CN102360526B - Real-time monitoring method for road section state of high road - Google Patents
Real-time monitoring method for road section state of high road Download PDFInfo
- Publication number
- CN102360526B CN102360526B CN 201110302386 CN201110302386A CN102360526B CN 102360526 B CN102360526 B CN 102360526B CN 201110302386 CN201110302386 CN 201110302386 CN 201110302386 A CN201110302386 A CN 201110302386A CN 102360526 B CN102360526 B CN 102360526B
- Authority
- CN
- China
- Prior art keywords
- traffic flow
- traffic
- accident
- flow modes
- modes
- Prior art date
- Legal status (The legal status 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 status listed.)
- Active
Links
Images
Landscapes
- Traffic Control Systems (AREA)
Abstract
The invention relates to a real-time monitoring method for a road section state of a high road. The method comprises the following steps of: arranging traffic flow detecting points on the high road to be detected, and spacing the detecting points by 800m, wherein each detecting point covers 400m of upstream road sections and 400m of downstream road sections; acquiring historical traffic flow dataand accident data; extracting dangerous traffic flow operation states by using clustering analysis and regression analysis; and giving a method for monitoring and identifying the traffic flow operation states by using a decision tree approach. The method is easy to process and operate and is favorable for real-time identification of high road accident 'black spots'. The real-time monitoring method has practical engineering utilization value of monitoring accident dangerous road sections of the high road in real time.
Description
Technical field
The invention belongs to traffic intelligent management and control technology field, be the method for real-time of a kind of through street highway section state, can differentiate in real time the highway section that the accident risk is arranged on the through street.
Background technology
Along with the in-depth of reform and deepening continuously of urbanization, motorization process, from 20th century the mid-80s, China Transportation Industry develops rapidly, and road network structure obtains constantly perfect, the traffic infrastructure condition is constantly improved, and vehicle guaranteeding organic quantity and transport need increase sharply.But the inharmonious and ladder of management of vehicle and road growth rate improper also brought serious traffic safety problem.Under such background, the traffic safety problem becomes the hot issue that more and more receives publicity, and improves China's traffic safety management working level, and for reducing casualties and property loss, the development new period, harmonious traffic had important practical significance.
In order to prevent and to reduce traffic hazard quantity, how to differentiate that accident highway section occurred frequently is the primary problem that solves.At present, traffic administration relevant departments mainly realize discriminating to accident black-spot by some traditional methods, these methods often need to gather casualty data, traffic data, weather data and the road parameters in longer a period of time, and the general in the world casualty data in 3 years of recommend adoption is analyzed.Although these methods are being that effectively these methods are difficult at short notice accident easily to be sent out the highway section and differentiate aspect the discriminating through street accident black-spot.
And on the other hand, in the research of traffic safety forward position, scholars begin to pay close attention to the relation between real-time traffic stream information and the traffic safety.At present, whether the different running statuses of traffic flow are relevant with traffic safety, and the degree of association is actually much between them, and whether accident exists certain to can be used as " danger " traffic flow modes of accident sign before occuring, and is never well answered.The present invention proposes the method for a kind of extraction through street " danger " traffic flow modes, and the method is applied in the real-time discriminating in through street accident hazard highway section, by detecting the traffic flow modes of each check point on the through street, differentiate in real time the accident hazard highway section, thereby solve the defective that existing accident black-spot discrimination method is difficult to differentiate at short notice Dangerous Area.
Summary of the invention
The problem to be solved in the present invention is: overcome the deficiency of existing accident black-spot discrimination method, the real-time traffic flow data that utilizes existing intelligent transportation system to detect proposes a kind of method that can real time discriminating through street Dangerous Area.
