CN101783074A - Method and system for real-time distinguishing traffic flow state of urban road - Google Patents

Abstract

The invention discloses a method and a system for distinguishing urban road traffic flow states in real time, and belongs to the field of urban road traffic monitoring and intelligent traffic identification. The method and the system of the invention divide the traffic flow mode of the urban road into four traffic states, analyze and process the collected traffic flow data, and judge various traffic flow data by establishing the recognition model of the traffic flow mode, thereby accurately obtaining the four traffic states of the urban road traffic flow. The method and the system can make the judgment of the urban road traffic flow more reasonable and accurate, and provide beneficial help for the driver to select the optimal path, thereby relieving the urban traffic jam problem and improving the urban traffic operation efficiency.

Description

A kind of distinguishing urban road traffic flow modes in real time method and system

Technical field

The present invention relates to a kind of distinguishing urban road traffic flow modes in real time method and system, be the differentiation of urban road traffic state, belong to urban highway traffic monitoring and intelligent transportation recognition technology field.

Background technology

Thereby carrying out urban highway traffic induces the distribution alleviation urban traffic blocking that equilibrium traffic is flowed in city road network to have vital role, present common traffic guidance mode is in the through street or traffic guidance screen and variable information boards such as VMS are installed in the major trunk roads roadside, the equipment of this class often shows with the traffic behavior segmentation near several the main trunk roads that link to each other of the form of signal picture, and use red, yellow, green three kinds of colors represent respectively unimpeded, state crowds, blocks the traffic.Wherein, unimpeded state is meant: vehicle ' is not influenced by other vehicles, and the speed average is higher; Congestion state is meant: traffic flow speed is lower, and flow is big, but not all right phenomenon does not appear stopping fully in vehicle; Blocked state is meant: vehicle is in complete halted state.And in existing road traffic state recognition technology, also defined normal condition usually for the domain of traffic congestion state fuzzy set, and it is meant: other vehicles influenced around vehicle ' was subjected to, and the speed average decreases but still higher situation.

But effective application prerequisite of prior art scheme and equipment is: can carry out real-time analysis to the traffic flow data of being gathered and handle, and then can obtain the real-time traffic states of each arterial traffic comparatively exactly.The division of traffic behavior at present with often differentiate with the highway section wagon flow average velocity that detected as foundation, there are the following problems for this mode:

At first, existing City road traffic system is a very complicated big system, only can not depict the complex state of traffic flow complete, exactly with the wagon flow average velocity in each highway section; Secondly, traffic detects data may exist deviation, only may cause erroneous results with one of them or several detection data as the analyzing and processing foundation.Therefore, for more accurate and characterize traffic flow modes all sidedly, in identifying, should consider as much as possible to comprise all kinds of influence factors relevant, and require more influence factor should not produce too much influence the accuracy of computation complexity and net result with traffic flow modes.For this reason, be necessary to seek a kind of new method and carry out the identification of traffic behavior.

Summary of the invention

In order to overcome the deficiencies in the prior art, the invention provides a kind of distinguishing urban road traffic flow modes in real time method and system.The inventive method is at first selected to characterize a plurality of influence factors of urban road traffic flow characteristic and is determined the traffic behavior classification that ultimate demand is divided, then the mutual relationship between each influence factor and the traffic behavior is analyzed the final urban road traffic state comparatively accurately that obtains.The technical scheme that the present invention is adopted for its technical matters of solution is as follows:

A kind of distinguishing urban road traffic flow modes in real time method is sampled to each traffic flow data, and carries out following steps in each traffic flow data sampling instant:

Set up the step of classical territory matter-element and joint territory matter-element: with traffic flow pattern and traffic flow data is object, based on the Matter Analysis method set up the Different Traffic Flows pattern the classical territory matter-element and the joint territory matter-element of corresponding described traffic flow data, and therefrom obtain corresponding classical thresholding and joint thresholding;

Set up the matter-element model step of traffic flow pattern to be discriminated: the described traffic flow data of gathering according to real-time detection, set up the matter-element model of traffic flow pattern to be discriminated, and obtain the measured value of the influence factor of described traffic flow pattern to be discriminated, wherein, described influence factor is the corresponding traffic flow data of traffic flow pattern to be discriminated;

Ask for correlation function value step: according to classical territory matter-element and the classical thresholding that saves the territory matter-element and the joint thresholding of described traffic flow data, and, ask for the correlation function value of the measured value of each influence factor about various traffic flow patterns according to the measured value of the described influence factor of described traffic flow pattern to be discriminated;

Correlation function value weighted sum step:, ask for the weighted sum of each influence factor about the correlation function value of every kind of traffic flow pattern according to the measured value of described each influence factor correlation function value about various traffic flow patterns;

Comparison step: about the size between the weighted sum of the correlation function value of every kind of traffic flow pattern, the selection pairing traffic flow pattern of the maximum wherein is as the traffic behavior of current traffic flow pattern by more described each influence factor; And,

The issuing traffic state step: issue the traffic behavior of described current traffic flow pattern, it is a kind of state in unimpeded state, normal condition, congestion state and the blocked state.

