CN104157139B - A kind of traffic congestion Forecasting Methodology and method for visualizing - Google Patents

A kind of traffic congestion Forecasting Methodology and method for visualizing Download PDF

Info

Publication number
CN104157139B
CN104157139B CN201410381904.3A CN201410381904A CN104157139B CN 104157139 B CN104157139 B CN 104157139B CN 201410381904 A CN201410381904 A CN 201410381904A CN 104157139 B CN104157139 B CN 104157139B
Authority
CN
China
Prior art keywords
traffic
traffic congestion
traffic behavior
behavior
congestion
Prior art date
Application number
CN201410381904.3A
Other languages
Chinese (zh)
Other versions
CN104157139A (en
Inventor
何兆成
叶伟佳
沙志仁
Original Assignee
中山大学
广东方纬科技有限公司
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 中山大学, 广东方纬科技有限公司 filed Critical 中山大学
Priority to CN201410381904.3A priority Critical patent/CN104157139B/en
Publication of CN104157139A publication Critical patent/CN104157139A/en
Application granted granted Critical
Publication of CN104157139B publication Critical patent/CN104157139B/en

Links

Abstract

The invention discloses a kind of traffic congestion Forecasting Methodology and method for visualizing, the gps data that the present invention utilizes taxi to return, by map-matching method, gps data is associated with the road on electronic chart; Data according to coupling calculate section speed, utilize the traffic behavior in velocity estimated section, section; Utilize historical data to extract the Evolution of traffic congestion, comprise formation and the dissipation of traffic congestion; Association Real-time Traffic Information storehouse, utilizes sliding time window mechanism to carry out traffic congestion prediction; Based on the intensity of blocking up calculating congested link that predicts the outcome, finally by block up intensity and the coverage of thermal map performance congested link.Forecasting Methodology of the present invention can realize the prediction of blocking up of high-accuracy, and thermodynamic chart method for visualizing makes traffic congestion visual and understandable, is convenient to understand traffic congestion position and coverage.

Description

A kind of traffic congestion Forecasting Methodology and method for visualizing

Technical field

The present invention relates to field of traffic, more specifically, relate to a kind of traffic congestion Forecasting Methodology and method for visualizing.

Background technology

Traffic congestion is the key factor affecting people's go off daily quality and travel cost.Traffic congestion prediction is the core content of traffic congestion management review, and accurately real-time traffic congestion prediction can relieving traffic jam, improves the traffic capacity and the section speed of road network.

Chinese scholars has done large quantity research to traffic congestion prediction algorithm.Some of them scholar chooses vehicle flowrate, occupation rate, flow speeds and vehicle density as parameter, utilizes the learning algorithm based on bayes rule, calculates the possibility that traffic congestion occurs.Also there is scholar to consider the parameter such as flow, speed, occupation rate, average track flow, use maximum entropy model training to obtain the weight of each parameter, thus prediction traffic congestion.Above-mentioned algorithm all considers multiple parameter, obtains the parameter combinations corresponding to traffic congestion by training study.But the input parameter that said method needs when predicted congestion is too much, there is significant limitation in actual applications, except section speed can utilize floating car data and obtains, other parameters are mainly derived from coil checker, there is the shortcomings such as installation cost is large, installation and maintenance are difficult.

As following several Forecasting Methodology: 1) research uses multiparameter method for estimating state more more reasonable than single parameter state method of estimation, more tallies with the actual situation.Consider the magnitude of traffic flow, average velocity, density 3 parameters, by clustering method, three-dimensional parameter is converted to One-dimension Time Series, adopt BP neural network to carry out the prediction of short time traffic conditions, but cluster result is easily subject to the impact of raw data.2) based on the traffic flow modes fuzzy reasoning method of neural network.But the accuracy rate of the method can decline with predicting the increase of step number continuously.3) consider that traffic behavior itself has fuzzy uncertainty, propose based on Adaptive Fuzzy Neural-network inference system.The traffic behavior of the method is divided into 4 classes, and employing fuzzy C-means clustering obtains the cluster centre corresponding to often kind of traffic behavior, then calculates degree of membership, and the maximum corresponding traffic behavior of degree of membership is and predicts the outcome.But fuzzy C-means clustering is easily absorbed in local optimum, cannot ensure that cluster centre can reflect the feature of traffic behavior well.

Conveniently traffic congestion situation is observed, have scholar to propose to express traffic congestion situation by modes such as indexes.Chinese scholars major part adopts the mode of service level class or 0-10 exponentiate to represent, U.S. 1985 HCM (HCM) is given chapter and verse the volume of traffic the earliest, and Assessment of Serviceability of Roads is divided into A-F6 grade by the sensation of road user; The span that some scholars then define traffic congestion index is 0-10, and the larger expression of value is blocked up more serious.In order to presentation (or road) is blocked up the situation of change of intensity, at present some scholars adopt sequential chart to represent regional area blocks up the Changing Pattern of intensity.But sequential chart can only reflect the temporal behavior of a certain region, a certain congestion in road intensity, this region (this road) impact on adjacent area (adjacent road) can not be reflected.

Summary of the invention

In order to overcome the deficiencies in the prior art, first the present invention proposes a kind of traffic congestion Forecasting Methodology, and the parameter of this method input only has section speed, and section speed obtains by floating car data.Compared to the traffic information collection such as coil checker, video means, floating car technology has that data volume is large, broad covered area, low cost and other advantages.

