CN111724592B - Highway traffic jam detection method based on charging data and checkpoint data - Google Patents

Highway traffic jam detection method based on charging data and checkpoint data Download PDF

Info

Publication number
CN111724592B
CN111724592B CN202010537175.1A CN202010537175A CN111724592B CN 111724592 B CN111724592 B CN 111724592B CN 202010537175 A CN202010537175 A CN 202010537175A CN 111724592 B CN111724592 B CN 111724592B
Authority
CN
China
Prior art keywords
data
time
toll
gate
travel time
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
Application number
CN202010537175.1A
Other languages
Chinese (zh)
Other versions
CN111724592A (en
Inventor
赵敏
孙棣华
唐毅
王卫平
廖孝勇
吴霄
陈星州
刘廷让
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Chongqing University
Original Assignee
Chongqing University
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 Chongqing University filed Critical Chongqing University
Priority to CN202010537175.1A priority Critical patent/CN111724592B/en
Publication of CN111724592A publication Critical patent/CN111724592A/en
Application granted granted Critical
Publication of CN111724592B publication Critical patent/CN111724592B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/01—Detecting movement of traffic to be counted or controlled
    • G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
    • G08G1/0125—Traffic data processing
    • G08G1/0133—Traffic data processing for classifying traffic situation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/23—Clustering techniques
    • G06F18/232—Non-hierarchical techniques
    • G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
    • G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
    • G—PHYSICS
    • G08—SIGNALLING
    • G08G—TRAFFIC CONTROL SYSTEMS
    • G08G1/00—Traffic control systems for road vehicles
    • G08G1/065—Traffic control systems for road vehicles by counting the vehicles in a section of the road or in a parking area, i.e. comparing incoming count with outgoing count

Landscapes

  • Engineering & Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Artificial Intelligence (AREA)
  • Probability & Statistics with Applications (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Traffic Control Systems (AREA)
  • Devices For Checking Fares Or Tickets At Control Points (AREA)

Abstract

The invention discloses a highway traffic jam detection method based on charging data and gate data, which considers that the charging data and the gate data are comprehensively utilized to predict traffic jam, the data source has diversity, the charging data is supplemented by the gate data, the problem of less data between adjacent intercommunicated charging stations can be well solved, and the prediction is more accurate; the running state of the road is analyzed by adopting a Gaussian mixture clustering algorithm, the consideration is thorough, different traffic forms of the road can be comprehensively simulated and analyzed, and the scientificity and comprehensiveness are higher.