Technical scheme of the present invention is: the method for real-time of through street accident section state may further comprise the steps:
1) on through street, every 800 meters a Traffic Flow Detector is set, collect the casualty data of the through street of examine, establish the scene that each plays traffic hazard, thereby determine a Traffic Flow Detector nearest with the traffic hazard scene, extract this check point traffic data in half an hour before accident occurs, comprise traffic flow rate and speed, its sampling precision is to gather once in per 5 minutes;
2) for every traffic hazard, adopt the case-control study method to choose under the normal traffic stream mode of incident highway section the traffic occupation rate data when namely not having an accident, described 1: 4 finger in 1: 4 ratio, corresponding to each accident example, choose 4 corresponding normal traffic stream mode examples;
Traffic data is combined into the traffic flow data sample in the normal traffic flow status data of the correspondence of 3) traffic data and the step 2 before accident is occured) choosing and the minimum check point 2 months of having an accident on every side, adopt K-means Dynamic Cluster Analysis method, select 2 traffic flow parameters of traffic flow rate and speed that the traffic flow data sample is carried out cluster analysis;
4) carry out the Logistic recurrence to divided each traffic flow modes that obtains by K-means Dynamic Cluster Analysis method, independent variable is traffic flow modes, whether dependent variable for accident, whether the level of signifiance P-value according to each traffic flow modes front coefficient in the Logistic recurrence comes significantly to determine whether traffic flow modes is dangerous traffic flow running rate, when level of signifiance P-value less than 0.05 the time, representing this traffic flow modes is dangerous traffic flow modes, otherwise, when level of signifiance P-value greater than 0.05 the time, representing this traffic flow modes is not dangerous traffic flow modes;
The border of 5) adopting the QUEST decision-tree model to establish various traffic flow modes is used for differentiating the real-time traffic stream mode: with the explanatory variable of traffic density as the QUEST decision-tree model, take traffic flow modes as target variable, establish the border of various traffic flow modes, be traffic density scopes corresponding to various traffic flow modes, traffic density=traffic flow rate/speed wherein, traffic flow rate and speed are obtained by check point;
6) gathered traffic flow rate and the speed of each check point every 5 minutes, calculate the traffic density of each point, traffic density scope corresponding to various traffic flow modes according to the establishment of QUEST decision-tree model, judge whether the whole piece through street has check point dangerous traffic flow modes to occur, if the traffic density that certain check point obtains drops on traffic density scope corresponding to dangerous traffic flow modes, dangerous traffic flow modes appearred near this check point was described, show that then 800 meters highway sections that this check point covers are Dangerous Area, the danger that accident is arranged is carried out early warning.
Whether the traffic data that the present invention utilizes on the through street each check point to gather dangerous traffic flow modes occurs near detecting in real time each check point, thereby reaches the purpose that real-time discriminating accident is easily sent out the highway section.For traditional through street accident black-spot discrimination method, this method can be found the accident hazard highway section within a short period of time, and processes simple.Thereby, use this method to differentiate that the dangerous traffic flow running rate of through street has actual engineering application and is worth.
Description of drawings
Fig. 1 is the check point schematic diagram of the embodiment of the invention.
Fig. 2 is the process flow diagram of monitoring method of the present invention.
Fig. 3 is the modeling program block diagram of integrated use K-means cluster of the present invention, Logistic recurrence and QUEST decision-tree model.
Fig. 4 is phase identification figure after the cluster of the embodiment of the invention.
