CN103886747B - Road traffic running similarity metrics - Google Patents

Road traffic running similarity metrics Download PDF

Info

Publication number
CN103886747B
CN103886747B CN201410095850.4A CN201410095850A CN103886747B CN 103886747 B CN103886747 B CN 103886747B CN 201410095850 A CN201410095850 A CN 201410095850A CN 103886747 B CN103886747 B CN 103886747B
Authority
CN
China
Prior art keywords
time series
vector
traffic
missing
step
Prior art date
Application number
CN201410095850.4A
Other languages
Chinese (zh)
Other versions
CN103886747A (en
Inventor
祁宏生
王殿海
许骏
金盛
马东方
叶盈
韦薇
蔡正义
郑正非
Original Assignee
浙江大学
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 浙江大学 filed Critical 浙江大学
Priority to CN201410095850.4A priority Critical patent/CN103886747B/en
Publication of CN103886747A publication Critical patent/CN103886747A/en
Application granted granted Critical
Publication of CN103886747B publication Critical patent/CN103886747B/en

Links

Abstract

本发明公开了一种路网中路段交通运行相似度度量方法。 The present invention discloses a road network link traffic running similarity metrics. 本发明利用路网中路段的交通运行参数时间序列,首先对时间序列进行重构,转换成不缺失的值向量、缺失值的下标向量、前两者的长度向量;随后对其进行截断,使得截断后的时间序列长度等同,最后对截断后的时间序列计算其相似性。 The present invention utilizes road traffic network operating parameter time series of sections, the first reconstructed time series, is converted into vectors not missing, the missing value of the subscript a vector, the vector length of the former two; subsequently truncated, such that the length of the time series after truncation equivalents, calculated last time sequence similarity truncated. 本发明在部分数据缺失的情况下仍然能得到结果,可靠性较高。 The present invention still results obtained in the case where the missing part of the data, the higher the reliability.

Description

路段交通运行相似度度量方法 Road traffic running similarity metrics

技术领域 FIELD

[0001] 本发明涉及一种用于城市交通管理中的路段交通运行相似度度量方法,具体来说是及利用路段交通运行时间序列数据对路段的交通运行相似度进行度量的方法。 [0001] The present invention relates to a road traffic in urban traffic management operation similarity metrics, and particularly a method using time series data link traffic running on the road traffic operation performed similarity metric.

背景技术 Background technique

[0002] 由于我国城市的交通拥堵越来越严重,从而识别交通拥堵的原因对交通管理有重要的意义。 [0002] Because of the city's traffic congestion is getting worse, thus identifying the cause of traffic congestion has important implications for traffic management. 而交通拥堵是一定时空范围内的交通现象,会反映在交通拥堵覆盖的连线的交通运行参数的时间序列(典型的例如速度时间序列)上,从而,利用时间序列相似性度量技术识别交通拥堵是一种可行的方法。 Traffic congestion and traffic within a certain time and space of the phenomenon, will be reflected in the operating parameters of the time series covered traffic congestion traffic connection (e.g. a typical velocity time series), so that by the time sequence similarity techniques to identify congestion metric It is a feasible method. 但是由于交通运行时间序列常常存在缺失值,从而,一般的时间序列相似性度量方法无法应用。 However, because there is often a time-series traffic operation missing values, thus, the general time series similarity measure can not be applied.

发明内容 SUMMARY

[0003] 本发明针对现有技术的不足,提供了一种路网中路段交通运行相似度度量方法。 [0003] The present invention addresses deficiencies in the prior art, there is provided a road network link traffic running similarity metrics.