Description

Highway traffic jam detection method based on charging data and checkpoint data
Technical Field
The invention relates to the field of traffic data analysis and processing, and discloses a method for detecting traffic jam of highway sections based on networking charging data and access data.
Background
The accurate and timely detection of the occurrence of the traffic jam on the highway is a precondition that relevant departments adopt management measures to prevent secondary accidents and reduce economic loss. Because the highway of our country carries out the closed management, the road toll system is very perfect, thus has accumulated the magnanimity charging data and data of the card gate. The abundant single-vehicle information in the charging data and the gate data and the detailed record of the vehicle running on the expressway can reflect the traffic running state of the expressway section to a certain extent. For a long time, the charging data and the gate data are only used as the basis for the related departments to carry out highway charging service, how to mine the traffic state information contained in the charging data and the gate data and use the information to realize the detection of traffic jam on highway sections have important theoretical and practical research significance for traffic managers to implement traffic control and traffic travelers to plan travel schemes.
In fact, some studies have been made to detect traffic congestion and determine traffic state by using highway networking charging data, for example, yang Jufen extracts road section travel time from charging data, and designs a traffic congestion index reflecting road traffic state based on the travel time. Meanwhile, in order to solve the problem of less data between adjacent toll stations, yang Jufen provides a traffic jam detection algorithm for fusing basic road sections and composite road sections of an expressway by combining the space-time layout characteristics of the toll stations. On a highway, most toll stations are intercommunicating toll stations, and for two adjacent intercommunicating toll stations, the vehicles are required to go up from the upstream intercommunicating toll station and then go down from the downstream intercommunicating toll station in order to obtain the travel time of the vehicles on the road section between the two toll stations. In practice, however, there are usually not so many vehicles going on and off from adjacent intercommunicating toll booths that the amount of toll data between two adjacent intercommunicating toll booths in a day is small, and the lack of data is a key factor that makes the toll data difficult to apply.
In order to solve the ambiguous problem of a vehicle driving path on an expressway and realize accurate toll collection, china sets an ambiguous card port on a road and determines the specific driving road of the vehicle by the matching between the card port and a toll station. The toll data records the information of the vehicle on the expressway, and also records the detailed bicycle information including the license plate of the vehicle, and the gate data also records the license plate of the passing vehicle, the passing time and other information. And the driving record of the vehicle on the road section between the gate and the toll station can be obtained according to the matching of the charging data and the license plate information of the gate data. The toll data is supplemented by the toll gate data, the problem that data between adjacent intercommunicated toll stations is less can be well solved, and the detection of the traffic jam of the highway section is feasible by combining the toll data and the toll gate data.
Disclosure of Invention
In view of the above, in order to solve the above mentioned problems, the present invention provides a method for detecting traffic jam of a highway based on charging data and access data.
The purpose of the invention is realized by the following technical scheme:
a highway traffic jam detection method based on charging data and access data comprises the following steps:
the method comprises the following steps: the method for extracting the travel time of the road section by combining the charging data and the checkpoint data comprises the following specific steps:
1) Extracting travel time of a single vehicle on a road section between toll stations according to the toll data;
2) The travel time of the vehicle on the toll station road section is extracted by combining the charging data and the toll data;
step two: the method comprises the following steps of carrying out space matching on a road section between adjacent toll stations and a road section between a gate and the toll station, and calculating the average travel time of the road section, wherein the specific method comprises the following steps:
1) And carrying out space matching on the road sections between the adjacent toll stations and the road sections of the bayonet-toll station. The difference of the driving time of vehicles which pass through the gate and get off the road from different toll stations at the same time is used as the travel time on the road section between the two corresponding toll stations;
2) Detecting and eliminating abnormal values in the travel time of the road section; the abnormal value with negative travel time is directly deleted, and the abnormal value with overlong travel time is detected by adopting a Grubbs (Grubbs) criterion method to the travel time of the vehicle passing through the exit toll station and the next lane within a certain time range;
step three: synthesizing original vehicle travel time data by using a rolling time interval;