Embodiment
K-means cluster, Logistic are returned in the present invention and the QUEST decision-tree model is easily sent out the highway section for the Real-Time Monitoring accident, and the method for real-time in a kind of through street accident hazard highway section may further comprise the steps:
1) on through street, every 800 meters a Traffic Flow Detector is set, collect the casualty data of the through street of examine, establish the scene that each plays traffic hazard, thereby determine a Traffic Flow Detector nearest with the traffic hazard scene, extract this check point traffic data in half an hour before accident occurs, comprise traffic flow rate and speed, its sampling precision is to gather once in per 5 minutes;
2) for every traffic hazard, adopt the case-control study method to choose under the normal traffic stream mode of incident highway section the traffic occupation rate data when namely not having an accident, described 1: 4 finger in 1: 4 ratio, corresponding to each accident example, choose 4 corresponding normal traffic stream mode examples;
Traffic data is combined into the traffic flow data sample in the normal traffic flow status data of the correspondence of 3) traffic data and the step 2 before accident is occured) choosing and the minimum check point 2 months of having an accident on every side, adopt K-means Dynamic Cluster Analysis method, select 2 traffic flow parameters of traffic flow rate and speed that the traffic flow data sample is carried out cluster analysis;
4) carry out the Logistic recurrence to divided each traffic flow modes that obtains by K-means Dynamic Cluster Analysis method, independent variable is traffic flow modes, whether dependent variable for accident, whether the level of signifiance P-value according to each traffic flow modes front coefficient in the Logistic recurrence comes significantly to determine whether traffic flow modes is dangerous traffic flow running rate, when level of signifiance P-value less than 0.05 the time, representing this traffic flow modes is dangerous traffic flow modes, otherwise, when level of signifiance P-value greater than 0.05 the time, representing this traffic flow modes is not dangerous traffic flow modes;
5) because clustering method is that the historical data of collecting is classified, can only provide the cluster centre of each state, which kind of state can't differentiate current traffic flow is under actually, thereby the present invention adopts the QUEST decision-tree model to establish the border of various traffic flow modes, be used for differentiating the real-time traffic stream mode: with the explanatory variable of traffic density as the QUEST decision-tree model, take traffic flow modes as target variable, establish the border of various traffic flow modes, be traffic density scopes corresponding to various traffic flow modes, traffic density=traffic flow rate/speed wherein, traffic flow rate and speed are obtained by check point;
6) gathered traffic flow rate and the speed of each check point every 5 minutes, calculate the traffic density of each point, traffic density scope corresponding to various traffic flow modes according to the establishment of QUEST decision-tree model, judge whether the whole piece through street has check point dangerous traffic flow modes to occur, if the traffic density that certain check point obtains drops on traffic density scope corresponding to dangerous traffic flow modes, dangerous traffic flow modes appearred near this check point was described, show that then 800 meters highway sections that this check point covers are Dangerous Area, the danger that accident is arranged is carried out early warning.
Step 1)-4) different conditions is divided in traffic flow, and utilized P-value therefrom to establish dangerous traffic flow modes, finish the division to traffic flow modes; Step 5) establishes traffic density scope corresponding to dangerous traffic flow modes, when current traffic density drops on traffic density scope corresponding to dangerous traffic flow modes, judge that namely current traffic flow is precarious position.
The below uses accompanying drawings the present invention.In through street regional upstream and downstream to be detected check point is set, at the measuring points placement checkout equipment, for example electromagnetic induction coil or other traffic flow checkout equipments are separated by between each check point 800 meters, as shown in Figure 1.
Check point is take 5 minutes as a sample unit collection traffic flow rate and speed.Make that X is traffic flow parameter traffic flow rate and speed in above-mentioned 5 minutes, whether y for accident, and the y value is 1 during accident, and the y value is not-1 when having accident, then,
The practice process is divided into model calibration and model uses two processes.
Model calibration: collect or gather the traffic data in above-mentioned each check point a period of time, contain accident sample and non-accident sample.In order to guarantee the accuracy of model, guarantee that model can reflect the relation between traffic flow detected parameters and the traffic hazard generation, sample should be enough large, should be less than 50 according to existing research accident example.According to abovementioned steps 1)～5), calculate phase identification, and establish therein dangerous traffic flow modes, finally obtained the boundary value of dangerous traffic flow modes by classification tree.
Model uses: 5 minutes traffic flow rates of each check point of Real-time Collection and speed, the dangerous traffic flow modes boundary value of establishing in the traffic flow rate that each point is detected and speed and the model calibration compares, when finding that traffic flow rate that certain point detects and speed are in dangerous traffic flow modes in-scope, there is the danger of accident in 800 meters highway sections that show this point covering.Should have by through street variable information display board the danger of accident this moment to the highway section that the driver issues the covering of this point, the reminding driver vehicle that drives with caution.
