道平均时间占有率),Oi—表示较长时间间隔内第i个原始数据采集时间间隔的有效时间占有率)。 Interval between the i-th effective traffic (vehicles / hour) of the raw data acquisition time interval, n-- represents valid data over longer time intervals the raw data acquisition time interval the number of records, N-- represents a longer time interval desired original data acquisition time interval of data records, V- represents a longer time interval vehicle travel lane average speed (km / h), Vi-- longer time interval represents the i-th original data acquisition time interval the effective average speed of the vehicle running (km / h), ο- lane represents the time interval longer average time occupancy), Oi- longer represents the effective occupancy time interval i-th original data acquisition time interval).

[0132] 步骤5.分车道交通流基础数据单个车辆行驶方向断面空间汇集。 [0132] Step 5. lane-based traffic data from a single vehicle traveling direction of the space section together. 若空间汇集结果为无效数据,即得到单个车辆行驶断面汇集时间间隔的有效交通流基础数据,若空间汇集结果为无效数据,则将其视为缺失数据,转入步骤6。 If the result is invalid data collection space, i.e., effectively a single vehicle traffic data base section together with the time interval, if the result is invalid data collection space, it is treated as missing data, it proceeds to step 6. 交通流基础数据空间汇集指的是在时间汇集基础上,将汇集时间间隔的分车道交通流量、车辆行驶平均速度数据、平均时间占有率数据汇集成单个车辆行驶方向断面的交通流量、车辆行驶平均速度与平均时间占有率,具体方法如下: Basic data collection traffic flow space refers to the collection of time based on the traffic lane-dividing pooled time interval, the average traveling speed of the vehicle data, the average time occupancy data together into a single section of the traffic traveling direction of the vehicle, the vehicle travel average the average time occupancy rate, as follows:

[0133] a)若进行空间汇集的分车道交通流基础数据均为无效数据,则标记该单个车辆行驶方向断面空间汇集结果为无效数据,若存在有效数据,否则转入b); [0133] a) If the traffic lane dividing space collection of basic data are invalid data, the single mark is a cross-sectional direction of vehicle travel as the result invalid data collection space, if there is valid data, into or B);

[0134] b)基于有效数据的分车道交通流基础数据空间汇集计算公式如下: [0134] b) together is calculated as follows based on traffic flow data lane-dividing spatial valid data:

Figure CN102800197AD00142

[0138] 式中,q—表示单个车辆行驶方向断面交通流量(辆/小时),qi—表示第i个车道的有效交通流量(辆/小时),n——表示车道有效数据记录数,N——表示单个车辆行驶方向断面总的车道数,V—表示单个车辆行驶方向断面车辆行驶平均速度(千米/小时),Vi—表示第i个车道的有效车辆行驶平均速度(千米/小时),ο—表示单个车辆行驶方向断面平均时间占有率(%),0i—表示第i个车道的有效时间占有率(%)。 [0138] wherein, q- represents a single cross-sectional direction of travel of the vehicle traffic (vehicles / hour), qi- represents the i-th effective traffic lanes (units / hour), n-- represents the number of valid data records lanes, N - represents the total number of lanes in a single direction of travel of the vehicle section, V- represents a single cross-sectional direction of the vehicle with the vehicle running average speed (km / h), Vi- effectively represents the i-th vehicle traveling lane average speed (km / h ), ο- represents a single vehicle traveling direction sectional average time occupancy (%), 0i- represents the effective time of the i th lane occupancy (%).

[0139] 步骤6.单个车辆行驶方向断面动态交通流缺失数据估计,即得到单个车辆行驶断面汇集时间间隔的估计交通流基础数据,具体实施方法如下: [0139] Step 6. A single cross-sectional direction of travel of the vehicle dynamic traffic estimate missing data, i.e. to obtain an estimated data traffic flow basis with a single vehicle collection time interval section, the specific embodiments as follows:

[0140] 6-1)判断连续缺失时间间隔,若时间间隔小于15分钟,则转入步骤6-2),否则转入步骤6-3); [0140] 6-1) is determined continuously deletion interval, if the time interval is less than 15 minutes, the process proceeds to step 6-2), otherwise it proceeds to step 6-3);

[0141] 6-2)采用随机漫步法估计缺失数据,其计算表达式为: [0141] 6-2) using a random walk method to estimate missing data, the calculation of the expression:

_2] _2]

[0143] 式中,;^——表示t时刻交通流基础数据的估计值,Xt^1——表示t-Ι时刻交通流基础数据的真实值; [0143] In the formula,; ^ - represents the estimated value at time t based traffic data, Xt ^ 1-- represent the true value of the t-Ι time traffic data base;

[0144] 6-3)若连续缺失时间间隔大于等于15分钟,小于等于30分钟,则转入步骤6_4),否则转入步骤6-5); [0144] 6-3) If time intervals greater than or equal consecutive deletions 15 minutes, 30 minutes or less, the process proceeds to step 6_4), otherwise proceeds to step 6-5);

[0145] 6-4)采用历史平均结合实时调整法估计缺失数据,其计算表达式为: [0145] 6-4) is estimated using the historical average combined real-time adjustment method for missing data, the calculation expression is:

[0146] [0146]

Figure CN102800197AD00151

[0147] [0147]

[0148] [0148]