step four: constructing and calculating a distinguishing characteristic; the specific method comprises the following steps:
1) Calculating the average travel speed v (t) of the road section at the current moment;
2) Setting proper time interval number n, calculating the average travel speed v (t) of the road section at the current moment and the average value of the previous n time intervals
Figure BDA0002537415500000022
Deviation of (2);
3) Calculating the standard deviation SD of the average travel speed of the previous n time interval road sections;
step five: performing Gaussian mixture clustering analysis to divide different road traffic running states; the specific method comprises the following steps:
1) Determining the number K of clustering categories;
2) Random selection of initial model parameter pi k ,μ k ,σ k Calculating the observed data x j From a Gaussian distribution model
Figure BDA0002537415500000021
Probability of (gamma) of jk ；
3) Determining an objective function Q (theta ) (i) )；
4) In order to make the objective function Q (theta ) (i) ) Reaching the maximum value, updating the clustering model parameter pi k ,μ k ,σ k ；
5) Repeating the steps 2) and 3) to continue iteration until the iteration termination condition is met, and obtaining the final clustering model parameters
Figure BDA0002537415500000031
Wherein each set of parameters represents a category; the iteration termination condition is
||Q(θ,θ (i+1) )-Q(θ,θ (i) ) | | < epsilon or t = t max
Wherein Q (θ, θ) (i+1) ) Is the value of the objective function at the i +1 th iteration, Q (theta ) (i) ) Is the value of the objective function at the ith iteration, epsilon is the iteration termination condition threshold, t max Is the maximum number of iterations;
step six: and detecting the traffic jam of the road sections between the adjacent toll stations of the highway by using the clustering result.
Further, the calculation formula in step 1) of the first step is as follows:
T＝t out -t in
where T is the travel time of the vehicle on the section between the entrance and exit toll stations, T out Time of vehicle passing through exit toll station, t in The time when the vehicle passes through the entrance toll station;
the calculation formula of the step 2) in the first step is as follows:
T c ＝t out -t c
in the formula, T c For the travel time, t, of the vehicle on the route between the toll gate and toll station out Time of vehicle passing through exit toll station, t c The time for the vehicle to pass the gate.
Further, the specific method in the step 1) in the step two is as follows:
the difference of the driving time of vehicles passing through the gate and leaving the lane from different toll stations at the same time is used as the travel time of the road section between the two toll stations:
T(k+1,k)＝T c (k)-T c (k+1)|t c (k)＝t c (k+1)
in the formula, T c (k + 1) travel time on the section of toll-gate k +1 for a vehicle passing through the toll gate and descending from the toll-gate k +1, t c (k) And t c (k + 1) are the times when vehicles descending from the toll station k and the toll station k +1 pass the gate, respectively.
Further, the expression of the objective function in the fifth step is
Figure BDA0002537415500000032
In the formula, n k Represents the number of data generated by the k-th Gaussian distribution model in the N observed data, pi k The k-th class distribution probability can also be regarded as the weight coefficient of each Gaussian distribution; mu.s k Is the mean, σ, of the kth class k Is the standard deviation of the kth class.
Further, the probability γ in the step five jk Is calculated by the formula
Figure BDA0002537415500000041
In the formula, pi k Is the k-th class distribution probability, and can also be regarded as the weight of each Gaussian distributionA weight coefficient; mu.s k Is the mean, σ, of the kth class k Is the standard deviation of the kth class.
Further, the updating of the cluster model parameters in the fifth step is performed according to the following formula:
Figure BDA0002537415500000042
Figure BDA0002537415500000043
Figure BDA0002537415500000044
in the formula, N is the number of observed data,
Figure BDA0002537415500000045
for the parameter theta of the next iteration (i+1) And K is the number of the clustering categories.
Due to the adoption of the technical scheme, the invention has the following beneficial effects:
the traffic jam prediction method considers that the traffic jam is predicted by comprehensively utilizing the charging data and the card port data, the data source has diversity, the charging data is supplemented by utilizing the card port data, the problem that the number of data between adjacent intercommunicated charging stations is small can be well solved, and the prediction is more accurate; the running state of the road is analyzed by adopting a Gaussian mixture clustering algorithm, the consideration is thorough, different traffic forms of the road can be comprehensively simulated and analyzed, and the scientificity and comprehensiveness are higher.
Additional advantages, objects, and features of the invention will be set forth in part in the description which follows and in part will become apparent to those having ordinary skill in the art upon examination of the following or may be learned from practice of the invention. The objectives and other advantages of the invention may be realized and attained by the means of the instrumentalities and combinations particularly pointed out hereinafter.
Drawings
The drawings of the present invention are described below.
FIG. 1 is a schematic view of a traffic jam detection process for a highway section based on toll data and gate data;
FIG. 2 is a schematic diagram of the distribution of toll stations and gates for detecting road sections;