Another object of the present invention proposes a kind of traffic congestion method for visualizing, relative timing figure can not reflecting regional (road) on the impact of periphery road network, method for visualizing proposed by the invention is the distribution adopting thermodynamic chart to express traffic congestion, make traffic congestion visual and understandable, be convenient to understand traffic congestion position and coverage.

To achieve these goals, technical scheme of the present invention is:

A kind of traffic congestion Forecasting Methodology, comprising:

Adopt Floating Car to obtain gps data, by map-matching method, gps data is associated with the road on electronic chart; According to the data estimation section speed of coupling, utilize the traffic behavior in velocity estimated section, section, described traffic behavior comprises and to block up and unimpeded;

Historical traffic status information is utilized to build traffic congestion model; Association real-time traffic states information bank carries out traffic congestion prediction;

Traffic congestion model comprises traffic congestion Model I, traffic congestion modelⅱ and traffic congestion model III;

Suppose interim one day of this week traffic behavior and on it cycle traffic behavior on the same day similar or identical,

Building traffic congestion Model I concrete mode is: trained as training sample by the traffic behavior in n cycle, add up each cycle on the same day t there is the probability that blocks up;

Target road section each cycle on the same day t there is the probability that blocks up:

P i t = Σ j = 1 n σ i j t n - - - ( 1 )

&sigma; i j t = 0 v i j t &GreaterEqual; v T 1 v i j t < v T - - - ( 2 )

Wherein: P i trefer to that the probability blocked up appears in i-th section t; refer to the section speed of i-th section in a jth cycle t; v tit is congestion status discrimination threshold; refer to the traffic behavior of i-th section in a jth cycle t, 0 represents smooth and easy, and 1 expression is blocked up;

The probabilistic determination traffic behavior blocked up is there is according to above-mentioned target road section;

Building the concrete mode of traffic congestion modelⅱ is: in training sample, statistics target road section is when developing the traffic behavior of t from the traffic behavior in t-1 moment, adjacent segments is in the first traffic behavior regularity of distribution in t-1 moment, and described adjacent segments refers to the section adjacent with target road section;

Association real-time traffic states information bank carries out traffic congestion prediction; Be obtain target road section according to the first traffic behavior regularity of distribution to block up probability, then occur the traffic behavior of the probabilistic determination target road section of blocking up according to this target road section;

Building the concrete mode of traffic congestion model III is:

In training sample, statistics adjacent segments is when developing the traffic behavior of t from the traffic behavior in t-1 moment, and the section, upstream of this adjacent segments is in the second traffic behavior regularity of distribution in t-1 moment, and section, described upstream refers to the section of the upstream of adjacent segments;

In training sample, statistics adjacent segments is when developing the traffic behavior of t from the traffic behavior in t-1 moment, and the downstream road section of this adjacent segments is in the 3rd traffic behavior regularity of distribution in t-1 moment, and described downstream road section refers to the section in the downstream of adjacent segments;

In training sample, when statistics target road section develops the traffic behavior of t from the traffic behavior in t-1 moment, adjacent segments is in the 4th traffic behavior regularity of distribution of t;

Association real-time traffic states information bank carries out traffic congestion prediction; Obtain t adjacent segments according to the second traffic behavior regularity of distribution to block up probability, then the traffic behavior of this adjacent segments of probabilistic determination that blocks up according to this adjacent segments; Simultaneously obtain t adjacent segments according to the 3rd traffic behavior regularity of distribution to block up probability, again according to the traffic behavior of this adjacent segments of probabilistic determination that blocks up of this adjacent segments, finally, according to the traffic behavior of the t adjacent segments predicted by second and third traffic behavior regularity of distribution, and the 4th the traffic behavior regularity of distribution obtain t target congested link probability, then the traffic behavior of this target road section of probabilistic determination of blocking up according to this target road section;

When predicting traffic congestion, adopting above-mentioned traffic congestion Model I, traffic congestion modelⅱ and traffic congestion model III three kinds of models couplings, adopting during three kinds of models couplings and predicting with majority principle.

A kind of traffic congestion prediction method for visualizing, is the thermodynamic chart drawing traffic congestion intensity, before drafting thermodynamic chart, needs to determine hotspot location, focus coverage, traffic congestion strength retrogression rule and scheme of colour;

Calculate the intensity C that blocks up of traffic i;

Hotspot location, the position of focus is that the terminus in each section respectively arranges a focus, and from the starting point of section, often distance M rice arranges a focus along section distribution;

Focus coverage, the coverage of focus is the circle of radius M/2 rice, and in units of section, the traffic congestion intensity in this section is evenly distributed on all focuses in this section; If the traffic congestion intensity in section is less than the focus number in this section, be then assigned randomly to C ion individual focus;

Traffic congestion strength retrogression rule, for each focus, take the center of circle as starting point, traffic congestion intensity along with from distance of center circle from increase and linear attenuation;

Scheme of colour, is divided into N number of grade by traffic congestion intensity, and adjacent rank adopts different colours mark.

Compared with prior art, beneficial effect of the present invention is:

Traffic congestion Forecasting Methodology of the present invention, only need to obtain section data by Floating Car, and using section data as inputting unique input parameter, the prediction of blocking up of high-accuracy can be realized, with the traffic information collection such as coil checker, video means, floating car technology has that data volume is large, broad covered area, low cost and other advantages.

The traffic congestion method for visualizing that the present invention proposes is showed by thermodynamic chart, adopts thermodynamic chart to express the distribution of traffic congestion, makes traffic congestion visual and understandable, be convenient to understand traffic congestion position and coverage.