[0149] 式中,Xt4i——表示第di天t时刻交通流基础数据的估计值,X(t_l)A——表示第Cli天t-Ι时刻交通流基础数据的真实值,M—分别表示为第Cii天t-Ι和t时刻交通流基础数据的历史平均值,_Λ和及^——分别表示第d(g天t-ι和t时刻交通流基础数据的真实值,N——表示固定时间窗口长度; [0149] wherein, Xt4i-- represents the estimated value of the underlying data traffic di day time t, X (t_l) A-- represents t-Ι Cli day real time based data traffic value, M-, respectively Cii historical average for the first data traffic flow based day time t and t-Ι, _Λ and and ^ - respectively denote d (g t-ι days and the true value at time t based data traffic, N-- represents a fixed time window length;

[0150] 6-5)采用历史平均法估计缺失数据,其计算表达式为: [0150] 6-5) using the historical average method to estimate missing data, the calculation expression is:

[0151] [0151]

Figure CN102800197AD00152

[0152] 式中,尤,4——表示第Cli天t时刻交通流基础数据的估计值,丨“——表示第Cl(^j)天的t时刻交通流基础数据的真实值,N—表示固定时间窗口长度; [0152] wherein, in particular, represents the estimated value of 4-- t Cli day time base data traffic, | 'or - represent the true value at time t of Cl (^ j) days basic data traffic flow, N- It represents a fixed time window length;

[0153] 本实施例定义历史平均结合实时调整法和历史平均法中的固定时间窗口长度均为30天。 [0153] This embodiment defines in conjunction with the historical average of the historical and real-time adjustment method averaging fixed time window length is 30 days.

[0154] 基于实际采集数据,下述内容将从数据有效性检验结果、车辆行驶平均速度阈值检验效果、相邻速度变化率阈值检验效果、平均有效车长阈值检验效果以及不同缺失时间间隔下缺失数据估计方法效果评估四个方面进一步说明和验证一种城市道路路段动态交通流基础数据的预处理方法。 [0154] Based on the actual data collected, the following data from the content validity test results, the average speed of the vehicle with the effect of the threshold test, the threshold rate of change of adjacent test results, the average effective length of the car threshold and the test results at different time intervals missing missing data estimation method to evaluate the effect of four further illustrate and validate the pretreatment method for dynamic traffic flow in urban road sections basic data. [0155] 数据有效性检验结果 [0155] Data validation results

[0156] 为了检验动态交通流基础数据有效性检验规则的性能,本实施例列举了北京市快速路(编号为2010、2014、2017、2035)和主干道(编号为22002、22005、24003、24005)如图2所示,一共8个断面2009年9月I日至9月30日的数据有效性规则检验结果,如表I所示: [0156] In order to test the effectiveness of inspection rules based dynamic traffic performance data, the present embodiment include the Expressway Beijing (No. 2010,2014,2017,2035) and trunk (numbered 22002,22005,24003,24005 ) 2, a total of 8 data section I September 2009 to September 30th validation rules test results as shown in table I:

[0157] 表I北京市快速路和主干道8个断面9月份动态交通流基础数据有效性检验结果汇总表 [0157] Table I expressways and main roads of Beijing section 8 September dynamic traffic flow data base validation results summary

[0158] [0158]

Figure CN102800197AD00161

[0159] 由表I可以看出规则4、规则5和规则6较容易检验出错误数据,其他规则检验出的错误量相对较少。 [0159] As can be seen from Table I Rule 4, and Rule 5 rules 6 easier inspection data error, the error amount of the other rule checking relatively small. 以下将重点说明阈值法检验中的车辆行驶平均速度阈值检验、相邻速度变化率检验以及交通流法检验中的平均有效车长检验的效果: The following description will focus on the threshold method of testing vehicle traveling an average speed threshold test, the rate of change of adjacent speed test and the average effective traffic law test vehicle length of the test results:

[0160] 车辆行驶平均速度阈值检验效果 [0160] with an average vehicle speed threshold test results

[0161] 。 [0161].

[0162] 表2和错误! [0162] Table 2 and error! 未找到引用源。 Reference source not found. 3分别为北京市快速路2035号断面内环进京方向第一车道2009年9月8日16:06:00至17:00:00的交通流基础数据及对应的流量速度图。 3 are expressway Beijing section No. 2035 Beijing direction of the first inner lane 8 September 2009 traffic flow data base and the corresponding flow velocity map of 16:06:00 to 17:00:00. 其中,。 among them,. 表2中标记为灰色的3条记录为车辆行驶平均速度阈值规则检验出的错误或可疑的数据,并对应于图3圆圈当中的三点。 Table 2 marked with gray 3 records the average vehicle speed threshold rule is checked out errors or questionable data, corresponding to Figure 3 and a circle among the three points.