The true traffic data that gathers on the I-880 highway of the present embodiment with SF Bay area, California and casualty data (I-880 data) are showed the workflow of this method.Data comprise traffic flow rate and the speed of examine point, gather once every 5 minutes.
80 traffic hazard examples in the I-880 database, have altogether been extracted, choose the traffic flow datas (comprising traffic flow rate and speed) that 80 traffic hazards occur, utilized simultaneously the case-control study method to choose the traffic data that plays (not having accident) under the normal traffic stream mode of traffic hazard corresponding to each in 1: 4 ratio.Traffic flow data in accident group data and normal traffic flow data and the check point 2 months is combined into the traffic flow data sample.
Utilize to use K-means dynamic clustering method at first traffic flow modes to be divided into such as 6 states among Fig. 4, and utilize Logistic to return Discovery Status 5 and state 6 for traffic hazard front dangerous traffic flow modes occurs.
After establishing dangerous traffic flow modes, utilize the QUEST decision-tree model to establish boundary value such as the table 1 of 6 class traffic flow modes:
The boundary value of table 16 class traffic flow modes
| Traffic flow modes | Traffic density (traffic flow rate is divided by speed) boundary value |
| Traffic flow modes 1 | 0＜traffic density≤17.6 |
| Traffic flow modes 2 | 17.6＜traffic density≤27.2 |
| Traffic flow modes 3 | 27.2＜traffic density≤36.8 |
| Traffic flow modes 4 | 36.8＜traffic density≤56.0 |
| Traffic flow modes 5 (dangerous traffic flow modes) | 56.0＜traffic density≤76.8 |
| Traffic flow modes 6 (dangerous traffic flow modes) | Traffic density 〉=76.8 |
After the boundary value of establishing each traffic flow modes, gathered traffic flow rate and the speed of each check point every 5 minutes, the real-time traffic flow data that each point is detected and the boundary value of dangerous traffic flow modes compare, when the real-time traffic flow data that detects of finding certain point was in traffic flow modes 5 or 6 (dangerous traffic flow modes) in-scope, there was the danger of accident in 800 meters highway sections that show this point covering.Should issue the danger that this point has accident to the driver this moment by the variable display board of highway, the reminding driver vehicle that drives with caution.
Because it is simple that this method is processed, and is not affected by human factors, and is convenient to differentiate in real time through street accident hazard highway section; Therefore, using this method monitoring through street accident hazard highway section to have actual engineering application is worth.
Claims (1)
1. the method for real-time of a through street highway section state is characterized in that may further comprise the steps:
1) on through street, every 800 meters a Traffic Flow Detector is set, collect the casualty data of the through street of examine, establish the scene that each plays traffic hazard, thereby determine a Traffic Flow Detector nearest with the traffic hazard scene, extract this check point traffic data in half an hour before accident occurs, comprise traffic flow rate and speed, its sampling precision is to gather once in per 5 minutes;
2) for every traffic hazard, adopt the case-control study method to choose under the normal traffic stream mode of incident highway section in the 1:4 ratio, the traffic occupation rate data when namely not having an accident, described 1:4 refers to, corresponding to each accident example, choose 4 corresponding normal traffic stream mode examples;
The normal traffic flow status data of the correspondence of 3) traffic data and the step 2 before accident is occured) choosing, and traffic data is combined into the traffic flow data sample in the minimum check point 2 months of having an accident on every side, adopt K-means Dynamic Cluster Analysis method, select 2 traffic flow parameters of traffic flow rate and speed that the traffic flow data sample is carried out cluster analysis;