fig. 3 is a schematic diagram of a rolling time interval composition.
Detailed Description
The invention is further illustrated by the following figures and examples.
Example 1
As shown in fig. 1 to 3, the method for detecting traffic congestion of a highway based on charging data and gate data according to this embodiment includes the following steps:
the method comprises the following steps: extracting the travel time of the road section by combining the charging data and the checkpoint data; the specific method comprises the following steps:
1) The travel time of the individual vehicle on the section between the toll stations is extracted on the basis of the toll data. The charging data records information such as entrance and exit toll stations and passing time of motor vehicles on and off the highway, and if the time of the vehicles driving on the ramp is ignored, the difference value of the time of the vehicles going down and the time of the vehicles going up is the travel time of the vehicles on the road section between the entrance toll station and the exit toll station:
T＝t out -t in
where T is the travel time of the vehicle on the road between the entrance and exit toll booths, T out Time of vehicle passing through exit toll station, t in The time the vehicle passes through the entrance tollgate.
2) And extracting the travel time of the vehicle on the toll station road section by combining the charging data and the toll data. The toll gate data records information such as license plate and passing time of passing vehicles like the charging data, and the driving records of the vehicles on the road section between the toll gate and the toll station can be obtained according to the matching of the charging data and the license plate information of the toll gate data. Similarly, the running time of the vehicle on the ramp of the toll station is ignored, and the difference value of the time passing through the toll gate and the time passing through the toll station is the travel time of the matched vehicle on the road section between the toll gate and the toll station.
T c ＝t out -t c
In the formula, T c For the travel time, t, of the vehicle on the section between the toll gate and toll station out Time of vehicle passing through exit toll station, t c The time for the vehicle to pass the gate.
Step two: carrying out space matching on the road sections between adjacent toll stations and the road sections of the bayonet-toll station, and calculating the average travel time of the road sections; the specific method comprises the following steps:
1) And carrying out space matching on the road sections between adjacent toll stations and the road sections of the toll gate-toll station. The gate is generally not installed at the toll station location and therefore the gate-toll station path is not perfectly matched to the two toll stations path. Since the gate data records information of all passing vehicles, vehicles passing the gate at the same time may get off the lane from any downstream toll station. Combining the characteristics, the difference of the driving time of vehicles which pass through the gate and leave the lane from different toll stations at the same time is used as the travel time of the road section between the two corresponding toll stations:
T(k+1,k)＝T c (k)-T c (k+1)|t c (k)＝t c (k+1)
in the formula, T c (k + 1) travel time on the section of toll-gate k +1 for a vehicle passing through the toll gate and descending from the toll-gate k +1, t c (k) And t c (k + 1) is the time for the vehicle descending from the toll station k and the toll station k +1 to pass the gate, respectively.
In consideration of the fact that conditions for vehicles passing through the gate at the same time to leave the gate k and the gate k +1 are too strict, the same time point may be expanded to the same time window, and the average travel time of vehicles passing through the gate and leaving the gate k +1 within a predetermined time before and after the time point of vehicles passing through the gate k and leaving the gate k +1 may be calculated as T c (k + 1) approximation.
Figure BDA0002537415500000061
2) Abnormal values in the section travel time are detected and removed. Two main abnormal data exist in the license plate matching result. One is that the travel time of the vehicle on the toll-gate road section is negative. This may be the case because the vehicle travels from another route to the toll station to get off the lane and then passes through the gate, resulting in the time to get off the lane from the toll station being earlier than the time to pass through the gate. Deletion is made directly for such outliers. The other abnormal data is that the travel time of the road section is too long and is obviously different from other vehicles in the same time period. This may be due to an abnormal event such as a vehicle break. Such abnormal data also exists in the charging data, and the travel time at that time cannot reflect the traffic state of the link, and needs to be eliminated as well. However, a long travel time may be caused by traffic congestion, and in order to ensure that data in a congestion state is not deleted by mistake, abnormal values of travel times of vehicles passing the lower lane of an exit toll gate within a certain time range are detected by using a Grubbs (Grubbs) criterion method.
The Grabbs criterion is based on the assumption of a normal, and the detected univariate data set should approximately follow a normal distribution. Let the test data set be X = [ X = [ ] 1 ,x 2 ,…,x n ]Firstly, arranging X from small to large in sequence:
x(1),x(2),…,x(n)
if there is anomalous data, it must be one of the order statistics x (1) and x (n). Calculating the mean X and standard deviation s of X:
Figure BDA0002537415500000062
Figure BDA0002537415500000063
deviation values of the order statistics x (1) and x (n) from the average value x are calculated, respectively.
Figure BDA0002537415500000064
Figure BDA0002537415500000071
If G is 1 ＞G n Then x (1) is more likely to be an outlier, otherwise x (n) is detected. According to the number n of data in the detection data set X and the significance level alpha, consulting the Grabbs critical table to obtain a corresponding critical value G (n, alpha), if G is 1 Or G n If the value is larger than the critical value G (n, alpha), the corresponding data is an abnormal value, the abnormal value is removed from the detection data set X, and the steps are repeated until no abnormal value is detected.
Step three: synthesizing original vehicle travel time data by using a rolling time interval; as shown in fig. 3, in order to reduce the negative effect of the random fluctuation of the traffic state parameters on the algorithm detection performance, the average travel time of all vehicles descending from the exit toll station in the composite time interval is calculated as the travel time value of the vehicles on the road section in the output time interval after a certain output time interval by taking the time of the vehicles descending from the exit toll station as a reference.
Step four: constructing and calculating a distinguishing characteristic; the specific method comprises the following steps:
1) Calculating the average travel speed v (t) of the road section at the current moment:
Figure BDA0002537415500000072
in the formula, L is the mileage length of a road section between entrance and exit toll stations, and T is the average travel time.
2) Setting a proper time interval number n, generally taking a value of 3-10, and calculating the average travel speed v (t) of the road section at the current moment and the average value of the previous n time intervals
Figure BDA0002537415500000073
Deviation of (2):
Figure BDA0002537415500000074
Figure BDA0002537415500000076
3) Calculating the standard deviation SD of the average travel speed of the previous n time interval road sections:
Figure BDA0002537415500000075
step five: performing Gaussian mixture clustering analysis to divide different road traffic running states; the specific method comprises the following steps:
1) And determining the number K of the clustering categories. The average travel speed of the road section at the current moment reflects the running state of the road traffic flow, the deviation of the average travel speed of the road section at the current moment and the average value of the previous n time intervals represents the change degree of the average travel speed of the road section at the current moment relative to the previous n time intervals, and the standard deviation of the average travel speed of the road section at the previous n time intervals represents the stability of the average travel speed of the road section in the previous n time intervals. And (3) dividing the road traffic state into 4 conditions of smooth traffic, transition from smooth to congested, traffic jam, transition from congested to smooth and the like by combining the meanings of the three distinguishing features, and thus determining the number K =4 of the categories of the Gaussian mixture cluster.
2) Random selection of initial model parameter pi k ,μ k ,σ k Calculating the observed data x j From Gaussian distribution models
Figure BDA0002537415500000081
Probability of (gamma) jk ：
Figure BDA0002537415500000082
In the formula, pi k The class distribution probability can also be regarded as the weight coefficient of each Gaussian distribution; mu.s k Is the kth classOther mean value, σ k Is the standard deviation of the kth class.
3) Determining an expected Q (theta ) of a log-likelihood function with an objective function being the full data (i) )：
Figure BDA0002537415500000083
In the formula, n k The number of data generated by the k-th gaussian distribution model among the N observed data is represented.
4) In order to make the objective function Q (theta ) (i) ) When the maximum value is reached, updating the clustering model parameters according to the following formula:
Figure BDA0002537415500000084
Figure BDA0002537415500000085
Figure BDA0002537415500000086
in the formula (I), the compound is shown in the specification,
Figure BDA0002537415500000087
for the parameter theta of the next iteration (i+1) 。
5) After obtaining the updated clustering model parameters, repeating the steps 2) and 3) to continue iteration until the iteration termination condition is met, and obtaining the final clustering model parameters
Figure BDA0002537415500000088
Where each set of parameters represents a category. The iteration termination condition is as follows:
||Q(θ,θ (i+1) )-Q(θ,θ (i) ) | | < epsilon or t = t max
Wherein Q (theta ) (i+1) ) Is the target at the i +1 th iterationValue of function, Q (theta ) (i) ) Is the value of the objective function at the ith iteration, epsilon is the iteration termination condition threshold, t max The maximum number of iterations.
Step six: and detecting the traffic jam of the road sections between the adjacent toll stations of the expressway by using the clustering result. For the actual data x, the probabilities that it comes from 4 classes are calculated respectively:
Figure BDA0002537415500000091
in the formula, gamma k Is the probability that x is from the kth class.
After the probability that the actual data x belongs to each category is obtained, the category corresponding to the maximum probability is taken as the final category of the actual data x. And if the actual data do not belong to the traffic unblocked state, judging that the traffic jam occurs.