[0163] 表2北京市快速路2035断面内环进京方向第一车道交通流基础数据表 [0163] Table 2 expressway Beijing section 2035 Beijing direction of the first inner lane traffic flow data base table

[0164] [0164]

Figure CN102800197AD00162

[0165] [0165]

Figure CN102800197AD00171

[0166] 相邻速度变化率阈值检验效果 [0166] adjacent speed change rate threshold value test results

[0167] 表3和图4显示是2035号断面2009年9月4日中午微波车辆检测器实际采集的交通流基础数据和流量-速度时间序列关系图。 [0167] Table 3 and FIG. 4 shows a cross-sectional No. 2035 September 2009 4 at noon traffic data based microwave vehicle detector actually collected and flow - speed time sequence diagram. 其中,表4中标记为灰色的两条记录为经过相邻时间速度变化率阈值检验后出现的错误或可疑数据,并对应于图4圆圈当中的两点。 Wherein, marked in gray in Table 4 of two adjacent recording the time elapsed after the error rate of change occurring or threshold value check suspect data, and corresponds to FIG. 4 circles among the two points.

[0168] 表5实时数据与不同交通状态类别聚类中心欧式距离表 [0168] Table 5 real-time traffic data with different status categories cluster centers Euclidean distance table

[0169] [0169]

Figure CN102800197AD00172

[0170] [0170]

Figure CN102800197AD00181

[0171] 平均有效车长阈值检验效果 [0171] The average effective vehicle length threshold test results

[0172] 错误! [0172] Error! 未找到引用源。 Reference source not found. 5为2035号断面2009年9月11日与9月12日车辆行驶平均速度与平均有效车长之间的散点关系图。 5 is a cross-section No. 2035 September 11, 2009 scatter the relationship between the average rate and the average effective vehicle length and 12 September vehicle. 由图可以看出,实际行驶车辆的平均有效车长极少大于18米或小于2. 4米,三角标志表示的数据为经过平均有效车长阈值检验后出现的错误或可疑数据,表6列举了部分平均有效车长检验后出现的错误数据。 As can be seen from the figure, the actual traveling of the vehicle is greater than the average effective minimum vehicle length 18 m or less than 2.4 m, errors or questionable data triangle represented by average effective after the vehicle length threshold test appear in Table 6 include some erroneous data average effective length of the car after the test.

[0173] 表7北京市西二环2041断面2009年9月11日平均有效车长检验错误数据示例 [0173] Table 7 West Second Ring Road of Beijing 2041 section September 11, 2009 the average effective vehicle length data validation error example

Figure CN102800197AD00182
Figure CN102800197AD00191

[0176] 不同缺失时间间隔下缺失数据估计方法效果评估 [0176] at different time intervals missing missing data estimation Effect Evaluation

[0177] 为了评估不同缺失时间间隔下各类方法的估计性能,本实施例在保证所选断面9月份数据完整性的基础上,抽取所选断面10月份的部分数据,人工生成十一种不同的连续缺失水平,分别为连续缺失10分钟、15分钟、20分钟、30分钟、45分钟、I小时、3小时、5小时、8小时、12小时和I天,并将估计结果与真实数据进行对比,分别计算估计平均绝对误差(MAE)、平均绝对百分误差(MAPE)以及平均绝对误差标准差(RMSE)值,部分计算结果如图6、7、8所示。 [0177] To assess the different deletion time interval estimation performance under various types of methods, in this embodiment, selected to ensure the integrity of the data base section September on the extracted data portion of the selected section in October, artificially generated eleven different continuous deletions level, continuous deletions were 10 minutes, 15 minutes, 20 minutes, 30 minutes, 45 minutes, I hour, 3 hours, 5 hours, 8 hours, 12 hours and days I, and the estimation result with the real data comparison, calculates the mean absolute error estimation (MAE), mean absolute percentage error (MAPE) standard deviation and the mean absolute error (RMSE) value, calculated results are shown partially 6,7,8 FIG.

[0178] 根据图表综合分析,得出第7位连续缺失水平(即3个小时)之前,随机漫步和历史平均结合实时调整法总体优于历史平均法;同时在第2位连续缺失水平之前随机漫步优于历史平均结合实时调整法;第2位连续缺失水平至第7位连续缺失水平之间,历史平均结合实时调整法总体优于随机漫步模型;第7位连续缺失水平之后,历史平均估计法优于随机漫步估计法和历史平均结合实时调整法。 [0178] The comprehensive analysis of the chart, before drawing level position 7 consecutive deletions (i.e., three hours), and the historical average random walk method was better than real-time adjustment in conjunction with the historical average method; simultaneously randomizing prior to the deletion of two consecutive horizontal Walking binding than the historical average real-time adjustment method; between two consecutive second level to the bit deletion of 7 consecutive deletions level, the historical average binding was better than the random walk model in real-time adjustment method; after 7 consecutive deletions bit level, the historical average estimated random walk estimation method is better than the historical average and combine real-time adjustment method.

Claims (8)