Preferably, the place is stated in the step of setting up classical territory matter-element and joint territory matter-element, described traffic flow pattern comprises: unimpeded state, normal condition, congestion state and four kinds of traffic behaviors of blocked state, and the classical territory matter-element of the pairing traffic flow data of various traffic flow patterns is set up as follows:

Wherein, P _0jBe traffic flow pattern, and j=(1,2 ..., m), m=4, P ₀₁, P ₀₂, P ₀₃And P ₀₄The corresponding unimpeded state of difference, normal condition, congestion state and blocked state; C is that traffic flow data is all, c _iBe described traffic flow pattern P _0jCorresponding traffic flow data, and i=(1,2 ..., n), n=3, c ₁, c ₂, c ₃Represent average wagon flow speed, roadway occupancy and average stop delay time respectively; V _0jiBe described traffic flow pattern P _0jCorresponding to described traffic flow data c _iClassical territory＜q _0ji, b _0ji;

The joint territory matter-element of described traffic flow data is set up as follows:

Wherein, set P is all of road traffic stream mode; V _Pj1, V _Pj2..., V _PjnBe respectively that described set P is about described traffic flow data c ₁, c ₂..., c _nThe joint territory of P, obviously have after the contrast

(i=1,2 ..., n).

Preferably, in the described matter-element model step of setting up urban road traffic flow pattern to be discriminated, the matter-element model of described urban road traffic flow pattern to be discriminated is set up as follows:

Wherein, R _dThe matter-element to be identified that is called traffic flow pattern, P _dRepresent traffic flow pattern to be identified, V _iRepresent described traffic flow pattern P to be identified _dAbout described influence factor c _DiMeasured value.

Preferably, in the correlation function value of measured value about various traffic flow patterns of described described each influence factor of asking for correlation function value step, i influence factor is as follows about the correlation function value representation of j kind traffic flow pattern:

$K_j\left(v_i\right)=\left\{\begin{array}{c}\frac{-\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})}{\left|V_{0\mathrm{ji}}\right|},v_i\∈V_{0\mathrm{ji}}\\ \frac{\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})}{\mathrm{\ρ}(v_i,V_{\mathrm{pji}})-\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})},v_i\∉V_{0\mathrm{ji}}\end{array}\right.$

Wherein:

$\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})=|v_i-\frac{a_{0\mathrm{ji}}+b_{0\mathrm{ji}}}{2}|-\frac{b_{0\mathrm{ji}}-a_{0\mathrm{ji}}}{2},(i=\mathrm{1,2},...,n)$

$\mathrm{\ρ}(v_i,V_{\mathrm{pji}})=|v_i-\frac{a_{\mathrm{pji}}+b_{\mathrm{pji}}}{2}|-\frac{b_{\mathrm{pji}}-a_{\mathrm{pji}}}{2},(i=\mathrm{1,2},...,n)$

Preferably, in described correlation function value weighted sum step, described each influence factor of asking for is finished as follows about the weighted sum of the correlation function value of every kind of traffic flow pattern:

$K_j\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\λ}}_i{K}_j\left(v_i\right)$

Wherein, λ ₁, λ ₂..., λ _nBe the weight coefficient of each feature, and have:

$\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\λ}}_i=1$

Preferably, in the comparison step,, get K corresponding to the described correlation function value sum in the described correlation function value weighted sum step _J0=maxK _j(p), determine that then p belongs to traffic flow pattern p _0j, and then judge traffic flow pattern p _0jCorresponding traffic behavior.

A kind of distinguishing urban road traffic flow modes in real time system comprises:

The structure module of classical territory matter-element and joint territory matter-element, this module is an object with traffic flow pattern and traffic flow data, sets up the classical territory matter-element and the joint territory matter-element of described traffic flow data according to the Matter Analysis method, and therefrom obtains corresponding classical thresholding and joint thresholding;

The matter-element model of traffic flow pattern to be discriminated makes up module, this module is according to detecting the matter-element model that the traffic flow data of being gathered is set up traffic flow pattern to be discriminated in real time, and obtain the measured value of the influence factor of described traffic flow pattern to be discriminated, wherein, described influence factor is the corresponding traffic flow data of traffic flow pattern to be discriminated;

The correlation function value is asked for module, this module is according to the classical territory matter-element of described traffic flow data and the classical thresholding and the joint thresholding of the structure module output of joint territory matter-element, and make up the influence factor that module provides according to the matter-element model of described traffic flow pattern to be discriminated, ask for the correlation function value of each influence factor about various traffic flow patterns;

Correlation function value weighted sum module: this module is asked for the weighted sum of each influence factor about the correlation function value of every kind of traffic flow pattern according to the measured value of described each influence factor correlation function value about various traffic flow patterns;

Comparison module: about the size between the weighted sum of the correlation function value of every kind of traffic flow pattern, the selection pairing traffic flow pattern of the maximum wherein is as the traffic behavior of current traffic flow pattern by more described each influence factor for this module; And,

The issuing traffic block of state: this module is responsible for issuing the traffic behavior of described current traffic flow pattern, and it is a kind of state in unimpeded state, normal condition, congestion state and the blocked state.

Preferably, in the structure module of the classical territory matter-element of described traffic flow data and joint territory matter-element, described traffic flow pattern comprises: unimpeded state, normal condition, congestion state and four kinds of traffic behaviors of blocked state; Described traffic flow data comprises: average wagon flow speed, roadway occupancy and average stop delay time; The classical territory matter-element of the pairing traffic flow data of described Different Traffic Flows pattern is set up as follows:

Wherein, P _0jBe traffic flow pattern, and j=(1,2 ..., m), m=4; c _iBe described traffic flow pattern P _0jTraffic flow data, i=(1,2 ..., n), n=3; V _0jiBe described traffic flow pattern P _0jCorresponding to described influence factor c _iThe value scope, promptly classical territory＜a _0ji, b _0ji;

The joint territory matter-element of described traffic flow data is set up as follows:

Wherein, P is all of described traffic flow pattern; V _Pj1, V _Pj2..., V _PjnBe respectively that P is about c ₁, c ₂..., c _nThe joint territory, obviously have after the contrast

(i=1,2 ..., n).