Claims (5)

1. A highway traffic jam detection method based on charging data and checkpoint data is characterized by comprising the following steps:
the method comprises the following steps: the method for extracting the travel time of the road section by combining the charging data and the checkpoint data comprises the following specific steps:
1) Extracting travel time of a single vehicle on a road section between toll stations according to the toll data;
2) The travel time of the vehicle on the section from the gate to the charging station is extracted by combining the charging data and the gate data;
step two: the method comprises the following steps of carrying out space matching on road sections between adjacent toll stations and road sections between a gate and the toll stations, and calculating the average travel time of the road sections, and specifically comprises the following steps:
1) Space matching is carried out on the road sections between the adjacent toll stations and the road sections from the toll gate to the toll station, and the difference of the driving time of vehicles passing through the toll gate and leaving the road from different toll stations at the same time is used as the travel time of the road sections between the two corresponding toll stations;
2) Detecting and eliminating abnormal values in the travel time of the road section; the abnormal value with negative travel time is directly deleted, and the abnormal value with overlong travel time is detected by adopting a Grubbs (Grubbs) criterion method to the travel time of the vehicle passing through the exit toll station and the next lane within a certain time range;
step three: synthesizing original vehicle travel time data by using a rolling time interval; the method comprises the following steps of calculating the average travel time of all vehicles descending from an exit toll station in a synthetic time interval after a certain output time interval is passed by taking the time of the vehicles descending from the exit toll station as a reference, and taking the average travel time as the travel time value of the vehicles on a road section in the output time interval;
step four: constructing and calculating a distinguishing characteristic; the specific method comprises the following steps:
1) Calculating the average travel speed v (t) of the road section at the current moment;
2) Setting proper time interval number n, calculating the average travel speed v (t) of the road section at the current moment and the average value of the previous n time intervals
Figure FDA0003879644970000011
Deviation of (2);
3) Calculating the standard deviation SD of the average travel speed of the previous n time interval road sections;
step five: performing Gaussian mixture clustering analysis to divide different road traffic running states; the specific method comprises the following steps:
1) Determining the number K of clustering categories;
2) Random selection of initial model parameter pi k ,μ k ,σ k Calculating the observed data x j From a Gaussian distribution model
Figure FDA0003879644970000021
Probability of (gamma) jk ；
3) Determining an objective function Q (theta ) (i) )；
4) In order to make the objective function Q (theta ) (i) ) Reaching the maximum value, updating the clustering model parameter pi k ,μ k ,σ k ；
5) Repeating the steps 2) and 3) to continue iteration until the iteration termination condition is met to obtain the final productCluster model parameter of (n) k ,μ k ,σ k Wherein each set of parameters represents a category; the iteration termination condition is
||Q(θ,θ (i+1) )-Q(θ,θ (i) ) | | < epsilon or t = t max
Wherein Q (θ, θ) (i+1) ) Is the value of the objective function at the i +1 th iteration, Q (theta ) (i) ) Is the value of the objective function at the ith iteration, epsilon is the iteration termination condition threshold, t max Is the maximum number of iterations;
wherein n is k Represents the number of data generated by the k-th Gaussian distribution model in the N observed data, and is pi k The k-th class distribution probability can also be regarded as the weight coefficient of each Gaussian distribution; mu.s k Is the mean, σ, of the kth class k Standard deviation for the kth class;
step six: and detecting the traffic jam of the road sections between the adjacent toll stations of the highway by using the clustering result.
2. The method according to claim 1, wherein the calculation formula of step 1) of step one is:
T＝t out -t in
where T is the travel time of the vehicle on the section between the entrance and exit toll stations, T out Time of vehicle passing through exit toll station, t in The time when the vehicle passes through the entrance toll station;
the calculation formula of the step 2) in the step one is as follows:
T c ＝t out -t c
in the formula, T c For the travel time, t, of the vehicle on the route between the gate and the toll station out Time of vehicle passing through exit toll station, t c The time for the vehicle to pass the gate.
3. The method of claim 1, wherein the target function of step five is expressed as
Figure FDA0003879644970000031
In the formula, n k Represents the number of data generated by the k-th Gaussian distribution model in the N observed data, pi k The k-th class distribution probability can also be regarded as the weight coefficient of each Gaussian distribution; mu.s k Is the mean, σ, of the kth class k Is the standard deviation of the kth class.
4. The method according to claim 1, wherein the probability γ in the step five jk Is calculated by the formula
Figure FDA0003879644970000032
In the formula, pi k The k-th class distribution probability can also be regarded as the weight coefficient of each Gaussian distribution; mu.s k Mean, σ, of the kth class k Is the standard deviation of the kth class.
5. The method of claim 1, wherein the updating cluster model parameters in step five is performed according to the following formula:
Figure FDA0003879644970000033
Figure FDA0003879644970000034
Figure FDA0003879644970000035
in the formula, N is the number of observed data,
Figure FDA0003879644970000036
is pi k ,μ k ,σ k The updated parameters of the clustering model are used,
Figure FDA0003879644970000041
are all the parameter theta of the next iteration (i+1) ", where θ (i+1) Is Q (theta ) of an objective function (i+1) ) Parameter value of (1), Q (theta ) (i+1) ) Is the value of the objective function at the (i + 1) th iteration; and K is the number of the clustering categories.
CN202010537175.1A 2020-06-12 2020-06-12 Highway traffic jam detection method based on charging data and checkpoint data Active CN111724592B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010537175.1A CN111724592B (en) 2020-06-12 2020-06-12 Highway traffic jam detection method based on charging data and checkpoint data