1. ー种城市道路路段动态交通流基础数据的预处理方法,其特征是包括如下步骤: 1)以一定时间间隔获取固定式车辆检测器采集的分车道交通流基础数据,规整数据的时间戳属性为最近相邻的整分钟时刻; 2)分车道交通流基础数据有效性检验,依次对各车道的交通流基础数据进行有效性检验: 2-1)数据非空检验,若数据为空,则标记为无效数据,转入步骤2-7); 2-2)日期时间戳有效性检验,若为日期时间戳错误数据,则标记为无效数据,转入步骤2-7); 2-3)非重复数据检验,若为重复数据,则标记为无效数据,转入步骤2-7); 2-4)车辆存在检验,若交通流基础数据属性变量值均为零,则标记为无效数据,转入步骤2-7); 2-5)单车道交通流基础数据阈值法检验; 2-6)单车道交通流基础数据交通流理论法检验; 2-7)若所有车道交通流基础数据均已完成有效性检验 1. ー pretreatment methods Urban road traffic dynamically based data, characterized by comprising the steps of: 1) obtaining a time interval based data traffic lane-dividing stationary vehicle detector collected data structured stamp recent property is adjacent to the full minute time; 2) sub-lane traffic flow basic data validation, followed by basic traffic flow data for each lane validity test: 2-1) test data is not empty, if data is empty, the data marked invalid, go to step 2-7); 2-2) validity check date stamp, the date stamp if the erroneous data, the data is marked as invalid, go to step 2-7); 2-3 ) non-repetitive test data, if the repetitive data, the data is marked as invalid, go to step 2-7); verifying the presence of 2-4) of the vehicle, if the traffic attribute data based variable values ​​are zero, the data marked invalid , proceeds to step 2-7); 2-5) based lane data traffic threshold is examined; 2-6) based data traffic lane traffic flow theory is examined; 2-7) if all lanes of traffic data base have been completed validation 则转入步骤3),否则进行下一车道交通流基础数据有效性检验; 3)判断原始数据累计采集时间间隔是否等于汇集时间间隔,若满足,则将原始数据累计采集时间间隔归零,并转入步骤4); 4)分车道交通流基础数据时间汇集,即得到各车道汇集时间间隔的有效交通流基础数据; 5)分车道交通流基础数据单个车辆行驶方向断面空间汇集,若空间汇集结果为有效数据,即得到单个车辆行驶断面汇集时间间隔的有效交通流基础数据,若空间汇集结果为无效数据,则将其视为缺失数据,转入步骤6); 6)单个车辆行驶方向断面交通流缺失数据估计,即得到单个车辆行驶断面汇集时间间隔的估计交通流基础数据: 6-1)判断连续缺失时间间隔,若时间间隔小于15分钟,则转入步骤6-2),否则转入步骤6-3); 6-2)采用随机漫歩法估计缺失数据; 6-3)若连续缺失时间间隔大于 Then go to step 3), otherwise the validity test the next lane traffic flow data base; 3) determine the original data acquisition time interval is equal to the cumulative collection interval, if satisfied, the cumulative raw data collection interval to zero, and go to step 4); 4) traffic lane-dividing the basic data collection time, i.e., the effective data of each traffic lane based collection time interval; 5) spatial traffic lane-sectional data of a single vehicle traveling direction based pooled together if space the result is valid data, i.e., effectively a single vehicle traffic data base section together with the time interval, if the result is invalid data collection space, it is treated as missing data, go to step 6); section 6) a single vehicle traveling direction missing data traffic flow estimation, i.e., obtain an estimated traffic flow data from a single vehicle traveling section based collection interval: 6-1) is determined continuously deletion interval, if the time interval is less than 15 minutes, the process proceeds to step 6-2), otherwise turn to the step 6-3); 6-2) were randomly diffuse ho method to estimate missing data; 6-3) If the interval is greater than the continuous deletions 于15分钟,小于等于30分钟,则转入步骤6-4),否则转入步骤6-5); 6-4)采用历史平均结合实时调整法估计缺失数据; 6-5)采用历史平均法估计缺失数据。 In 15 minutes, 30 minutes or less, the process proceeds to step 6-4), otherwise it proceeds to step 6-5); 6-4) adjusted in real time using the historical average binding method to estimate missing data; 6-5) using the historical average method estimate the missing data.
2.根据权利要求I所述的ー种城市道路路段动态交通流基础数据的预处理方法,其特征为: 步骤I)中,一定时间间隔为固定式车辆检测器的原始数据采集时间间隔;动态交通流基础数据是交通流量、车辆行驶平均速度、平均时间占有率三个属性变量数据的总称; 步骤2-2)中,日期时间戳错误数据定义为数据日期时间戳不等于有效值的数据; 步骤2-3)中,重复数据定义为与上一原始数据采集时间间隔的交通流基础数据属性变量值相同的数据; 所述有效数据和无效数据指有效交通流基础数据和无效交通流基础数据,交通流基础数据属性变量具有一致的有效性。 The pretreatment method according ー I type of dynamic traffic data based Urban road claim, wherein: step I), the predetermined time interval is fixed vehicle detector raw data acquisition time interval; dynamic traffic flow based traffic data, the average speed of the vehicle is traveling, the average time occupancy variable data attribute three generic name; step 2-2), error data is defined as the date stamp date stamp data is not equal to the effective value of the transactions; in step 2-3), repeating the same data is defined as basic data traffic attribute value of the variable time interval data acquisition and data on an original; valid data and invalid data traffic flow refers to the effective basis of traffic flow data and invalid data base , traffic flow data base attribute variables with the same effectiveness.
3.根据权利要求I所述的ー种城市道路路段动态交通流基础数据的预处理方法,其特征为步骤2-5)中,单车道交通流基础数据阈值法检验的方法如下: a)交通流量阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); b)车辆行驶平均速度阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); c)平均时间占有率阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7); d)相邻速度变化率阈值检验,若超过阈值范围,则标记为无效数据,转入步骤2-7),其中相邻时间间隔(t-1,t)速度变化率T t_p t的计算公式如下: The pretreatment method according ー I type of dynamic traffic city road segment data base as claimed in claim, wherein step 2-5), the traffic lane based method of data threshold is examined as follows: a) Transport traffic threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7); B) with an average vehicle speed threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7) ; c) the average time occupancy threshold test, if the threshold value exceeds the range, the data is marked as invalid, go to step 2-7); d) adjacent to the rate of change of the threshold test, if the threshold value exceeds the range, the data is marked as invalid, proceeds to step 2-7), wherein adjacent time interval (t-1, t) calculated rate of change of T t_p t as follows:
Figure CN102800197AC00031
式中,Xt—表示车辆行驶平均速度序列,t = 1,2,3…;相邻速度变化率阈值检验的对象为快速路交通流基础数据; 上述步骤a)〜d)中,各阈值范围需要基于历史交通流基础数据的统计来进行标定。 Wherein, xt- sequence representing the vehicle running average speed, t = 1,2,3 ...; adjacent target speed change rate threshold test for Traffic Flow based transactions; the above step a) ~d), each of the threshold range needs based on statistical historical traffic flow data base to be calibrated.
4.根据权利要求I所述的ー种城市道路路段动态交通流基础数据的预处理方法,其特征为步骤2-6)中,单车道交通流基础数据交通流理论法检验的方法如下: a)判断是否满足车辆行驶平均速度值等于零而交通流量值不等于零,若满足,则标记为无效数据,转入步骤2-7); b)判断是否满足交通流量值等于零而车辆行驶平均速度值不等于零,若满足,则标记为无效数据,转入步骤2-7); c)判断是否满足交通流量值和车辆行驶平均速度值都等于零而平均时间占有率值不等于零,若满足,则标记为无效数据,转入步骤2-7); d)判断是否满足平均时间占有率值为零而交通流量值大于最大可能交通量值,若满足,则标记为无效数据,转入步骤2-7),其中最大可能交通量q_的计算公式如下: The pretreatment ー species Dynamic Urban road traffic data base according to claim I, wherein the step 2-6), the method lane traffic flow theory underlying data traffic flow is examined as follows: a ) is judged whether the vehicle running speed is equal to zero and the average value is not zero traffic flow, if yes, the data is marked as invalid, go to step 2-7); b) determining whether traffic flow satisfies is equal to zero and the vehicle running speed average value is not is equal to zero, if yes, the data is marked as invalid, go to step 2-7); c) determining whether traffic flow satisfies the mean value and the traveling speed of the vehicle is equal to zero and the value of the average time occupancy value is not equal to zero, if yes, then labeled invalid data, proceeds to step 2-7); D) is judged whether the average time occupancy value greater than zero and the maximum possible traffic flow of traffic value, if yes, then the data is marked as invalid, go to step 2-7) which is calculated the maximum possible amount of traffic q_ as follows:
Figure CN102800197AC00032
式中,——表示原始数据采集时间间隔内平均时间占有率为零时的最大可能交通量(辆/小吋),V—表示车辆行驶平均速度(千米/小吋),a。 Wherein - denotes the maximum possible amount of traffic (vehicles / hour inch) average time when the original data acquisition time interval of zero occupancy, V- representing the vehicle running average speed (kilometers / hour inch), a. 一表示平均时间占有率精度(% ),I—表示行驶车辆平均有效车长(米); e)判断车道交通流密度计算值是否超过最大交通流密度,若满足,则标记为无效数据,转入步骤2-7),其中车道交通流密度k的计算公式如下: k = q/v 式中,k——表示车道交通流密度(辆/车道/千米),q——表示车道交通流量(辆/小吋),V——表示车辆行驶平均速度(千米/小吋);最大交通流密度需要基于历史交通流基础数据的统计来进行标定; f)若平均时间占有率不为零,判断平均有效车长计算值是否超过一定的阈值范围,若满足,则标记该数据记录为无效,转入步骤2-7),其中平均有效车长I的计算公式如下: A represents the average time occupancy accuracy (%), I- average effective running of the vehicle represents the vehicle length (m); whether e) determining a lane of traffic density calculated value exceeds the maximum traffic density, if yes, the data is marked as invalid, transfer the step 2-7), wherein the calculated lane of traffic density k as follows: k = q / v where, K-- represents a lane of traffic density (units / lane / one thousand meters), q - represents a lane traffic (vehicles / hour inches), V-- representing the vehicle running average speed (km / h small inches); the maximum traffic density need to be calibrated based on statistical historical traffic flow data base; f) if the average time share is not zero , it is determined whether the average effective length Calcd vehicle exceeds a certain threshold range, if yes, the data recording is marked invalid, go to step 2-7), wherein the calculated average effective length I of the vehicle as follows:
Figure CN102800197AC00033
式中,I—表示平均有效车长(米),0—表示平均时间占有率(% ),V——表示车辆行驶平均速度(千米/小时),q——表示车道交通流量(辆/小时);平均有效车长阈值范围需要基于历史交通流基础数据的统计和交通流构成进行标定。 Wherein, I- car represents the average effective length (m), 0 represents the average time occupancy (%), V-- representing the vehicle running average speed (km / h), q - represents a lane traffic (vehicles / hours); the average effective vehicle length required threshold range based on statistical and historical traffic flow data constitute the basis of traffic flow calibration.
5.根据权利要求I所述的ー种城市道路路段动态交通流基础数据的预处理方法,其特征为步骤3)中,汇集时间间隔可根据需要定义为任意原始数据采集时间间隔的整数倍。 According to claim I of the method of pretreatment kind ー dynamic Urban road traffic data base, wherein in step 3), the collection of time interval may be an integral multiple of the acquisition time interval is defined according to any of the raw data required.
6.根据权利要求I所述的ー种城市道路路段动态交通流基础数据的预处理方法,其特征为所述步骤4)中,时间汇集指的是在有效性检验的基础上,将原始数据采集时间间隔的分车道交通流基础数据汇集成较长时间间隔的分车道交通流基础数据,具体方法为:依次对各车道的交通流基础数据进行时间汇集,若汇集时间间隔内不存在有效数据,则标记该车道时间汇集结果为无效数据,若存在有效数据,则单车道交通流基础数据时间汇集计算公式如下: 6. The pretreatment method according ー I type of dynamic traffic city road segment data base as claimed in claim 4 wherein said step), the time of collection refers to a validity check on the basis of raw data acquisition time interval based data traffic lane-dividing lane-dividing traffic flow together into a longer interval data basis, specific methods of: sequentially traffic lane data of each basic time collection, the collection interval if the valid data does not exist , then the lane mark time for the aggregation result is invalid data, if valid data is present, then the base data traffic lane collection time is calculated as follows:
Figure CN102800197AC00041
式中,q——表示较长时间间隔内车道交通流量(辆/小吋),Qi——表示较长时间间隔内第i个原始数据采集时间间隔的有效交通流量(辆/小时),n——表示较长时间间隔内原始数据采集时间间隔的有效数据记录数,N——表示较长时间间隔内期望的原始数据采集时间间隔数据记录数,V——表示较长时间间隔内车道车辆行驶平均速度(千米/小吋),Vi——表示较长时间间隔内第i个原始数据采集时间间隔的有效车辆行驶平均速度(千米/小时),O——表示较长时间间隔内车道平均时间占有率),Oi——表示较长时间间隔内第i个原始数据采集时间间隔的有效时间占有率)。 Wherein, q-- represents a longer time interval within a lane of traffic flow (cars / hour inch), Qi-- represents an effective traffic flow within a time interval longer the i-th original data acquisition time interval (veh / h), n - represents a time interval longer valid data in the original data acquisition time interval the number of records, N-- represents a desired longer time interval of the original data acquisition time interval of data records, V-- represents a longer time interval within a lane of the vehicle with average speed (kilometers / hour inch), Vi-- represents an effective vehicle for a longer time interval i-th original data acquisition time interval running average speed (km / h), O-- represents a longer time interval the average time occupancy lane), Oi-- longer time interval represents the i-th original collection effective occupancy time interval).
7.根据权利要求I所述的ー种城市道路路段动态交通流基础数据的预处理方法,其特征为所述步骤5)中,空间汇集指的是在时间汇集基础上,将分车道交通流基础数据汇集成单个车辆行驶方向断面的交通流基础数据,具体方法如下: a)若进行空间汇集的分车道交通流基础数据均为无效数据,则标记该单个车辆行驶方向断面空间汇集结果为无效数据,若存在有效数据,则转入b); b)基于有效数据的分车道交通流基础数据空间汇集计算公式如下: The pretreatment method according ー I type of dynamic traffic city road segment data base as claimed in claim 5 wherein said step), the collection space refers to the collection time on the basis of the lane-dividing the traffic flow basic data together into a single section of the vehicle traveling direction based data traffic, as follows: a) If the traffic lane dividing space collection of basic data are invalid data, is marked with the single direction cross-sectional space of the vehicle together result invalid data, if valid data is present, the process proceeds to b); b) on the basis of data traffic lane-dividing space efficient collection of data is calculated as follows:
Figure CN102800197AC00042
式中,q——表示单个车辆行驶方向断面交通流量(辆/小吋),Qi——表示第i个车道的有效交通流量(辆/小时),n——表示车道有效数据记录数,N——表示单个车辆行驶方向断面总的车道数,V——表示单个车辆行驶方向断面车辆行驶平均速度(千米/小吋),Vi——表示第i个车道的有效车辆行驶平均速度(千米/小时),0——表示单个车辆行驶方向断面平均时间占有率),oi——表示第i个车道的有效时间占有率(%)。 Wherein, q-- represents a single cross-sectional direction of travel of the vehicle traffic (vehicles / hour inch), Qi-- represents the i-th effective traffic lanes (units / hour), n-- represents the number of valid data records lanes, N - represents the total number of lanes in a single direction of travel of the vehicle section, V-- represents a single cross-sectional direction of the vehicle with the vehicle running average speed (kilometers / hour inch), Vi-- effectively represents the i-th vehicle traveling lane average speed (in thousands m / hr), represents a single vehicle traveling direction 0-- sectional average time occupancy), oi-- represents the effective time of the i th lane occupancy (%).
8.根据权利要求I所述的ー种城市道路路段交通流基础数据的预处理方法,其特征为所述步骤6-2)中,随机漫歩法估计缺失数据的表达式为:名=ん式中,——表示t时刻交通流基础数据的估计值,Xw——表示t-1时刻交通流基础数据的真实值; 所述步骤6-4)中,历史平均结合实时调整法估计缺失数据的表达式为: 8. ー pretreatment kind Urban road traffic data base as claimed in claim I, wherein said step 6-2), the random diffuse expression ho method is to estimate missing data: name = san formula in - time value t represents an estimate of the underlying data traffic, Xw-- t-1 represents the value of the real time traffic data base; in the step 6-4), real-time adjustment in conjunction with the historical average method to estimate missing data expression is:
Figure CN102800197AC00051
式中,文tA——表示第di天t时刻交通流基础数据的估计值,—表示第Cli天t-1时刻交通流基础数据的真实值,HIST±、t—lldp HIST±tA—分别表示为第Cli天t-1和t时刻交通流基础数据的历史平均值,和XtAl_r>——分别表示第d(g天t-1和t时刻交通流基础数据的真实值,N——表示固定时间窗ロ长度; 所述步骤6-5)中,历史平均法估计缺失数据的表达式为: In the formula, represents text tA-- traffic flow estimated value of the underlying data di day time t - represents the true value Cli basic data traffic time t-1 day, HIST ±, t-lldp HIST ± tA- represent Cli day for the first t-1 and time t based historical average traffic flow data, and XtAl_r> - respectively denote d (g t-1 and time t, the actual value of the basic data traffic days, the N-- represents a fixed ro length of the time window; step 6-5), the method to estimate the historical average expression of missing data:
Figure CN102800197AC00052
式中,——表示第Cii天t时刻交通流基础数据的估计值,xtAj_r>——表示第屯づ)天的t时刻交通流基础数据的真实值,N——表示固定时间窗ロ长度。 Wherein, - t represents Cii day time traffic flow estimated value of the underlying data, xtAj_r> - represents Tun づ) true value at time t-day basic data traffic flow, the N-- represents a fixed length time window ro.
CN201210045464.5A 2012-02-27 2012-02-27 Preprocessing method of road section dynamic traffic stream essential data of urban road CN102800197B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201210045464.5A CN102800197B (en) 2012-02-27 2012-02-27 Preprocessing method of road section dynamic traffic stream essential data of urban road