Preferably, described correlation function value is asked for classical thresholding and the joint thresholding of obtaining in the structure module of module according to the classical territory matter-element of described traffic flow data and joint territory matter-element, and according to the measured data that makes up the described influence factor that module obtains from the matter-element model of described traffic flow pattern to be discriminated, it is as follows about the correlation function value of j class traffic flow pattern to ask for i influence factor:

$K_j\left(v_i\right)=\left\{\begin{array}{c}\frac{-\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})}{\left|V_{0\mathrm{ji}}\right|},v_i\∈V_{0\mathrm{ji}}\\ \frac{\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})}{\mathrm{\ρ}(v_i,V_{\mathrm{pji}})-\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})},v_i\∉V_{0\mathrm{ji}}\end{array}\right.$

Wherein, v _iDescribed influence factor is described traffic flow data, and has:

$\mathrm{\ρ}(v_i,V_{0\mathrm{ji}})=|v_i-\frac{a_{0\mathrm{ji}}+b_{0\mathrm{ji}}}{2}|-\frac{b_{0\mathrm{ji}}-a_{0\mathrm{ji}}}{2},(i=\mathrm{1,2},...,n)$

$\mathrm{\ρ}(v_i,V_{\mathrm{pji}})=|v_i-\frac{a_{\mathrm{pji}}+b_{\mathrm{pji}}}{2}|-\frac{b_{\mathrm{pji}}-a_{\mathrm{pji}}}{2},(i=\mathrm{1,2},...,n)$

Preferably, it is as follows at the correlation function value sum of every kind of traffic flow pattern that described correlation function value weighting block is asked for all kinds of traffic flow datas that detected:

$K_j\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\λ}}_i{K}_j\left(v_i\right)$

Wherein, λ ₁, λ ₂..., λ _nBe the weight coefficient of each feature, and have:

$\underset{i=1}{\overset{n}{\mathrm{\Σ}}}{\mathrm{\λ}}_i=1$

Described comparison module relatively corresponding to the correlation function value sum of every kind of traffic flow pattern, is got K _J0=max K _j(p), determine that then p belongs to traffic flow pattern p _0j, and then judge traffic flow pattern p _0jCorresponding traffic behavior.

The inventive method is divided into four kinds of traffic behaviors by the traffic flow pattern with urban road, and the traffic flow data that collects is carried out analyzing and processing, and then obtains four kinds of traffic behaviors of urban road traffic flow exactly.By traffic behavior being published on the corresponding VMS, the current transport information of decorrelation road that can make the driver provides useful help for the driver selects suitable path, gathers around resistance thereby urban transportation is reduced, and improves the running efficiency of urban transportation system.

Description of drawings

Following accompanying drawing of the present invention is used to understand the present invention at this as a part of the present invention.Embodiments of the invention and description thereof have been shown in the accompanying drawing, have been used for explaining principle of the present invention.

The workflow diagram of Fig. 1 urban road traffic state method for real time discriminating according to a particular embodiment of the invention;

Fig. 2 is the modular structure synoptic diagram of urban road traffic state real time discriminating system according to a particular embodiment of the invention.

Embodiment

Below in conjunction with the drawings and specific embodiments the present invention is described in further detail:

The inventive method is defined as the traffic flow pattern of urban road: unimpeded state, normal condition, congestion state and four kinds of traffic behaviors of blocked state; The traffic flow data of traffic flow pattern correspondence comprises: average wagon flow speed, roadway occupancy and average stop delay time.Be example with the crossing, Beijing and the highway section that links to each other now, the traffic flow data that this traffic route of a certain sampling instant is gathered is: average wagon flow speed v=2km/h; Roadway occupancy η=0.18%; And, average stop delay time τ=30s.On this basis, the pairing traffic flow data of each traffic flow pattern is sampled, Fig. 1 is the workflow diagram according to urban road traffic state method for real time discriminating of the present invention, as shown in Figure 1, carries out following steps in each traffic flow data sampling instant:

Step 1: set up classical territory matter-element and joint territory matter-element, it is an object with traffic flow pattern and traffic flow data, based on the Matter Analysis method set up the Different Traffic Flows pattern the classical territory matter-element and the joint territory matter-element of corresponding described traffic flow data, and therefrom obtain corresponding classical thresholding and joint thresholding.The classical territory matter-element of the traffic status identification condition of traffic flow pattern can be defined as follows:

$R_{d1}=(P_{d1},C,V_{d1})=\left[\begin{array}{cc}{P}_d,& \stackrel{\&OverBar;}{v},<\mathrm{41,80}>\\ & \mathrm{\&eta;},<\mathrm{0,0.05}>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{0,14}>\end{array}\right]$

$R_{d2}=(P_{d2},C,V_{d2})=\left[\begin{array}{cc}{P}_d,& \stackrel{\&OverBar;}{v},<\mathrm{33,39}>\\ & \mathrm{\&eta;},<\mathrm{0.1,0.15}>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{16,24}>\end{array}\right]$

$R_{d3}=(P_{d3},C,V_{d3})=\left[\begin{array}{cc}{P}_d,& \stackrel{\&OverBar;}{v},<\mathrm{25,31}>\\ & \mathrm{\&eta;},<0.2,0.3>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{26,39}>\end{array}\right]$

$R_{d4}=(P_{d4},C,V_{d4})=\left[\begin{array}{cc}{P}_d,& \stackrel{\&OverBar;}{v},<\mathrm{0,23}>\\ & \mathrm{\&eta;},<0.35,1>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{41,200}>\end{array}\right]$

R _D1, R _D2, R _D3, R _D4The classical matter-element of representing unimpeded state, normal condition, congestion state and these four kinds of traffic flow patterns of blocked state under speed, occupation rate, the three kinds of traffic flow influence factors effect of stop delay time respectively.