Accompanying drawing explanation

Fig. 1 is traffic congestion Forecasting Methodology implementation framework figure of the present invention.

Fig. 2 is for building candidate road section collection schematic diagram when realizing map match.

Fig. 3 be target road section with adjacent segments associate schematic diagram.

Fig. 4 be section, upstream with adjacent segments and target road section associate schematic diagram.

Fig. 5 be downstream road section with adjacent segments and target road section associate schematic diagram.

Fig. 6 adopts traffic congestion Forecasting Methodology of the present invention to carry out the process flow diagram predicted.

Fig. 7 is the focus coverage schematic diagram of traffic congestion of the present invention prediction method for visualizing.

Fig. 8 is traffic congestion strength retrogression rule schematic diagram.

Fig. 9 is 8:00 Guangzhou traffic congestion thermodynamic chart on April 2

Figure 10 is 8:05 Guangzhou traffic congestion thermodynamic chart on April 2.

Embodiment

Below in conjunction with accompanying drawing, the present invention will be further described, but embodiments of the present invention are not limited to this.

As Fig. 1, the gps data that the present invention utilizes taxi to return, is associated gps data with the road on electronic chart by map-matching method; Data according to coupling calculate section speed, utilize the traffic behavior in velocity estimated section, section; Utilize historical data to extract the Evolution of traffic congestion, comprise formation and the dissipation of traffic congestion; Association Real-time Traffic Information storehouse, utilizes sliding time window mechanism to carry out traffic congestion prediction; Based on the intensity of blocking up calculating congested link that predicts the outcome, finally by block up intensity and the coverage of thermal map performance congested link.

map match

The object of map match determines the position in section, vehicle place.Map-matching method selects matching process (point is to lines matching and point sequence coupling) adaptively according to road mileage in the present embodiment, considers the factors such as distance, course, topology connectivity simultaneously, realize global path coupling.

This matching process comprises two steps, first obtains candidate road section collection, screening possible coupling section, also referred to as candidate road section, seeing Fig. 2 by building an error band around anchor point.Its mid point is determined to mate section to the distance of each candidate road section and the course of GPS point to lines matching according to GPS point; Point sequence coupling, using topology connectivity as the rational constraint condition of coupling, chooses the maximum section combination of coupling weight summation as final matching sequence.

Map matching result stores in a database, as shown in table 1, tables of data comprises the fields such as the number-plate number, GPS positioning time, longitude, latitude, speed, direction, coupling road section ID, vehicle-state (wherein " 4 " represent unloaded, and " 5 " represent carrying).

Table 1 map matching result tables of data

velocity estimation

Velocity estimation utilizes map matching result to calculate section speed.Velocity estimation algorithm divides the distribution situation of single Floating Car GPS point in section, according to the adaptively selected velocity estimation model of distribution situation in the present embodiment.

This algorithm adopts four kinds of velocity estimation models, respectively: distance verses time model, Velocity-time trapezoidal integration model, vehicle tracking model and ground spot speed harmonic average model.

Distance verses time model chooses first and last GPS point of target road section, utilizes the travel speed of range difference divided by mistiming estimating vehicle of 2 GPS points.

The distance that Velocity-time trapezoidal integration model utilizes instantaneous velocity sequence scores accumulated to obtain single Floating Car to pass by, divided by the travel speed of a mistiming pro form bill Floating Car of GPS point sequence first and last point.

There is crossing scope in two ends, vehicle tracking model hypothesis section, Floating Car completes even deceleration, parking and even boost phase within the scope of crossing, and Floating Car at the uniform velocity travels outward in crossing scope.Wherein the ground spot speed at crossing scope and edge, crossing obtains by history gps data.According to above hypothesis, obtain the travel speed of vehicle in conjunction with real gps data.

Ground spot speed harmonic average model is using the harmonic-mean of the instantaneous velocity of all vehicles of target road section as section speed.

Velocity estimation result stores in a database, and as shown in table 2, tables of data comprises the fields such as road section ID, start time, finish time, speed, vehicle number.

Table 2 speed calculation result data table

traffic state judging (threshold speed)

Definition average stroke speed of a motor vehicle in Beijing's is the heavy congestion section on through street lower than the section of 30 kilometers/hour, judges that trunk roads, secondary distributor road and branch road are that the threshold speed of congested link is respectively: 20 kilometers/hour, 15 kilometers/hour, 15 kilometers/hour.Present embodiment adopts above-mentioned speed as the threshold speed of traffic state judging.

Table 3 traffic behavior velocity range table

extract traffic behavior Evolution

Traffic congestion refers under a certain specific space-time condition, due to the traffic trapping phenomena that imbalance between supply and demand produces, comprises category of roads, number of track-lines and the transport need etc. in certain moment.Category of roads and number of track-lines are less likely to occur to change, and transport need has regularity.Therefore in present embodiment, each cycle is a week, suppose some day traffic circulation state and its last week traffic circulation state on the same day same or similar, statistics weekly on the same day t there is the probability that blocks up.

Traffic congestion Model I, rule 1: the probability that blocks up appears in t on the same day weekly.

P i t = &Sigma; j = 1 n &sigma; i j t n - - - ( 1 )

&sigma; i j t = 0 v i j t &GreaterEqual; v T 1 v i j t < v T - - - ( 2 )

Wherein: P i trefer to that the probability blocked up appears in i-th section t; refer to the section speed of i-th section in jth week t; refer to the traffic behavior of i-th section in jth week t, 0 represents smooth and easy, and 1 represents and blocks up; N refers to all numbers of training sample.