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201210045464.5A CN102800197B (en) 2012-02-27 2012-02-27 Preprocessing method of road section dynamic traffic stream essential data of urban road

Publications (2)

Publication Number Publication Date
CN102800197A true CN102800197A (en) 2012-11-28
CN102800197B CN102800197B (en) 2014-07-16

Family

ID=47199290

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201210045464.5A CN102800197B (en) 2012-02-27 2012-02-27 Preprocessing method of road section dynamic traffic stream essential data of urban road

Country Status (1)

Country Link
CN (1) CN102800197B (en)

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103064974A (en) * 2013-01-10 2013-04-24 东南大学 Method for cleaning traffic flow data on basis of time-space analysis
CN103632545A (en) * 2013-11-22 2014-03-12 北京世纪高通科技有限公司 Method and device for acquiring traffic information
CN103971516A (en) * 2014-05-13 2014-08-06 紫光捷通科技股份有限公司 Traffic data preprocessing method and road condition detection method
CN104316328A (en) * 2014-05-29 2015-01-28 潍柴动力股份有限公司 Measuring and determining method for vehicle ride comfort
CN104715605A (en) * 2015-02-16 2015-06-17 北京交通大学 VSP-distribution-based traffic operation data and emission data coupling method and system
CN105931458A (en) * 2016-05-05 2016-09-07 杭州智诚惠通科技有限公司 Method for evaluating credibility of road traffic flow detection device

Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001084479A (en) * 1999-09-14 2001-03-30 Matsushita Electric Ind Co Ltd Method and device for forecasting traffic flow data
CN1560756A (en) * 2004-03-09 2005-01-05 北京交通大学 Intelligent traffic data processing method
US20050091176A1 (en) * 2003-10-24 2005-04-28 Mitsubishi Denki Kabushiki Kaisha Forecasting apparatus
CN1645402A (en) * 2005-01-19 2005-07-27 北京交通大学 Road traffic flow data quality controlling method and apparatus
JP2007241429A (en) * 2006-03-06 2007-09-20 Sumitomo Electric Ind Ltd Traffic flow parameter calculation system, method, and program
US20080221783A1 (en) * 2006-04-18 2008-09-11 Gregory Jensen Boss Intelligent redirection of vehicular traffic due to congestion and real time performance metrics
CN101286270A (en) * 2008-05-26 2008-10-15 北京捷讯畅达科技发展有限公司 Traffic flow forecasting method combining dynamic real time traffic data
CN101794510A (en) * 2009-12-30 2010-08-04 北京世纪高通科技有限公司 Data processing method and device of floating cars
CN102169630A (en) * 2011-03-31 2011-08-31 上海电科智能系统股份有限公司 Quality control method of road continuous traffic flow data