The joint territory matter-element of traffic status identification condition can be defined as follows:

$R_{p1}=(P_{p1},C,V_{p1})=\left[\begin{array}{cc}{P}_p,& \stackrel{\&OverBar;}{v},<\mathrm{39,80}>\\ & \mathrm{\&tau;},<\mathrm{0,0.1}>\\ & \stackrel{\&OverBar;}{\mathrm{\&eta;}},<\mathrm{0,16}>\end{array}\right]$

$R_{p2}=(P_{p2},C,V_{p2})=\left[\begin{array}{cc}{P}_p,& \stackrel{\&OverBar;}{v},<\mathrm{31,41}>\\ & \mathrm{\&eta;},<\mathrm{0.05,0.2}>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{14,26}>\end{array}\right]$

$R_{p3}=(P_{d3},C,V_{d3})=\left[\begin{array}{cc}{P}_p,& \stackrel{\&OverBar;}{v},<\mathrm{23,33}>\\ & \mathrm{\&eta;},<0.15,0.35>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{24,41}>\end{array}\right]$

$R_{p4}=(P_{d4},C,V_{d4})=\left[\begin{array}{cc}{P}_p,& \stackrel{\&OverBar;}{v},<\mathrm{0,25}>\\ & \mathrm{\&eta;},<0.3,1>\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},<\mathrm{39,200}>\end{array}\right]$

R _P1, R _P2, R _P3, R _P4The joint territory matter-element of representing unimpeded state under three kinds of influence factors, normal condition, congestion state and these four kinds of traffic flow patterns of blocked state respectively.

The numerical value in wherein classical territory and joint territory is drawn by the following step:

(1) determine average speed of traffic flow, occupation rate, average stop delay time degree of membership to blocked state, congestion state, normal condition and these four kinds of traffic behaviors of unimpeded state:

${\mathrm{\&mu;}}_{\mathrm{ji}}\left(x\right)=\left\{\begin{array}{cc}0& x<a_{\mathrm{ij}}\\ (x-a_{\mathrm{ij}})/(b_{\mathrm{ij}}-a_{\mathrm{ij}})& a_{\mathrm{ij}}\&le;x\&le;b_{\mathrm{ij}}\\ 1& x>b_{\mathrm{ij}}\end{array}\right.$

I represents that traffic behavior is unimpeded state in the following formula, and j represents the traffic flow velocity characteristic, or i represents that traffic behavior is a blocked state, and j represents traffic flow occupation rate or stop delay temporal characteristics.

${\mathrm{\&mu;}}_{\mathrm{ji}}\left(x\right)=\left\{\begin{array}{cc}1& x<a_{\mathrm{ij}}\\ (b_{\mathrm{ij}}-a_{\mathrm{ij}})/(x-a_{\mathrm{ij}})& a_{\mathrm{ij}}\&le;x\&le;b_{\mathrm{ij}}\\ 0& x>b_{\mathrm{ij}}\end{array}\right.$

I represents that traffic behavior is unimpeded state in the following formula, and j represents traffic flow occupation rate or stop delay temporal characteristics, or i represents that traffic behavior is a blocked state, and j represents the traffic flow velocity characteristic.

${\mathrm{\&mu;}}_{\mathrm{ji}}\left(x\right)=\left\{\begin{array}{cc}0& x<a_{\mathrm{ij}}\\ \left({x-a}_{\mathrm{ij}}\right)/(b_{\mathrm{ij}}-a_{\mathrm{ij}})& a_{\mathrm{ij}}\&le;x\&le;b_{\mathrm{ij}}\\ 1& b_{\mathrm{ij}}<x<c_{\mathrm{ij}}\\ (d_{\mathrm{ij}}-x)/(d_{\mathrm{ij}}-c_{\mathrm{ij}})& c_{\mathrm{ij}}\&le;x\&le;d_{\mathrm{ij}}\\ 0& x>d_{\mathrm{ij}}\end{array}\right.$

I represents that traffic behavior is normal or congestion state in the following formula, and j represents traffic flow occupation rate, stop delay time or velocity characteristic.

(2) the threshold value a in above-mentioned each membership function _Ij, b _Ij, c _Ij, d _IjDetermine as follows:

The threshold value of speed, occupation rate and the stop delay of table 1 under various urban road traffic states

Step 2: the matter-element model of setting up traffic flow pattern to be discriminated, it is according to detecting the described traffic flow data of being gathered in real time, set up the matter-element model of traffic flow pattern to be discriminated, and obtain the measured value of the influence factor of described traffic flow pattern to be discriminated, wherein, described influence factor is the corresponding traffic flow data of traffic flow pattern to be discriminated;

If road traffic state matter-element to be identified is as follows:

$R_p=\left[\begin{array}{cc}{P}_p,& \stackrel{\&OverBar;}{v},32\\ & \mathrm{\&eta;},0.18\\ & \stackrel{\&OverBar;}{\mathrm{\&tau;}},30\end{array}\right]$

Adopting the extreme difference method to carry out dimensionless to the value of the influence factor that influences traffic flow pattern handles:

$V_i^{\&prime;}=\frac{V_i-V_{\min }}{V_{\max }-V_{\min }}$

Wherein, V _iBe that i influence factor is at certain (as j) the joint territory of pattern or the raw data in classical territory, V ' _iBe the data after the extreme differenceization, V _Imax, V _IminBe respectively the maximum evaluation criterion value and the minimum evaluation criterion value of i factor.