Table 4 rule 1 storage organization

The generation of traffic congestion is caused by multifactor, except road self attributes, also relevant to the traffic behavior of adjacent segments, as shown in Figure 3.Therefore the traffic behavior of adjacent segments must be considered when predicting traffic congestion.

The traffic behavior regularity of distribution of rule 2:t-1 moment adjacent segments

The traffic behavior regularity of distribution in rule 2 is added up in target road section traffic behavior evolution process (traffic behavior in t-1 moment is to the traffic behavior of t) adjacent segments t-1 moment.

The traffic behavior of t-1 moment target road section can be divided into 4 kinds of situations to the traffic behavior of t target road section, is respectively:

Table 5 target road section traffic behavior

Add up the traffic behavior of t-1 moment adjacent segments in often kind of situation, statistics is as shown in table 6

Table 6 rule 2 storage organization

Consider that the traffic behavior of target road section is not only relevant to adjacent segments, also may be subject to the impact in the section, upstream (or downstream road section) of adjacent segments, the impact in section, adjacent segments upstream (or downstream road section) during target of prediction road section traffic volume state, should be considered.As shown in Figure 4, Figure 5, dotted lines represents the section blocked up, and dotted line represents the coverage that congested link is possible.The definition in section, upstream is except adjacent segments and target road section, is positioned at the section of adjacent segments upstream; The definition of downstream road section is except adjacent segments and target road section, is positioned at the section in adjacent segments downstream.

Rule 3, rule 4 and rule 5 are it is considered that the section, upstream (or downstream road section) of adjacent segments is on the impact of target road section.Suppose that the impact of section, upstream (or downstream road section) on adjacent segments and target road section can be embodied within a time interval (i.e. Δ t), as blocking up appears in t-1 moment downstream road section, the traffic behavior of t adjacent segments and target road section becomes thereupon and blocks up.Section, upstream (or downstream road section) first affects adjacent segments, then target road section is affected, this process comprises two rules: the traffic behavior regularity of distribution of section, (1) t-1 moment upstream (or downstream road section), i.e. rule 4, rule 5, for judging the traffic behavior of t adjacent segments; (2) the traffic behavior regularity of distribution of t adjacent segments, i.e. rule 3, for judging the traffic behavior of t target road section.

The traffic behavior regularity of distribution of rule 3:t moment adjacent segments

The traffic behavior regularity of distribution of adjacent segments t that rule 3 is added up in target road section traffic behavior evolution process (traffic behavior in t-1 moment is to the traffic behavior of t).

Equally with rule 2 divide 4 kinds of situations, add up the traffic behavior of t adjacent segments in often kind of situation, statistics is as shown in table 7.

Table 7 rule 3 storage organization

The traffic behavior regularity of distribution in section, rule 4:t-1 moment upstream

The traffic behavior regularity of distribution of section, upstream t that rule 4 is added up in adjacent segments traffic behavior evolution process (traffic behavior in t-1 moment is to the traffic behavior of t).

Equally with rule 2 divide 4 kinds of situations, add up the traffic behavior in section, t-1 moment upstream in often kind of situation, statistics is as shown in table 8.

Table 8 rule 4 storage organization

The traffic behavior regularity of distribution of rule 5:t-1 moment downstream road section

The traffic behavior regularity of distribution of downstream road section t that rule 5 is added up in adjacent segments traffic behavior evolution process (traffic behavior in t-1 moment is to the traffic behavior of t).

Equally with rule 2 divide 4 kinds of situations, add up the traffic behavior of t-1 moment downstream road section in often kind of situation, statistics is as shown in table 9.

Table 9 rule 5 storage organization

traffic congestion Forecasting Methodology

Traffic congestion Forecasting Methodology process flow diagram as shown in Figure 6.This Forecasting Methodology is made up of three methods, and each method obtains the traffic status prediction value of target road section, adopts the traffic behavior that obtained by most method as the traffic behavior of target road section.

1) method 1

Step 1-1: laws of use 1, searches the probability P that blocking up appears in t i t.

Step 1-2: if P i tbe more than or equal to 50%, then judge traffic behavior TS ifor blocking up, otherwise be smooth and easy.

2) method 2

Step 2-1: the traffic behavior obtaining the adjacent segments t-1 moment.

Step 2-2: the traffic behavior obtaining the target road section t-1 moment.

Step 2-3: the probability P of blocking up calculating target road section i t, laws of use 2 calculates and to block up probability here, and as shown in table 10, the traffic behavior in hypothetical target section t-1 moment is smooth and easy, the traffic behavior in adjacent segments t-1 moment be respectively block up, smooth and easy, block up, block up, smooth and easy, smooth and easy.

Table 10 calculate target road section block up probability time rule (rule 2)

Target road section is blocked up probability P i tfor:

P i t = 983 564 + 983 = 63.5 %

The probability that blocks up is greater than 50%, and the traffic behavior of prediction t is for blocking up.

3) method 3

Step 3-1: the traffic behavior obtaining the adjacent segments t-1 moment.

Step 3-2: laws of use 4 and rule 5 predict the traffic behavior of adjacent segments t.As shown in table 11, suppose that the traffic behavior in adjacent segments t-1 moment is smooth and easy, the traffic behavior in section, upstream t-1 moment is respectively smooth and easy, blocks up, smooth and easy.

Table 11 calculate adjacent segments block up probability time rule (rule 4,5)

Adjacent segments blocks up probability P i tfor:

P i t = 290 1318 + 290 = 18 %

The probability that blocks up is less than 50%, and the traffic behavior of prediction t is smooth and easy.

Step 3-3: the traffic behavior obtaining the target road section t-1 moment.

Step 3-4: the probability P of blocking up calculating target road section i t, laws of use 3 calculates the probability that blocks up here, and as shown in table 12, the traffic behavior in hypothetical target section t-1 moment is smooth and easy, and the traffic behavior of adjacent segments t is respectively smooth and easy, blocks up, blocks up, blocks up, blocks up, blocks up.

Table 12 calculate adjacent segments block up probability time rule (rule 3)

Target road section is blocked up probability P i tfor:

P i t = 403 403 + 314 = 56.2 %

The probability that blocks up is greater than 50%, and the traffic behavior of prediction t is for blocking up.

traffic congestion Strength co-mputation

The calculation process of the traffic congestion intensity of present embodiment is as follows:

C i = &Sigma; t = 1 T x t i a i l t i s t - - - ( 3 )

x t i = 1 , v &OverBar; t i < v T 0 , v &OverBar; t i &GreaterEqual; v T - - - ( 4 )

Wherein, C ifor the traffic congestion intensity of section i. for the discriminant function that blocks up, for the average speed of section i within t the time interval, v tfor congestion status discrimination threshold, a ifor the number of track-lines of section i, for the block up length of section i within t the time interval, s tit is the duration in t the time interval.

draw traffic congestion intensity thermodynamic chart

Traffic behavior expression way (congestion index, sequential chart) for current main flow can only reflect the temporal behavior of regional area traffic behavior, the problems such as overall traffic behavior Changing Pattern can not be reflected, thermodynamic chart is adopted to express the distribution of traffic congestion herein, and by several thermodynamic charts synthesis Dynamic Graph.Thermodynamic chart can show the generation position of traffic congestion and the intensity of traffic congestion on map, make traffic congestion visual and understandable, Dynamic Graph is convenient to research jam road (congestion regions) to the impact of peripheral path (neighboring area), grasps the Evolution of road net traffic state.Before drawing thermodynamic chart, need first to determine hotspot location, focus coverage, traffic congestion strength retrogression rule and scheme of colour.

Hotspot location

Because traffic congestion occurs on section, therefore the position of focus is that the terminus in each section respectively arranges a focus along section distribution, and from the starting point of section, often distance 100 meters arranges a focus.

Focus coverage

As shown in Figure 7, the coverage of focus is the circle of radius 50 meters.In units of section, the traffic congestion intensity in this section is evenly distributed on all focuses in this section.If the traffic congestion intensity in section is less than the focus number in this section, (traffic congestion intensity is C i), be then assigned randomly to C ion individual focus.

Traffic congestion strength retrogression rule

For each focus, take the center of circle as starting point, traffic congestion intensity along with from distance of center circle from increase and linear attenuation.Effect as shown in Figure 8.

Scheme of colour

Traffic congestion intensity is divided into 5 grades, and 1 represents that traffic congestion intensity is minimum, and 5 represent that traffic congestion intensity is maximum, and the color corresponding to grade 1-5 is: water white transparency → blueness → redness → yellow → white.

When carrying out map match, other high-precision matching process also can be adopted to realize, as point-to-point coupling, map-matching method based on topological analysis.

Same, also can also select other high-precision velocity estimation algorithms during velocity estimation practical application, as the section velocity estimation model based on neural network.

In the present invention, the division of traffic behavior is with reference to Beijing to the definition in heavy congestion section, also can select other traffic behavior division rules during practical application, as " urban traffic control assessment indicator system " (2008) that the Ministry of Public Security issues.

The comparison of traffic congestion prognostic experiment result

1) traffic congestion prediction

In table 13, method 1 represents the accuracy of only using method 1, method 2, the rest may be inferred for method 3, accuracy after the comprehensive three kinds of methods of method representation, namely three kinds of methods are first used to obtain the judged result of three traffic behaviors, using the traffic behavior obtained by most method as the traffic behavior of target road section.As shown in Table 13, the accuracy that traffic congestion judges reaches more than 85% substantially, and precision of prediction is satisfactory.

Table 13 traffic congestion predicts the outcome analysis

2) traffic congestion is visual

Fig. 9, Figure 10 are Guangzhou traffic congestion intensity thermodynamic charts, thermodynamic chart can intuitively show traffic congestion position and coverage thereof, thermodynamic chart is synthesized traffic congestion dynamic changes of strength figure, traffic congestion thermodynamic chart more in the same time, be convenient to grasp the Evolution that urban road network traffic blocks up.

Above-described embodiments of the present invention, do not form limiting the scope of the present invention.Any amendment done within spiritual principles of the present invention, equivalent replacement and improvement etc., all should be included within claims of the present invention.

Claims (8)

1. a traffic congestion Forecasting Methodology, is characterized in that, comprising:
Adopt Floating Car to obtain gps data, by map-matching method, gps data is associated with the road on electronic chart; According to the data estimation section speed of coupling, utilize the traffic behavior in velocity estimated section, section, described traffic behavior comprises and to block up and unimpeded;
Historical traffic status information is utilized to build traffic congestion model; Association real-time traffic states information bank carries out traffic congestion prediction;
Traffic congestion model comprises traffic congestion Model I, traffic congestion modelⅱ and traffic congestion model III;
Suppose that interim one day of this week traffic behavior is identical with cycle traffic behavior on the same day on it,
Building traffic congestion Model I concrete mode is: trained as training sample by the traffic behavior in n cycle, add up each cycle on the same day t there is the probability that blocks up;
Target road section each cycle on the same day t there is the probability that blocks up:
P i t = &Sigma; j = 1 n &sigma; i j t n - - - ( 1 )
&sigma; i j t = 0 v i j t &GreaterEqual; v T 1 v i j t < v T - - - ( 2 )
Wherein: refer to that the probability blocked up appears in i-th section t; refer to the section speed of i-th section in a jth cycle t; v tit is congestion status discrimination threshold; refer to the traffic behavior of i-th section in a jth cycle t, 0 represents smooth and easy, and 1 expression is blocked up;
The probabilistic determination traffic behavior blocked up is there is according to above-mentioned target road section;
Building the concrete mode of traffic congestion modelⅱ is: in training sample, statistics target road section is when developing the traffic behavior of t from the traffic behavior in t-1 moment, adjacent segments is in the first traffic behavior regularity of distribution in t-1 moment, and described adjacent segments refers to the section adjacent with target road section;
Association real-time traffic states information bank carries out traffic congestion prediction; Be obtain target road section according to the first traffic behavior regularity of distribution to block up probability, then occur the traffic behavior of the probabilistic determination target road section of blocking up according to this target road section;
Building the concrete mode of traffic congestion model III is:
In training sample, statistics adjacent segments is when developing the traffic behavior of t from the traffic behavior in t-1 moment, and the section, upstream of this adjacent segments is in the second traffic behavior regularity of distribution in t-1 moment, and section, described upstream refers to the section of the upstream of adjacent segments;
In training sample, statistics adjacent segments is when developing the traffic behavior of t from the traffic behavior in t-1 moment, and the downstream road section of this adjacent segments is in the 3rd traffic behavior regularity of distribution in t-1 moment, and described downstream road section refers to the section in the downstream of adjacent segments;
In training sample, when statistics target road section develops the traffic behavior of t from the traffic behavior in t-1 moment, adjacent segments is in the 4th traffic behavior regularity of distribution of t;
Association real-time traffic states information bank carries out traffic congestion prediction; Obtain t adjacent segments according to the second traffic behavior regularity of distribution to block up probability, then the traffic behavior of this adjacent segments of probabilistic determination that blocks up according to this adjacent segments; Simultaneously obtain t adjacent segments according to the 3rd traffic behavior regularity of distribution to block up probability, again according to the traffic behavior of this adjacent segments of probabilistic determination that blocks up of this adjacent segments, finally, according to the traffic behavior of the t adjacent segments predicted by second and third traffic behavior regularity of distribution, and the 4th the traffic behavior regularity of distribution obtain t target congested link probability, then the traffic behavior of this target road section of probabilistic determination of blocking up according to this target road section;
When predicting traffic congestion, adopting above-mentioned traffic congestion Model I, traffic congestion modelⅱ and traffic congestion model III three kinds of models couplings, adopting during three kinds of models couplings and predicting with majority principle;
Build in traffic congestion model, the evolution process developing the traffic behavior of t from the traffic behavior in t-1 moment is:
When the traffic behavior in t-1 moment is for blocking up, the traffic behavior of t is for block up or smooth and easy;
When the traffic behavior in t-1 moment is smooth and easy, the traffic behavior of t is for block up or smooth and easy.
2. traffic congestion Forecasting Methodology according to claim 1, is characterized in that, according to blocking up, probability carries out traffic behavior judgement, when the probability that blocks up is more than or equal to 50%, then and traffic behavior TS ifor blocking up, otherwise be smooth and easy.
3. traffic congestion Forecasting Methodology according to claim 2, is characterized in that, described map-matching method is specially: obtain candidate road section collection, screens possible coupling section, also referred to as candidate road section by building an error band around anchor point; Its mid point is determined to mate section to the distance of each candidate road section and the course of GPS point to lines matching according to GPS point; Point sequence coupling, using topology connectivity as the rational constraint condition of coupling, chooses the maximum section combination of coupling weight summation as final matching sequence.
4. traffic congestion Forecasting Methodology according to claim 3, it is characterized in that, carry out section velocity estimation according to the adaptively selected velocity estimation model of matching result, the appraising model of employing is: distance verses time model, Velocity-time trapezoidal integration model, vehicle tracking model or ground spot speed harmonic average model.
5. apply the method for visualizing of the traffic congestion Forecasting Methodology described in any one of the claims 1 to 4 for one kind, it is characterized in that, it is the thermodynamic chart drawing traffic congestion intensity, before drafting thermodynamic chart, need to determine hotspot location, focus coverage, traffic congestion strength retrogression rule and scheme of colour;
Calculate the intensity C that blocks up of traffic i;
Hotspot location, the position of focus is that the terminus in each section respectively arranges a focus, and from the starting point of section, often distance M rice arranges a focus along section distribution;
Focus coverage, the coverage of focus is the circle of radius M/2 rice, and in units of section, the traffic congestion intensity in this section is evenly distributed on all focuses in this section; If the traffic congestion intensity C in section ibe less than the focus number in this section, be then assigned randomly to C ion individual focus;
Traffic congestion strength retrogression rule, for each focus, take the center of circle as starting point, traffic congestion intensity along with from distance of center circle from increase and linear attenuation;
Scheme of colour, is divided into N number of grade by traffic congestion intensity, and adjacent rank adopts different colours mark.
6. method for visualizing according to claim 5, is characterized in that, the intensity C that blocks up of traffic irealize in the following ways:
C i = &Sigma; t = 1 T x t i a i l t i s t - - - ( 3 )
x t i = 1 , v &OverBar; t i < v T 0 , v &OverBar; t i &GreaterEqual; v T - - - ( 4 )
Wherein, C ifor the traffic congestion intensity of section i, for the discriminant function that blocks up, for the average speed of section i within t the time interval, v tfor congestion status discrimination threshold, a ifor the number of track-lines of section i, for the block up length of section i within t the time interval, s tit is the duration in t the time interval.
7. the method for visualizing according to claim 5 or 6, is characterized in that, described scheme of colour is that traffic congestion intensity is divided into 5 grades, and 1 represents that traffic congestion intensity is minimum, and 5 represent that traffic congestion intensity is maximum.
8. method for visualizing according to claim 7, is characterized in that, the color corresponding to grade 1-5 is: water white transparency → blueness → redness → yellow → white.
CN201410381904.3A 2014-08-05 2014-08-05 A kind of traffic congestion Forecasting Methodology and method for visualizing CN104157139B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201410381904.3A CN104157139B (en) 2014-08-05 2014-08-05 A kind of traffic congestion Forecasting Methodology and method for visualizing

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201410381904.3A CN104157139B (en) 2014-08-05 2014-08-05 A kind of traffic congestion Forecasting Methodology and method for visualizing

Publications (2)

Publication Number Publication Date
CN104157139A CN104157139A (en) 2014-11-19
CN104157139B true CN104157139B (en) 2016-01-13

Family

ID=51882626

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201410381904.3A CN104157139B (en) 2014-08-05 2014-08-05 A kind of traffic congestion Forecasting Methodology and method for visualizing

Country Status (1)

Country Link
CN (1) CN104157139B (en)

Families Citing this family (34)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN104573859A (en) * 2014-12-31 2015-04-29 合肥城市云数据中心有限公司 Human traffic prediction method based on Wifi positioning and cloud data processing technology
CN104766260A (en) * 2015-02-04 2015-07-08 北京奇虎科技有限公司 Passenger transport information display method and device
CN104933860B (en) * 2015-05-20 2017-07-11 重庆大学 Bus traffic congestion delay time at stop Forecasting Methodology based on gps data
CN104848871B (en) * 2015-06-09 2018-11-16 刘光明 Map and traffic route with traffic condition variation tendency
CN104851298B (en) * 2015-06-09 2017-06-27 刘光明 Prediction traffic and running time
CN105185107A (en) * 2015-07-23 2015-12-23 合肥革绿信息科技有限公司 GPS-based traffic running tendency prediction method
CN105095481B (en) * 2015-08-13 2018-06-26 浙江工业大学 Extensive taxi OD data visualization analysis methods
CN105184728A (en) * 2015-09-15 2015-12-23 广州地理研究所 Construction method of customized regular passenger coach transportation demand thermodynamic diagram
CN105184409A (en) * 2015-09-15 2015-12-23 广州地理研究所 Customized bus planned route travel demand thermodynamic diagram construction method
CN105139656B (en) * 2015-09-28 2017-09-29 百度在线网络技术(北京)有限公司 A kind of road condition Forecasting Methodology and device
CN105373805A (en) * 2015-10-09 2016-03-02 中国电子科技集团公司第二十八研究所 A multi-sensor maneuvering target tracking method based on the principle of maximum entropy
CN105390001B (en) * 2015-10-20 2016-07-06 北京长峰金鼎科技有限公司 The dynamic control method of traffic light
CN106683399A (en) * 2015-11-10 2017-05-17 湖南南车时代电动汽车股份有限公司 Method of acquiring road condition state of bus
CN105303839B (en) * 2015-12-04 2017-12-08 中国科学院遥感与数字地球研究所 The Forecasting Methodology and device in potential jam road crosspoint
CN105489006B (en) * 2015-12-15 2018-01-12 浙江工业大学 Multiple dimensioned vehicle flow visual analysis method based on GPS data from taxi
US9824580B2 (en) * 2015-12-17 2017-11-21 International Business Machines Corporation Method, computer readable storage medium and system for producing an uncertainty-based traffic congestion index
CN105741548A (en) * 2016-04-19 2016-07-06 上海理工大学 Method for generating traffic state cloud atlas
CN106874324B (en) * 2016-06-28 2020-09-08 阿里巴巴集团控股有限公司 Method and device for displaying data information association relation by using color
CN107689152A (en) * 2016-08-04 2018-02-13 西门子公司 Method and apparatus for carrying out traffic processing in a road network
CN106056954A (en) * 2016-08-16 2016-10-26 合肥东上多媒体科技有限公司 Vehicle traffic state display prompt system based on GPS
CN106530694B (en) * 2016-11-07 2019-01-11 深圳大学 A kind of traffic congestion prediction technique and system based on traffic congestion propagation model
CN106408945B (en) * 2016-11-28 2019-03-01 北京掌行通信息技术有限公司 A kind of traffic congestion evaluation method and system
CN107507442A (en) * 2017-06-29 2017-12-22 百度在线网络技术(北京)有限公司 Vehicle peccancy method for early warning and device, computer equipment and computer-readable recording medium
CN108257380B (en) * 2017-12-05 2020-11-10 北京掌行通信息技术有限公司 Method and system for detecting congestion event based on road condition information
CN108229724B (en) * 2017-12-06 2020-12-22 华南理工大学 Short-term traffic data flow prediction method based on temporal-spatial information fusion
CN108513676A (en) * 2018-04-25 2018-09-07 深圳市锐明技术股份有限公司 A kind of road conditions recognition methods, device and equipment
CN109754597A (en) * 2018-08-02 2019-05-14 银江股份有限公司 A kind of urban road area congestion regulating strategy recommender system and method
CN108898838B (en) * 2018-08-03 2020-08-28 首都经济贸易大学 Method and device for predicting airport traffic jam based on LSTM model
CN109410598B (en) * 2018-11-09 2021-01-15 浙江浩腾电子科技股份有限公司 Traffic intersection congestion detection method based on computer vision
CN109579861B (en) * 2018-12-10 2020-05-19 华中科技大学 Path navigation method and system based on reinforcement learning
CN109815279B (en) * 2018-12-27 2019-10-15 南京行者易智能交通科技有限公司 A kind of providing method and device of passenger flow distribution thermodynamic chart
CN109740811A (en) * 2018-12-28 2019-05-10 斑马网络技术有限公司 Passage speed prediction technique, device and storage medium
CN109754606A (en) * 2019-02-28 2019-05-14 山东浪潮云信息技术有限公司 A method of based on taxi location prediction congestion in road situation
CN110689719B (en) * 2019-05-31 2021-01-19 北京嘀嘀无限科技发展有限公司 System and method for identifying closed road sections