Patent Citations (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2001084479A (en) * 1999-09-14 2001-03-30 Matsushita Electric Ind Co Ltd Method and device for forecasting traffic flow data
US20050091176A1 (en) * 2003-10-24 2005-04-28 Mitsubishi Denki Kabushiki Kaisha Forecasting apparatus
CN1560756A (en) * 2004-03-09 2005-01-05 北京交通大学 Intelligent traffic data processing method
CN1645402A (en) * 2005-01-19 2005-07-27 北京交通大学 Road traffic flow data quality controlling method and apparatus
JP2007241429A (en) * 2006-03-06 2007-09-20 Sumitomo Electric Ind Ltd Traffic flow parameter calculation system, method, and program
US20080221783A1 (en) * 2006-04-18 2008-09-11 Gregory Jensen Boss Intelligent redirection of vehicular traffic due to congestion and real time performance metrics
CN101286270A (en) * 2008-05-26 2008-10-15 北京捷讯畅达科技发展有限公司 Traffic flow forecasting method combining dynamic real time traffic data
CN101794510A (en) * 2009-12-30 2010-08-04 北京世纪高通科技有限公司 Data processing method and device of floating cars
CN102169630A (en) * 2011-03-31 2011-08-31 上海电科智能系统股份有限公司 Quality control method of road continuous traffic flow data

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
成玉栋 等: "城市道路交通实施状况监测系统数据库设计", 《道路交通与安全》, vol. 9, no. 6, 31 December 2009 (2009-12-31), pages 28 - 33 *
朱雷雷 等: "干线公路交通流数据有效性检验规则", 《东南大学学报(自然科学版)》, vol. 41, no. 1, 31 January 2011 (2011-01-31), pages 194 - 198 *
陆振波 等: "城市道路交通流监测数据最优汇集时间间隔分析", 《东南大学学报(自然科学版)》, vol. 42, no. 5, 30 September 2012 (2012-09-30), pages 1000 - 1005 *

Cited By (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103064974A (en) * 2013-01-10 2013-04-24 东南大学 Method for cleaning traffic flow data on basis of time-space analysis
CN103064974B (en) * 2013-01-10 2016-05-04 东南大学 Traffic flow data cleaning method based on space-time analysis
CN103632545A (en) * 2013-11-22 2014-03-12 北京世纪高通科技有限公司 Method and device for acquiring traffic information
CN103971516A (en) * 2014-05-13 2014-08-06 紫光捷通科技股份有限公司 Traffic data preprocessing method and road condition detection method
CN103971516B (en) * 2014-05-13 2016-04-20 紫光捷通科技股份有限公司 Traffic data preprocess method and road conditions detection method
CN104316328A (en) * 2014-05-29 2015-01-28 潍柴动力股份有限公司 Measuring and determining method for vehicle ride comfort
CN104316328B (en) * 2014-05-29 2017-09-05 潍柴动力股份有限公司 A kind of assay method of vehicle ride comfort
CN104715605A (en) * 2015-02-16 2015-06-17 北京交通大学 VSP-distribution-based traffic operation data and emission data coupling method and system
CN105931458A (en) * 2016-05-05 2016-09-07 杭州智诚惠通科技有限公司 Method for evaluating credibility of road traffic flow detection device
CN105931458B (en) * 2016-05-05 2019-02-12 杭州智诚惠通科技有限公司 A kind of method of road traffic flow detection device reliability assessment

Also Published As

Publication number Publication date
CN102800197B (en) 2014-07-16

Similar Documents

Publication Publication Date Title
Dia An object-oriented neural network approach to short-term traffic forecasting
Zhan et al. Urban link travel time estimation using large-scale taxi data with partial information
Li et al. Evaluation of speed-based travel time estimation models
Santhi et al. Spatial Calibration and Temporal Validation of Flow for Regional Scale Hydrologic Modeling 1
Bhouri et al. Isolated versus coordinated ramp metering: Field evaluation results of travel time reliability and traffic impact
CN104157139B (en) A kind of traffic congestion Forecasting Methodology and method for visualizing
CN101794507B (en) Method for evaluating macroscopic road network traffic state based on floating car data
Rakha et al. Trip travel-time reliability: issues and proposed solutions
CN101373559A (en) Method for evaluating city road net traffic state based on floating vehicle data
CN102163225B (en) A fusion evaluation method of traffic information collected based on micro blogs
CN102629418B (en) Fuzzy kalman filtering-based traffic flow parameter prediction method
CN102169630B (en) Quality control method of road continuous traffic flow data
CN103279534B (en) Travel frequently OD distribution estimation method based on the mass transit card passenger of intelligent public transportation system data
CN102081846B (en) Expressway charge data track matching based traffic state recognition method
“Brian” Park et al. Stochastic optimization for sustainable traffic signal control
Tang et al. Dynamic analysis of traffic time series at different temporal scales: A complex networks approach
CN102968901B (en) Method for acquiring regional congestion information and regional congestion analyzing device
Gao et al. A Markov‐based road maintenance optimization model considering user costs
CN104408925B (en) Crossing evaluation of running status method based on display radar
CN101694743B (en) Method and device for predicting road conditions
CN102306450B (en) Layout method for traffic detectors of sparse road network
CN101739824B (en) Data fusion technology-based traffic condition estimation method
Long Cheu et al. Mobile sensor and sample-based algorithm for freeway incident detection
Abdel-Aty et al. Considering various ALINEA ramp metering strategies for crash risk mitigation on freeways under congested regime
Tong et al. Development of driving cycles for motorcycles and light-duty vehicles in Vietnam

Legal Events

Date Code Title Description
C06 Publication
C10 Entry into substantive examination
C14 Grant of patent or utility model
CF01