Step 3: ask for the correlation function value, it is according to the classical thresholding and the joint thresholding of the classical territory matter-element and the joint territory matter-element of described traffic flow data, and, ask for the correlation function value of the measured value of each influence factor about various traffic flow patterns according to the measured value of the described influence factor of described traffic flow pattern to be discriminated.In one embodiment, it is as follows about the correlation function value of j class traffic flow pattern to ask for i influence factor:

$K_j\left(v_i\right)=\left\{\begin{array}{c}\frac{-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\left|V_{0\mathrm{ji}}\right|},v_i\&Element;V_{0\mathrm{ji}}\\ \frac{\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})},v_i\&NotElement;V_{0\mathrm{ji}}\end{array}\right.$

Wherein, v _iDescribed influence factor is described traffic flow data, and has:

$\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})=|v_i-\frac{a_{0\mathrm{ji}}+b_{0\mathrm{ji}}}{2}|-\frac{b_{0\mathrm{ji}}-a_{0\mathrm{ji}}}{2},(i=\mathrm{1,2},...,n)$

$\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})=|v_i-\frac{a_{\mathrm{pji}}+b_{\mathrm{pji}}}{2}|-\frac{b_{\mathrm{pji}}-a_{\mathrm{pji}}}{2},(i=\mathrm{1,2},...,n)$

If v _iBe i influence factor; v ₁, v ₂, v ₃Value be by obtaining after and the normalization corresponding with table 1.Then the degree of association is calculated as follows:

$\mathrm{\&rho;}(v_1,V_{p21})=|v_1-\frac{a_{p21}+b_{p21}}{2}|-\frac{b_{p21}-a_{p21}}{2}=|0.4-\frac{0.39+0.51}{2}|-\frac{0.51-0.39}{2}=-0.01$

$\mathrm{\&rho;}(v_2,V_{p22})=|v_2-\frac{a_{p22}+b_{p22}}{2}|-\frac{b_{p22}-a_{p22}}{2}=|0.18-\frac{0.05+0.2}{2}|-\frac{0.51}{2}=-0.02$

$\mathrm{\&rho;}(v_3,V_{p23})=|v_3-\frac{a_{p23}+b_{p23}}{2}|-\frac{b_{p23}-a_{p23}}{2}=|0.15-\frac{0.07+0.13}{2}|-\frac{0.13-0.07}{2}=0.02$

$\mathrm{\&rho;}(v_1,V_{021})=|v_1-\frac{a_{021}+b_{021}}{2}|-\frac{b_{021}-a_{021}}{2}=|0.4-\frac{0.41+0.49}{2}|-\frac{0.49-0.41}{2}=0.01$

$\mathrm{\&rho;}(v_2,V_{022})=|v_2-\frac{a_{022}+b_{022}}{2}|-\frac{b_{022}-a_{022}}{2}=|0.18-\frac{0.1+0.15}{2}|-\frac{0.15-0.1}{2}=0.03$

$\mathrm{\&rho;}(v_3,V_{023})=|v_2-\frac{a_{023}+b_{023}}{2}|-\frac{b_{023}-a_{023}}{2}=|0.15-\frac{0.08+0.12}{2}|-\frac{0.12-0.08}{2}=0.03$

Then can calculate by the computing formula of asking for correlation function:

$K_2\left(v_1\right)=\frac{0.01}{-0.01-0.01}=-0.5K_2\left(v_2\right)=\frac{0.03}{-0.03-0.03}=-0.6K_2\left(v_3\right)=\frac{0.03}{0.02-0.03}=-3$

In like manner:

$\mathrm{\&rho;}(v_1,V_{p31})=|v_1-\frac{a_{p31}+b_{p31}}{2}|-\frac{b_{p31}-a_{p31}}{2}=|0.4-\frac{0.29+0.41}{2}|-\frac{0.41-0.29}{2}=-0.01$

$\mathrm{\&rho;}(v_2,V_{p32})=|v_2-\frac{a_{p32}+b_{p32}}{2}|-\frac{b_{p32}-a_{p32}}{2}=|0.18-\frac{0.15+0.35}{2}|-\frac{0.2}{2}=-0.03$

$\mathrm{\&rho;}(v_3,V_{p33})=|v_3-\frac{a_{p33}+b_{p33}}{2}|-\frac{b_{p33}-a_{p33}}{2}=|0.15-\frac{0.12+0.21}{2}|-\frac{0.21-0.12}{2}=-0.03$

$\mathrm{\&rho;}(v_1,V_{031})=|v_1-\frac{a_{031}+b_{031}}{2}|-\frac{b_{031}-a_{031}}{2}=|0.4-\frac{0.31+0.39}{2}|-\frac{0.39-0.31}{2}=0.01$

$\mathrm{\&rho;}(v_2,V_{032})=|v_2-\frac{a_{032}+b_{032}}{2}|-\frac{b_{032}-a_{032}}{2}=|0.18-\frac{0.2+0.3}{2}|-\frac{0.3-0.2}{2}=0.02$

$\mathrm{\&rho;}(v_3,V_{033})=|v_3-\frac{a_{033}+b_{033}}{2}|-\frac{b_{033}-a_{033}}{2}=|0.15-\frac{0.13+0.195}{2}|-\frac{0.195-0.13}{2}=-0.02$

And have:

$K_3\left(v_1\right)=\frac{0.01}{-0.01-0.01}=-0.5$ $K_3\left(v_2\right)=\frac{0.02}{-0.03-0.02}=-0.4$ $K_3\left(v_3\right)=\frac{0.02}{0.195-0.13}=0.31$

Step 4: the weighted sum of correlation function value, it asks for the weighted sum of each influence factor about the correlation function value of every kind of traffic flow pattern according to the measured value of described each influence factor correlation function value about various traffic flow patterns.Described each influence factor of asking for can be finished by following formula about the weighted sum of the correlation function value of every kind of traffic flow pattern:

$K_j\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i{K}_j\left(v_i\right)$

Wherein, λ ₁, λ ₂..., λ _nBe the weight coefficient of each feature, and have:

$\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i=1$

The result of calculation of integrating step 4 is utilized the weighted sum formula of step 5

Determine that by expert investigation weight is respectively 0.2,0.3,0.5, then specifically to ask for mode as follows for the weighted sum of the correlation function value of every kind of AC mode:

$K_2\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i{K}_2\left(v_i\right)=-1.78$

$K_3\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i{K}_3\left(v_i\right)=-0.065$

Step 5: comparison step, about the size between the weighted sum of the correlation function value of every kind of traffic flow pattern, the selection pairing traffic flow pattern of the maximum wherein is as the traffic behavior of current traffic flow pattern by more described each influence factor for it.