Family Cites Families (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20080033630A1 (en) * 2006-07-26 2008-02-07 Eun-Mi Lee System and method of predicting traffic speed based on speed of neighboring link
CN101739820B (en) * 2009-11-19 2012-09-26 北京世纪高通科技有限公司 Road condition predicting method and device
CN103942953A (en) * 2014-03-13 2014-07-23 华南理工大学 Urban road network dynamic traffic jam prediction method based on floating vehicle data

Also Published As

Publication number Publication date
CN104157139A (en) 2014-11-19

Similar Documents

Publication Publication Date Title
Yao et al. Short‐term traffic speed prediction for an urban corridor
Klappenecker et al. Finding available parking spaces made easy
CN105788260B (en) A kind of bus passenger OD projectional techniques based on intelligent public transportation system data
Aslani et al. Adaptive traffic signal control with actor-critic methods in a real-world traffic network with different traffic disruption events
CN103903433B (en) The Real-time and Dynamic method of discrimination of a kind of road traffic state and device
CN102110365B (en) Road condition prediction method and road condition prediction system based on space-time relationship
Li et al. An Open Traffic Light Control Model for Reducing Vehicles'$\hbox {CO} _ {2} $ Emissions Based on ETC Vehicles
Lee et al. A knowledge based real-time travel time prediction system for urban network
CN102509470B (en) System and method for realizing energy conservation and emission reduction of vehicle based on dynamic path planning
CN102436751B (en) Short-time forecasting method for traffic flow based on urban macroscopic road network model
CN104821080B (en) Intelligent vehicle traveling speed and time predication method based on macro city traffic flow
CN100444210C (en) Mixed controlling method of single dot signal controlling crossing
CN105825669B (en) A kind of system and method for identifying urban expressway traffic bottleneck
CN102855760B (en) On-line queuing length detection method based on floating vehicle data
CN102081859B (en) Control method of bus arrival time prediction model
CN106205114A (en) A kind of Freeway Conditions information real time acquiring method based on data fusion
CN100463009C (en) Traffic information fusion processing method and system
Mazloumi et al. An integrated framework to predict bus travel time and its variability using traffic flow data
CN101308604B (en) Traffic coordinating and controlling method with strategy of big range
Narzt et al. Self-organising congestion evasion strategies using ant-based pheromones
CN104299442B (en) A kind of city Route Travel Time Estimation Method based on pattern match
CN103927890B (en) A kind of Trunk Road Coordination signal control method based on dynamic O-D Matrix Estimation
Teodorović Fuzzy sets theory applications in traffic and transportation
CN102819955B (en) Road network operation evaluation method based on vehicle travel data
CN102646332B (en) Traffic state estimation device and method based on data fusion

Legal Events

Date Code Title Description
PB01 Publication
C06 Publication
SE01 Entry into force of request for substantive examination
C10 Entry into substantive examination
ASS Succession or assignment of patent right

Owner name: GUANGDONG FUNDWAY TECHNOLOGY CO., LTD.

Effective date: 20150902

CB03 Change of inventor or designer information

Inventor after: He Zhaocheng

Inventor after: Ye Weijia

Inventor after: Sha Zhiren

Inventor before: He Zhaocheng

Inventor before: Ye Weijia

TA01 Transfer of patent application right

Effective date of registration: 20150902

Address after: 510006 Panyu District City, Guangdong Province, University of outer ring road, No., No. 132

Applicant after: Sun Yat-sen University

Applicant after: Guangdong Fundway Technology Co., Ltd.

Address before: 510006 Panyu District City, Guangdong Province, University of outer ring road, No., No. 132

Applicant before: Sun Yat-sen University

C41 Transfer of patent application or patent right or utility model
C53 Correction of patent for invention or patent application
COR Change of bibliographic data

Free format text: CORRECT: INVENTOR; FROM: HE ZHAOCHENG YE WEIJIA TO: HE ZHAOCHENG YE WEIJIA SHA ZHIREN

GR01 Patent grant
C14 Grant of patent or utility model