4) carry out the Logistic recurrence to divided each traffic flow modes that obtains by K-means Dynamic Cluster Analysis method, independent variable is traffic flow modes, whether dependent variable for accident, whether the level of signifiance P-value according to each traffic flow modes front coefficient in the Logistic recurrence comes significantly to determine whether traffic flow modes is dangerous traffic flow running rate, when level of signifiance P-value less than 0.05 the time, representing this traffic flow modes is dangerous traffic flow modes, otherwise, when level of signifiance P-value greater than 0.05 the time, representing this traffic flow modes is not dangerous traffic flow modes;
The border of 5) adopting the QUEST decision-tree model to establish various traffic flow modes is used for differentiating the real-time traffic stream mode: with the explanatory variable of traffic density as the QUEST decision-tree model, take traffic flow modes as target variable, establish the border of various traffic flow modes, be traffic density scopes corresponding to various traffic flow modes, traffic density=traffic flow rate/speed wherein, traffic flow rate and speed are obtained by check point;
6) gathered traffic flow rate and the speed of each check point every 5 minutes, calculate the traffic density of each point, traffic density scope corresponding to various traffic flow modes according to the establishment of QUEST decision-tree model, judge whether the whole piece through street has check point dangerous traffic flow modes to occur, if the traffic density that certain check point obtains drops on traffic density scope corresponding to dangerous traffic flow modes, dangerous traffic flow modes appearred near this check point was described, show that then 800 meters highway sections that this check point covers are Dangerous Area, the danger that accident is arranged is carried out early warning.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110302386 CN102360526B (en) | 2011-09-28 | 2011-09-28 | Real-time monitoring method for road section state of high road |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN 201110302386 CN102360526B (en) | 2011-09-28 | 2011-09-28 | Real-time monitoring method for road section state of high road |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN102360526A CN102360526A (en) | 2012-02-22 |
| CN102360526B true CN102360526B (en) | 2013-09-18 |
Family
ID=45585849
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN 201110302386 Active CN102360526B (en) | 2011-09-28 | 2011-09-28 | Real-time monitoring method for road section state of high road |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN102360526B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102610098B (en) * | 2012-03-09 | 2014-05-07 | 西安费斯达自动化工程有限公司 | Method for setting road section traffic flow detector based on discrete model |
| CN103116979B (en) * | 2013-01-17 | 2015-01-28 | 东南大学 | Road accident dark spot identification system based on system safety index distribution method |
| CN103198713B (en) * | 2013-03-21 | 2016-01-06 | 东南大学 | Based on the vehicle regulate and control method of the minimizing traffic hazard of traffic data and weather data |
| CN103198707B (en) * | 2013-03-21 | 2016-03-02 | 东南大学 | A kind of vehicle regulate and control method based on traffic flow character dangerous under fine day situation |
| CN104182453A (en) * | 2014-06-20 | 2014-12-03 | 银江股份有限公司 | Distributed map matching method for massive historical floating car data |
| CN104408918B (en) * | 2014-11-20 | 2016-06-01 | 江苏物联网研究发展中心 | Based on the method that the floating car data of MapReduce Computational frame mates with road |
| CN105046952B (en) * | 2015-06-17 | 2017-12-08 | 天津职业技术师范大学 | The sparse air-ground associated detecting method of road traffic accident |
| CN106935052B (en) * | 2015-12-30 | 2020-12-18 | 沈阳美行科技有限公司 | Speed change prompting method and device based on driving data |
| CN105869398B (en) * | 2016-05-10 | 2018-10-23 | 浙江工业大学 | A kind of unimpeded degree judgment method of road traffic based on K-means clusters |
| CN106485919B (en) * | 2016-10-11 | 2019-06-18 | 东南大学 | A method of determining that through street fixed point tachymeter influences traffic accident quantity |
| CN106781474B (en) * | 2016-12-21 | 2019-03-12 | 东南大学 | A method of traffic accident spot self-healing ability is judged based on traffic video |
| CN107590766A (en) * | 2017-09-28 | 2018-01-16 | 东南大学 | A kind of method of discrimination of the land used combination form related to road traffic accident risk |
| CN108417091A (en) * | 2018-05-10 | 2018-08-17 | 武汉理工大学 | Driving risk section identification based on net connection vehicle and early warning system and method |