[0004] 本发明方法具体是: [0004] The method of the present invention, in particular:

[0005] 步骤1 :利用移动式交通检测器或者路段人工速度调查,在设定的时间间隔内获得交通运行参数,例如速度、流量、行程时间等,利用Xl]表达,其中i表示第i个时间间隔,j 表示第j个路段,时间序列长度为N,时间序列表达为X1=(Xu,xl2, . . . .xlN)。 [0005] Step 1: With the mobile detector or road traffic speed doing survey to obtain traffic operation parameters within a set time interval, such as speed, traffic, travel time, etc., using Xl] expression, where i denotes the i th time interval, j denotes the j-th link, the time series of length N, the time series is expressed as X1 = (Xu, xl2,... .xlN). 假设要计算路段i和路段j的交通运行相似度。 Suppose you want to calculate the similarity road traffic operations section i and j are.

[0006] 步骤2 :如果有缺失值,则按照缺失值情况,将Xi扩展成新的时间序列ti,每个新的时间序列分为三段:没有缺失的值组成的向量A;、缺失值的下标向量&、前两者的长度组成的向量Q,同理可得Aj、B#卩C。 [0006] Step 2: If there are missing values, according to the missing value, will expand into a new time series Xi ti, each new time series is divided into three sections: no deletion vector composed of values ​​A ;, missing values subscript & vector, the vector length of the former two compositions Q, Similarly available Aj, B # C. Jie .

[0007] 步骤4:求得两个时间序列)ΓjPrj之内A满A」的共同部分^pA'j,同理可得IVi、B'j。 [0007] Step 4: time series obtained by two) A common portions of the full ΓjPrj A "in ^ pA'j, Similarly available IVi, B'j.

[0008] 步骤3 :最终两个可能含有缺失的原始时间序列&和Xj的相似度度量S(XXJ的ifs(X1;Xj) =s(C1;Cj)Xs(ArdA'JXs®'dB'J)。 [0008] Step 3: The final two similarity may contain deletions original time series Xj and & metric S (XXJ of ifs (X1; Xj) = s (C1; Cj) Xs (ArdA'JXs®'dB'J ).

[0009] 至此,两个含有缺失值得交通运行时间序列相似度就得出来了。 [0009] At this point, two containing the deleted sequence similarity worth of traffic have run out of time.

[0010] 本发明的有益效果: [0010] Advantageous effects of the invention:

[0011] 1.对数据类型要求不高,只要是反映交通运行质量的都可以应用; [0011] 1. The type of data do not ask, as long as the operation is a reflection of the quality of traffic can be applied;

[0012] 2.对时间序列数据完整性没有要求。 [0012] 2. The time-series data integrity is not required.

附图说明 BRIEF DESCRIPTION

[0013] 图1为时间序列截断方法示意图。 [0013] FIG. 1 is a schematic time series truncation method.

具体实施方式 Detailed ways

[0014] 下面对本发明做进一步的详细描述。 [0014] The following further detailed description of the invention. 在获得了一个时段内(例如一天)每一个时间间隔(例如5分钟)的交通运行数据Xl](部分可能有缺失),按照下述步骤对两个路段的交通运行时间序列&和X(长度都为N)的相似度进行度量: Within a period obtained in each time interval (e.g., one day) (e.g., 5 minutes) operating data traffic Xl] (likely missing portions) according to the following sequence of steps run time & traffic and X (the length of the two sections We are N) similarity to measure:

[0015] 1)时间序列转换。 [0015] 1) time series conversion. 假设时间序列&中有部分缺失值,则首先将没有缺失的值重新组成一个向量A1;再将缺失数值的下标按照从小到大组成向量B1;将长度组成向量Q,则新的向量f (AdBdQ);举例来说,一个含有缺失值得向量 Suppose the time series & some missing values, no value is first deleted reconstitute a vector A1; then the missing values ​​in ascending index vector composition Bl; the length of the vector composed of Q, then a new vector f ( AdBdQ); for example, a vector comprising deletions worthy

[l,4,2,NaN,4,NaN]的[1,4,2,4],[4,6],[4,6],其中NaN表示数据缺失; 用同样的方法对X,进行转换得到t(ApB^C,);可知转换之后的时间序列长度为N+2。 [L, 4,2, NaN, 4, NaN] in [1,4,2,4], [4,6], [4,6], where NaN represents the missing data; using the same method for X, for converted t (ApB ^ C,); after a length of time series understood convert N + 2.