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010537175.1A CN111724592B (en) 2020-06-12 2020-06-12 Highway traffic jam detection method based on charging data and checkpoint data

Publications (2)

Publication Number Publication Date
CN111724592A CN111724592A (en) 2020-09-29
CN111724592B true CN111724592B (en) 2022-11-25

Family

ID=72566567

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010537175.1A Active CN111724592B (en) 2020-06-12 2020-06-12 Highway traffic jam detection method based on charging data and checkpoint data

Country Status (1)

Country Link
CN (1) CN111724592B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112419721B (en) * 2020-11-18 2021-06-08 交通运输部科学研究院 Method and device for calculating key indexes of transportation of vehicles on highway and electronic equipment
CN112767684A (en) * 2020-12-22 2021-05-07 重庆大学 Highway traffic jam detection method based on charging data
CN113674538B (en) * 2021-08-09 2023-04-18 南京领航交通科技有限公司 Section flow monitoring system

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN102768804B (en) * 2012-07-30 2014-03-26 江苏物联网研究发展中心 Video-based traffic information acquisition method
CN105608431A (en) * 2015-12-22 2016-05-25 杭州中威电子股份有限公司 Vehicle number and traffic flow speed based highway congestion detection method
SG11201805249RA (en) * 2016-03-03 2018-09-27 Mitsubishi Electric Corp Congestion prediction device and congestion prediction method
CN107610464B (en) * 2017-08-11 2019-02-12 河海大学 A kind of trajectory predictions method based on Gaussian Mixture time series models

Also Published As

Publication number Publication date
CN111724592A (en) 2020-09-29

Similar Documents

Publication Publication Date Title
CN111724592B (en) Highway traffic jam detection method based on charging data and checkpoint data
CN105590346B (en) The traffic information collection of turn pike net and inducible system based on path identifying system
CN111508094B (en) Highway congestion finding method based on ETC portal frame and gate traffic data
CN109191861B (en) System and method for detecting abnormal behavior of fee evasion vehicle on expressway based on video detector
CN112767684A (en) Highway traffic jam detection method based on charging data
CN111768619A (en) Express way vehicle OD point determining method based on checkpoint data
CN102081846A (en) Expressway charge data track matching based traffic state recognition method
CN111581538B (en) Expressway charging data-based expressway traffic flow state deducing method
CN114023073B (en) Expressway congestion prediction method based on vehicle behavior analysis
CN116631186B (en) Expressway traffic accident risk assessment method and system based on dangerous driving event data
CN111882858B (en) Multi-source data-based method for predicting queuing length of highway abnormal event
CN115691120A (en) Congestion identification method and system based on highway running water data
CN106919953A (en) A kind of abnormal trip Stock discrimination method based on track traffic data analysis
CN108447265A (en) Road traffic accident stain section discrimination method based on TOPSIS methods
CN112906945A (en) Traffic flow prediction method, system and computer readable storage medium
CN115311858A (en) Urban road section grading control method based on traffic flow toughness
CN111915880A (en) Method for estimating effective traffic capacity of accident point of highway abnormal event
CN116257797A (en) Single trip track identification method of motor vehicle based on Gaussian mixture model
Loubesi et al. Road trafficking description and short term travel time forecasting, with a classification method
CN113345252A (en) Short-time prediction method and device for lower-channel flow of toll station
CN113723459A (en) Bus operation time interval division and one-way point recommendation method based on time aggregation
CN114023085A (en) Intersection signal timing parameter inference method based on bayonet detection data
CN113723761A (en) Multi-dimensional urban bus operation service reliability evaluation method based on operation data
CN111724587A (en) Traffic flow density parameter extraction method based on RFID technology
CN116403432B (en) Smart city roadside parking space supervision system and method

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
GR01 Patent grant
GR01 Patent grant