Because for unimpeded and stop up two states, the numerical value of three kinds of influence factors of matter-element to be identified all is in outside the joint territory, thus above-mentioned traffic state judging only need carry out with crowded two states at normal because K ₃(p)＞K ₂(p), so finally can to differentiate this traffic behavior according to result of calculation be congestion state.

Step 6: the issuing traffic state, it issues the traffic behavior of described current traffic flow pattern, and it is a kind of state in unimpeded state, normal condition, congestion state and the blocked state.

On the center monitoring computer workstation, issue road traffic congestion information by corresponding VMS equipment such as serial ports or network interface.

Embodiment 2, and distinguishing urban road traffic flow modes in real time system as shown in Figure 2 to be example with Beijing's traffic intersection and the highway section that links to each other, is that the distinguishing urban road traffic flow modes in real time system that object is set up comprises with this traffic intersection and the highway section that links to each other still:

The structure module of classical territory matter-element and joint territory matter-element, this module is an object with traffic flow pattern and traffic flow data, sets up the classical territory matter-element and the joint territory matter-element of described traffic flow data according to the Matter Analysis method, and therefrom obtains corresponding classical thresholding and joint thresholding;

The matter-element model of traffic flow pattern to be discriminated makes up module, and this module is set up the matter-element model of traffic flow pattern to be discriminated according to detecting the traffic flow data of being gathered in real time, and obtains the measured value of influence factor;

The correlation function value is asked for module, this module is according to the classical territory matter-element of described traffic flow data and the classical thresholding and the joint thresholding of the structure module output of joint territory matter-element, and make up the influence factor that module provides according to the matter-element model of described traffic flow pattern to be discriminated, ask for the correlation function value of each influence factor about various traffic flow patterns;

Correlation function value weighting block, this module is asked for each influence factor that module asks for correlation function value about various traffic flow patterns according to described correlation function value, asks for all kinds of traffic flow datas that the detected correlation function value sum corresponding to every kind of traffic flow pattern;

Comparison module, this module are by the size between the correlation function value sum of the various traffic flow patterns of more described correlation function value weighting block output, and the selection pairing traffic flow pattern of the maximum wherein is as the residing traffic behavior of current traffic flow; And,

Traffic behavior release module, this module are responsible for issuing the traffic behavior of the current traffic flow pattern of described comparison module output: unimpeded state, normal condition, congestion state and blocked state a kind of state wherein.

This distinguishing urban road traffic flow modes in real time system further comprises following technical characterictic:

In the structure module of the classical territory matter-element of described traffic flow data and joint territory matter-element, described traffic flow pattern comprises: unimpeded state, normal condition, congestion state and four kinds of traffic behaviors of blocked state; Described traffic flow data comprises: average wagon flow speed, roadway occupancy and average stop delay time; The classical territory matter-element of the pairing traffic flow data of described Different Traffic Flows pattern is set up as follows:

Wherein, P _0jFor traffic flow pattern (j=1,2 ..., m), m=4; c _iBe described traffic flow pattern P _0jFeature, promptly the influence factor of described traffic flow pattern (i=1,2 ..., n), n=3; V _0jiBe described traffic flow pattern P _0jCorresponding to described influence factor c _iThe value scope, promptly classical territory＜a _0ji, b _0ji;

The joint territory matter-element of described traffic flow data is set up as follows:

Wherein, P is all of road traffic stream mode; V _P1, V _P2..., V _PnBe respectively that P is about c ₁, c ₂..., c _nSpan, the joint territory of promptly gathering P obviously has after the contrast

(i=1,2 ..., n).

Described correlation function value is asked for classical thresholding and the joint thresholding of obtaining in the structure module of module according to the classical territory matter-element of described traffic flow data and joint territory matter-element, and according to the measured data that makes up the described influence factor that module obtains from the matter-element model of traffic flow pattern to be discriminated, it is as follows about the correlation function value of j class traffic flow pattern to ask for i influence factor:

$K_j\left(v_i\right)=\left\{\begin{array}{c}\frac{-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\left|V_{0\mathrm{ji}}\right|},v_i\&Element;V_{0\mathrm{ji}}\\ \frac{\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})},v_i\&NotElement;V_{0\mathrm{ji}}\end{array}\right.$

Wherein:

$\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})=|v_i-\frac{a_{0\mathrm{ji}}+b_{0\mathrm{ji}}}{2}|-\frac{b_{0\mathrm{ji}}-a_{0\mathrm{ji}}}{2},(i=\mathrm{1,2},...,n)$

$\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})=|v_i-\frac{a_{\mathrm{pji}}+b_{\mathrm{pji}}}{2}|-\frac{b_{\mathrm{pji}}-a_{\mathrm{pji}}}{2},(i=\mathrm{1,2},...,n)$

It is as follows at the correlation function value sum of every kind of traffic flow pattern that described correlation function value weighting block is asked for all kinds of traffic flow datas that detected:

$K_j\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i{K}_j\left(v_i\right)$

Wherein, λ ₁, λ ₂..., λ _nBe the weight coefficient of each feature, and have:

$\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i=1$

Described comparison module relatively corresponding to the correlation function value sum of every kind of traffic flow pattern, is got K _J0=max K _j(p), determine that then p belongs to traffic flow pattern p _0j, and then judge traffic flow pattern p _0jCorresponding traffic behavior.

The present invention is illustrated by the foregoing description, but should be understood that, the foregoing description just is used for for example and illustrative purposes, but not is intended to the present invention is limited in the described scope of embodiments.It will be appreciated by persons skilled in the art that in addition the present invention is not limited to the foregoing description, can also make more kinds of variants and modifications according to instruction of the present invention, these variants and modifications all drop in the present invention's scope required for protection.Protection scope of the present invention is defined by the appended claims and equivalent scope thereof.

Claims (10)

1. distinguishing urban road traffic flow modes in real time method is characterized in that: each traffic flow data is sampled, and carry out following steps in each traffic flow data sampling instant:

Set up the step of classical territory matter-element and joint territory matter-element: with traffic flow pattern and traffic flow data is object, based on the Matter Analysis method set up the Different Traffic Flows pattern the classical territory matter-element and the joint territory matter-element of corresponding described traffic flow data, and therefrom obtain corresponding classical thresholding and joint thresholding;

Set up the matter-element model step of traffic flow pattern to be discriminated: the described traffic flow data of gathering according to real-time detection, set up the matter-element model of traffic flow pattern to be discriminated, and obtain the measured value of the influence factor of described traffic flow pattern to be discriminated, wherein, described influence factor is the corresponding traffic flow data of traffic flow pattern to be discriminated;

Ask for correlation function value step: according to classical territory matter-element and the classical thresholding that saves the territory matter-element and the joint thresholding of described traffic flow data, and, ask for the correlation function value of the measured value of each influence factor about various traffic flow patterns according to the measured value of the described influence factor of described traffic flow pattern to be discriminated;

Correlation function value weighted sum step:, ask for the weighted sum of each influence factor about the correlation function value of every kind of traffic flow pattern according to the measured value of described each influence factor correlation function value about various traffic flow patterns;

Comparison step: about the size between the weighted sum of the correlation function value of every kind of traffic flow pattern, the selection pairing traffic flow pattern of the maximum wherein is as the traffic behavior of current traffic flow pattern by more described each influence factor; And,

The issuing traffic state step: issue the traffic behavior of described current traffic flow pattern, it is a kind of state in unimpeded state, normal condition, congestion state and the blocked state.

2. distinguishing urban road traffic flow modes in real time method according to claim 1, it is characterized in that, the place is stated in the step of setting up classical territory matter-element and joint territory matter-element, described traffic flow pattern comprises: unimpeded state, normal condition, congestion state and four kinds of traffic behaviors of blocked state, and the classical territory matter-element of the pairing traffic flow data of various traffic flow patterns is set up as follows:

Wherein, P _0jBe traffic flow pattern, and j=(1,2 ..., m), m=4, P ₀₁, P ₀₂, P ₀₃And P ₀₄The corresponding unimpeded state of difference, normal condition, congestion state and blocked state; C is that traffic flow data is all, c _iBe described traffic flow pattern P _0jCorresponding traffic flow data, and i=(1,2 ..., n), n=3, c ₁, c ₂, c ₃Represent average wagon flow speed, roadway occupancy and average stop delay time respectively; V _0jiBe described traffic flow pattern P _0jCorresponding to described traffic flow data c _iClassical territory＜a _0ji, b _0ji;

The joint territory matter-element of described traffic flow data is set up as follows:

Wherein, set P is all of road traffic stream mode; V _Pj1, V _Pj2..., V _PjnBe respectively that described set P is about described traffic flow data c ₁, c ₂..., c _nThe joint territory of P, obviously have after the contrast

3. distinguishing urban road traffic flow modes in real time method according to claim 1, it is characterized in that, in the described matter-element model step of setting up urban road traffic flow pattern to be discriminated, the matter-element model of described urban road traffic flow pattern to be discriminated is set up as follows:

Wherein, R _dThe matter-element to be identified that is called traffic flow pattern, P _dRepresent traffic flow pattern to be identified, V _iRepresent described traffic flow pattern P to be identified _dAbout described influence factor c _DiMeasured value.

4. distinguishing urban road traffic flow modes in real time method according to claim 1, it is characterized in that, in the correlation function value of measured value about various traffic flow patterns of described described each influence factor of asking for correlation function value step, i influence factor is as follows about the correlation function value representation of j kind traffic flow pattern:

$K_j\left(v_i\right)=\left\{\begin{array}{cc}\frac{-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\left|V_{0\mathrm{ji}}\right|},& v_i\&Element;V_{0\mathrm{ji}}\\ \frac{\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})},& v_i\&NotElement;V_{0\mathrm{ji}}\end{array}\right.$

Wherein:

$\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})=|v_i-\frac{a_{0\mathrm{ji}}+b_{0\mathrm{ji}}}{2}|-\frac{b_{0\mathrm{ji}}-a_{0\mathrm{ji}}}{2}(i=\mathrm{1,2},...,n)$

$\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})=|v_i-\frac{a_{\mathrm{pji}}+b_{\mathrm{pji}}}{2}|-\frac{b_{\mathrm{pji}}-a_{\mathrm{pji}}}{2}(i=\mathrm{1,2},...,n).$

5. according to claim 1 or 4 described distinguishing urban road traffic flow modes in real time methods, it is characterized in that, in described correlation function value weighted sum step, described each influence factor of asking for is finished as follows about the weighted sum of the correlation function value of every kind of traffic flow pattern:

$K_j\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i{K}_j\left(v_i\right)$

Wherein, λ ₁, λ ₂..., λ _nBe the weight coefficient of each feature, and have:

$\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i=1.$

6. according to claim 1 or 4 described distinguishing urban road traffic flow modes in real time methods, it is characterized in that: in the described comparison step,, get K corresponding to the described correlation function value sum of described correlation function value weighted sum step _J0=max K _j(p), determine that then p belongs to traffic flow pattern p _0j, and then judge traffic flow pattern p _0jCorresponding traffic behavior.

7. a distinguishing urban road traffic flow modes in real time system is characterized in that, comprising:

The structure module of classical territory matter-element and joint territory matter-element, this module is an object with traffic flow pattern and traffic flow data, sets up the classical territory matter-element and the joint territory matter-element of described traffic flow data according to the Matter Analysis method, and therefrom obtains corresponding classical thresholding and joint thresholding;

The matter-element model of traffic flow pattern to be discriminated makes up module, this module is according to detecting the matter-element model that the traffic flow data of being gathered is set up traffic flow pattern to be discriminated in real time, and obtain the measured value of the influence factor of described traffic flow pattern to be discriminated, wherein, described influence factor is the corresponding traffic flow data of traffic flow pattern to be discriminated;

The correlation function value is asked for module, this module is according to the classical territory matter-element of described traffic flow data and the classical thresholding and the joint thresholding of the structure module output of joint territory matter-element, and make up the influence factor that module provides according to the matter-element model of described traffic flow pattern to be discriminated, ask for the correlation function value of each influence factor about various traffic flow patterns;

Correlation function value weighted sum module: this module is asked for the weighted sum of each influence factor about the correlation function value of every kind of traffic flow pattern according to the measured value of described each influence factor correlation function value about various traffic flow patterns;

Comparison module: about the size between the weighted sum of the correlation function value of every kind of traffic flow pattern, the selection pairing traffic flow pattern of the maximum wherein is as the traffic behavior of current traffic flow pattern by more described each influence factor for this module; And,

The issuing traffic block of state: this module is responsible for issuing the traffic behavior of described current traffic flow pattern, and it is a kind of state in unimpeded state, normal condition, congestion state and the blocked state.

8. distinguishing urban road traffic flow modes in real time according to claim 7 system, it is characterized in that, in the structure module of the classical territory matter-element of described traffic flow data and joint territory matter-element, described traffic flow pattern comprises: unimpeded state, normal condition, congestion state and four kinds of traffic behaviors of blocked state; Described traffic flow data comprises: average wagon flow speed, roadway occupancy and average stop delay time; The classical territory matter-element of the pairing traffic flow data of described Different Traffic Flows pattern is set up as follows:

Wherein, P _0jBe traffic flow pattern, and j=(1,2 ..., m), m=4; c _iBe described traffic flow pattern P _0jTraffic flow data, i=(1,2 ..., n), n=3; V _0jiBe described traffic flow pattern P _0jCorresponding to described influence factor c _iThe value scope, promptly classical territory＜a _0ji, b _0ji;

The joint territory matter-element of described traffic flow data is set up as follows:

Wherein, P is all of described traffic flow pattern; V _Pj1, V _Pj2..., V _PjnBe respectively that P is about c ₁, c ₂..., c _nThe joint territory, obviously have after the contrast

9. according to claim 7 or 8 described distinguishing urban road traffic flow modes in real time systems, it is characterized in that, described correlation function value is asked for classical thresholding and the joint thresholding of obtaining in the structure module of module according to the classical territory matter-element of described traffic flow data and joint territory matter-element, and according to the measured data that makes up the described influence factor that module obtains from the matter-element model of described traffic flow pattern to be discriminated, it is as follows about the correlation function value of j class traffic flow pattern to ask for i influence factor:

$K_j\left(v_i\right)=\left\{\begin{array}{cc}\frac{-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\left|V_{0\mathrm{ji}}\right|},& v_i\&Element;V_{0\mathrm{ji}}\\ \frac{\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})}{\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})-\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})},& v_i\&NotElement;V_{0\mathrm{ji}}\end{array}\right.$

Wherein, v _iDescribed influence factor is described traffic flow data, and has:

$\mathrm{\&rho;}(v_i,V_{0\mathrm{ji}})=|v_i-\frac{a_{0\mathrm{ji}}+b_{0\mathrm{ji}}}{2}|-\frac{b_{0\mathrm{ji}}-a_{0\mathrm{ji}}}{2}(i=\mathrm{1,2},...,n)$

$\mathrm{\&rho;}(v_i,V_{\mathrm{pji}})=|v_i-\frac{a_{\mathrm{pji}}+b_{\mathrm{pji}}}{2}|-\frac{b_{\mathrm{pji}}-a_{\mathrm{pji}}}{2}(i=\mathrm{1,2},...,n).$

10. according to claim 7,8 or 9 described distinguishing urban road traffic flow modes in real time systems, it is characterized in that it is as follows at the correlation function value sum of every kind of traffic flow pattern that described correlation function value weighting block is asked for all kinds of traffic flow datas that detected:

$K_j\left(p\right)=\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i{K}_j\left(v_i\right)$

Wherein, λ ₁, λ ₂..., λ _nBe the weight coefficient of each feature, and have:

$\underset{i=1}{\overset{n}{\mathrm{\&Sigma;}}}{\mathrm{\&lambda;}}_i=1$

Described comparison module relatively corresponding to the correlation function value sum of every kind of traffic flow pattern, is got K _J0=max K _j(p), determine that then p belongs to traffic flow pattern p _0j, and then judge traffic flow pattern p _0jCorresponding traffic behavior.

Images