[0016]2)对于转换之后的两个时间序列X'jPX'j。 [0016] 2) for the two time series X'jPX'j after conversion. 针对AJPAj,将长度较长的向量尾部截断,使得长度和较短向量的长度一致,长度较短的向量保持不变,得到A'pA',,如图1所示; For AJPAj, the longer the length of the tail cut vector, and a shorter length such that the vector consistent shorter length of the vector remains unchanged, resulting A'pA ',, shown in Figure 1;

[0017] 3)针对BjPBj,将长度较长的向量头部截断,使得长度和较短向量的长度一致,长度较短的向量保持不变,得到B'pB',,如图1所示; [0017] 3) for BjPBj, the longer length of the vector cut head, so that the vector length is short and consistent, shorter length of the vector remains unchanged, resulting B'pB ',, shown in Figure 1;

[0018] 4)则两个时间序列的相似度计算方法为S^Xj)= WCoCjXsWdA' )Xs(B,dB'其中,计算方法为: Similarity calculation method according to [0018] 4) for the two time series S ^ Xj) = WCoCjXsWdA ') Xs (B, dB' which is calculated as:

[0019]式中Clk和Cjk为时间序列C满C弟k个元素; Y丁 [0019] wherein Clk time series C and Cjk full brother C k elements; the Y-butoxy

Figure CN103886747BD00041

\ 丁- \ Ding-

[0020] 同理设 [0020] Likewise provided

Figure CN103886747BD00042

其中lk为时间序列A'i中的第k个元素,B'lk为时间序列B' 个元素。 Wherein lk A'i time series of k-th element, B'lk time series of elements B '.

Claims (1)

1.路段交通运行相似度度量方法,其特征在于该方法包括以下步骤: 步骤1.利用移动式交通检测器或者路段人工速度调查,在设定的时间间隔内获得时间间隔内的交通运行参数,包括速度、流量和行程时间,该交通运行参数利用X1]表达,其中i表示第i个时间间隔,j表示第j个路段,时间序列X1=(Xu,xl2,....xlN); 步骤2.按照缺失值情况,将&扩展成新的时间序列Xi',每个新的时间序列分为三段: 没有缺失的值组成的向量A;、缺失值的下标向量&、前两者的长度组成的向量Q; 步骤3.求得两个时间序列&'和X/,并进行截断,求得AJPA,的共同部分A/、A/,同理可得Β/、Β/ ; 1. similarity measure link traffic operation, characterized in that the method comprises the following steps: Step 1. With the mobile detector or road traffic speed doing survey to obtain traffic operation parameters within the time interval within a set time interval, including speed, traffic and travel time, vehicle operating parameters using the X1] expression, where i denotes the i th time interval, j denotes the j-th link, the time series X1 = (Xu, xl2, .... xlN); step 2. the missing value, the & extended to new time series Xi ', each new time series is divided into three sections: there is no missing index vector of the vector a ;, missing values ​​& composition, both before the length of the vector composed of Q; step 3. & determined two time series' and X /, and cut, to obtain AJPA, the common portion a /, a /, Similarly available Β /, Β /;
Figure CN103886747BC00021
步骤4.最终两个可能含有缺失的原始时间序列XjPXj的相似度度量S(XdXj) =s沁,C^XsiA,1 ,A/ )Xs(B1, ,B/ ) s(B/,B/ )类推。 Step 4. The final two original time series XjPXj similarity may contain deletions metric S (XdXj) = s Qin, C ^ XsiA, 1, A /) Xs (B1,, B /) s (B /, B / )analogy.
CN201410095850.4A 2014-03-14 2014-03-14 Road traffic running similarity metrics CN103886747B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410095850.4A CN103886747B (en) 2014-03-14 2014-03-14 Road traffic running similarity metrics

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410095850.4A CN103886747B (en) 2014-03-14 2014-03-14 Road traffic running similarity metrics

Publications (2)

Publication Number Publication Date
CN103886747A CN103886747A (en) 2014-06-25
CN103886747B true CN103886747B (en) 2016-03-09

Family

ID=50955615

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410095850.4A CN103886747B (en) 2014-03-14 2014-03-14 Road traffic running similarity metrics

Country Status (1)

Country Link
CN (1) CN103886747B (en)

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003085493A2 (en) * 2002-03-29 2003-10-16 Agilent Technologies, Inc. Method and system for predicting multi-variable outcomes
CN101930667A (en) * 2009-06-26 2010-12-29 歌乐牌株式会社 Apparatus and method for generating statistic traffic information
CN103020079A (en) * 2011-09-24 2013-04-03 国家电网公司 Industrial data supplementation method
KR101363171B1 (en) * 2012-05-30 2014-02-14 (주)유엠텍 Cosine similarity based expert recommendation technique using hybrid collaborative filtering

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2003085493A2 (en) * 2002-03-29 2003-10-16 Agilent Technologies, Inc. Method and system for predicting multi-variable outcomes
CN101930667A (en) * 2009-06-26 2010-12-29 歌乐牌株式会社 Apparatus and method for generating statistic traffic information
CN103020079A (en) * 2011-09-24 2013-04-03 国家电网公司 Industrial data supplementation method
KR101363171B1 (en) * 2012-05-30 2014-02-14 (주)유엠텍 Cosine similarity based expert recommendation technique using hybrid collaborative filtering

Also Published As

Publication number Publication date
CN103886747A (en) 2014-06-25

Similar Documents

Publication Publication Date Title
WO2013121298A3 (en) System and method for analyzing gis data to improve operation and monitoring of water distribution networks
CN103593430B (en) A segmented target temporal clustering information based on the movement trajectory method
RU2011102443A (en) Methods and systems for dynamically adaptive hierarchical representation of the road network and the laying of routes
CN102169630B (en) Quality control method of road continuous traffic flow data
CN103218398B (en) Intelligent substation SCL (substation configuration description language) file difference comparison method
CN102073785B (en) Daily gas load combination prediction method based on generalized dynamic fuzzy neural network
CN101853261A (en) Network public-opinion behavior analysis method based on social network
CN104156489B (en) Permanent mining method point of the driver based on the driver trajectory
CN101882372B (en) Short road chain real-time traffic information merging method, device and system
CN104361110A (en) Mass electricity consumption data analysis system as well as real-time calculation method and data mining method
CN103476051A (en) Method for evaluating importance of nodes in communication network
CN101996258B (en) Electronic map information grid processing and querying method facilitating quick query
CN103680157A (en) Vehicle queuing overflow anticipation method for city bottleneck road section
CN102831488A (en) Power load forecasting process based on improved grey forecasting method
CN101808351B (en) Method and system for business impact analysis
CN101286178A (en) Electric power automatization system real time database multi-domain group multi-section accomplishing method
CN104715610B (en) Oriented urban transport traffic index calculation method
EA201270589A1 (en) The system and method of modeling of the geological volume of interest
de Groot et al. Vehicle and crew scheduling: solving large real-world instances with an integrated approach
Varia et al. Application of genetic algorithms for joint optimization of signal setting parameters and dynamic traffic assignment for the real network data
CN102890866B (en) Traffic flow speed estimation method based on multi-core support vector regression machine
CN103440404B (en) Degradation Behavior warning method stiffening transverse bridge wind load based side effects of anti-Beam
CN102930111B (en) Scd model file generation method and apparatus substation
Nelson Assessment of harvest blocks generated from operational polygons and forest-cover polygons in tactical and strategic planning
Wang et al. Estimating safety performance trends over time for treatments at intersections in Florida

Legal Events

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