| CN108682149B (en) * | 2018-05-21 | 2021-01-05 | 东南大学 | Highway accident black point road section line shape cause analysis method based on binomial Logistic regression |
| CN115188158A (en) * | 2022-06-22 | 2022-10-14 | 中国电子科技集团公司电子科学研究院 | Intelligent safety early warning method and device based on historical case space distribution |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6177885B1 (en) * | 1998-11-03 | 2001-01-23 | Esco Electronics, Inc. | System and method for detecting traffic anomalies |
| CN101706996A (en) * | 2009-11-12 | 2010-05-12 | 北京交通大学 | Method for identifying traffic status of express way based on information fusion |
| CN101739815A (en) * | 2009-11-06 | 2010-06-16 | 吉林大学 | On-line recognition method of road traffic congestion state |
-
2011
- 2011-09-28 CN CN 201110302386 patent/CN102360526B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6177885B1 (en) * | 1998-11-03 | 2001-01-23 | Esco Electronics, Inc. | System and method for detecting traffic anomalies |
| CN101739815A (en) * | 2009-11-06 | 2010-06-16 | 吉林大学 | On-line recognition method of road traffic congestion state |
| CN101706996A (en) * | 2009-11-12 | 2010-05-12 | 北京交通大学 | Method for identifying traffic status of express way based on information fusion |
Non-Patent Citations (2)
| Title |
|---|
| 交通事故黑点智能排查***的研究与实现;张鹏;《中国优秀硕士学位论文全文数据库 信息科技辑》;20051115(第07期);I238-299 * |
| 张鹏.交通事故黑点智能排查***的研究与实现.《中国优秀硕士学位论文全文数据库 信息科技辑》.2005,(第07期),I238-299. |
Also Published As
| Publication number | Publication date |
|---|---|
| CN102360526A (en) | 2012-02-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN102360526B (en) | Real-time monitoring method for road section state of high road | |
| CN102360525B (en) | Discriminant analysis-based high road real-time traffic accident risk forecasting method | |
| CN103646534B (en) | A kind of road real-time traffic accident risk control method | |
| CN103021176B (en) | Discriminating method based on section detector for urban traffic state | |
| CN104408925B (en) | Crossing evaluation of running status method based on display radar | |
| CN102968901B (en) | Method for acquiring regional congestion information and regional congestion analyzing device | |
| CN104240499B (en) | A kind of abnormal congestion points method of discrimination based on microwave data | |
| CN102568226B (en) | High speed variable speed limit control method based on adverse weather conditions | |
| CN101465059B (en) | Early-warning system for inspecting and judging urban road traffic safety situation | |
| CN103198711B (en) | Vehicle regulating and controlling method of lowering probability of traffic accidents of different severity | |
| CN104484996A (en) | Road segment traffic state distinguishing method based on multi-source data | |
| CN103198713B (en) | Based on the vehicle regulate and control method of the minimizing traffic hazard of traffic data and weather data | |
| CN103150900A (en) | Traffic jam event automatic detecting method based on videos | |
| CN101783074A (en) | Method and system for real-time distinguishing traffic flow state of urban road | |
| CN102881162A (en) | Data processing and fusion method for large-scale traffic information | |
| CN101882374A (en) | System and method for analysis of comprehensive information on vehicle trip distribution | |
| CN102735320A (en) | Method for identifying weights of cars based on dynamic strain of bridges | |
| CN102360524B (en) | Automatic detection and confirmation method of dangerous traffic flow characteristics of highway | |
| CN102368354A (en) | Road security evaluation method based on floating vehicle data acquisition | |
| CN100481153C (en) | Method for automatically inspecting highway traffic event based on offset minimum binary theory | |
| CN108922168A (en) | A kind of mid-scale view Frequent Accidents road sentences method for distinguishing | |
| CN103198707B (en) | A kind of vehicle regulate and control method based on traffic flow character dangerous under fine day situation | |
| CN107784832A (en) | A kind of method and apparatus for being used to identify the accident black-spot in traffic route | |
| CN101551940A (en) | Urban high-speed road traffic state judging and issuing system and method thereof | |
| CN106530714A (en) | Secondary traffic accident time prediction method based on traffic flow data |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant |