CN111260922A - Ramp control method based on congestion situation classification - Google Patents
Ramp control method based on congestion situation classification Download PDFInfo
- Publication number
- CN111260922A CN111260922A CN202010064336.XA CN202010064336A CN111260922A CN 111260922 A CN111260922 A CN 111260922A CN 202010064336 A CN202010064336 A CN 202010064336A CN 111260922 A CN111260922 A CN 111260922A
- Authority
- CN
- China
- Prior art keywords
- ramp
- data
- control
- congestion
- congestion situation
- 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.)
- Granted
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
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/09—Arrangements for giving variable traffic instructions
- G08G1/096—Arrangements for giving variable traffic instructions provided with indicators in which a mark progresses showing the time elapsed, e.g. of green phase
Landscapes
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Chemical & Material Sciences (AREA)
- Analytical Chemistry (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention relates to a ramp control method based on congestion situation grading, which is characterized in that the congestion situation is divided again by utilizing Gaussian mixture model clustering through analyzing historical data, because the critical occupancy is difficult to obtain usually and the performance of an ALINEA algorithm is directly influenced due to inaccurate occupancy. The method takes the traffic flow as a control parameter of an ALINEA algorithm, adaptively selects the control rate according to the congestion situation, adopts a sectional control method to increase the queuing length constraint and transfers a control target to the ramp when the queuing is too long in order to avoid the influence of overflow generated by the overlong queuing of the ramp on ground traffic. The SUMO simulation software is used for carrying out simulation experiments, and results show that the method has different degrees of optimization on ramp queuing and waiting time of vehicles under the condition of ensuring main line traffic, and improves road traffic efficiency.
Description
Technical Field
The invention relates to the technical field of intelligent traffic, in particular to a ramp control method based on congestion situation grading.
Background
With the rapid development of economy, many cities construct urban expressways to meet the increasing traffic demands, but the quantity of vehicles kept increases year by year, and the problem of traffic jam becomes more and more serious. At present, the main control methods capable of effectively relieving the traffic jam problem of urban expressway include ramp control, variable speed limit and traffic guidance. The ramp control is the most widely applied and effective express way traffic control means, and the traffic flow which is converged into the main line is controlled by the signal lamps arranged on the ramp, so that the purpose of improving the traffic capacity of the main line is achieved. The method for single-point control of the entrance ramp mainly comprises the following steps: demand-capacity control, occupancy control, ALINEA, ZONE, fuzzy control, neural network control, reinforcement learning.
Although ALINEA is very effective in treating single-ramp congestion, the control method comprises ALINEA and a plurality of expansion algorithms thereof, wherein the control target is focused on the control of a main line, the problem that ramps are easy to overflow is ignored, the selection of the critical occupancy value of ALINEA has a great influence on the control, and if the critical occupancy value is unreasonable to set or cannot be obtained, the control effect is greatly reduced. In the aspect of congestion situation grading, the clustering is performed by using k-means, and optimization is performed on the aspects of selecting a central point and the number of clusters, reducing the iteration times of an algorithm or increasing the clustering accuracy and the like on the basis, but the k-means is a hard clustering method, each object is allocated to the nearest cluster by calculating the distance between the object and the central point of the cluster, and for some situations which are difficult to distinguish, the probability of cluster allocation is lacked, so that the distinguishing is difficult. Secondly, the boundary of the k-means cluster is circular, and the effect is not good when some non-circular data are fitted.
Disclosure of Invention
The invention aims to overcome the defects and provides a ramp control method based on congestion situation grading. The method takes the traffic flow as a control parameter of an ALINEA algorithm, adaptively selects the control rate according to the congestion situation, adopts a sectional control method to increase the queuing length constraint and transfers a control target to the ramp when the queuing is too long in order to avoid the influence of overflow generated by the overlong queuing of the ramp on ground traffic. Under the condition of ensuring the main line to pass, the invention optimizes the queuing of the ramps and the waiting time of the vehicles to different degrees, thereby improving the road passing efficiency.
The invention achieves the aim through the following technical scheme: a ramp control method based on congestion situation grading comprises the following steps:
(1) collecting microwave data of a microwave detector on a main line of the urban expressway and gate data detected by a ramp gate electric alarm, and counting the number of passing vehicles at the gates;
(2) preprocessing the acquired microwave data and the acquired checkpoint data, correcting abnormal data by adopting a historical mean value in a period, eliminating missing, repeating and abnormal conditions by adopting a license plate number matching method, and realizing detection and correction of the abnormal data;
(3) based on the data obtained in the step (2), clustering by using a Gaussian mixture clustering algorithm, and re-dividing the congestion situation of the main line of the express way by the vehicle speed according to a clustering result;
(4) performing ramp signal control by utilizing an ALINEA algorithm;
(5) and (4) increasing the restriction of the queuing length of the ramp on the basis of the step (4), setting a threshold value for the queuing length of the vehicles, and adopting a sectional control to adjust a long time timing scheme of a green light.
Preferably, in the step (1), in order to keep the time collection granularity of the gate data and the microwave data the same, the number of passing cars is counted according to the number of the license plate on the gate electric alarm record at regular time, so as to obtain the number of passing cars of the gate.
Preferably, in the step (2), microwave data are given according to lanes and need to be integrated, the data of each lane are complete, the flow is taken and the sum is obtained, and the speed is averaged; if only 1 lane has data at a certain moment, taking the product of the lane and the number of lanes as the upstream flow, and taking the measured value of the lane as the speed; if the speed flow data does not exist at a certain moment, replacing the speed flow data with the data at the previous moment; in the data of the card port, repeated data appearing within 5 minutes are removed, namely the condition that the same license plate number appears in a plurality of records is removed.
Preferably, in the step (3), the number K of the cluster clusters is determined, then gaussian mixture clustering is performed, and the current parameter pi is used by using the EM algorithmk、μk、ΣkCalculating posterior probability, and continuously iterating until the most appropriate parameter is obtained; dividing the microwave data set D into k congestion situations through Gaussian mixture clustering, wherein the congestion situations are divided according to posterior probability gammajkDetermining cluster labels of each sample
And finally, dividing the congestion situation into four levels by taking the speed as a dividing condition: unobstructed (v is more than or equal to v)1) Slight congestion [ v ]2,v1) Moderate congestion [ v ]3,v2) Severe congestion (v < v)3)。
Preferably, the gaussian mixture clustering algorithm specifically comprises: gaussian mixture clustering uses a probability model to represent a clustering prototype based on a gaussian mixture model, and the probability density function of the gaussian mixture distribution is shown as follows:
the distribution consists of K mixture components, each corresponding to a Gaussian distribution, N (x | μ)k,Σk) Is a probability density function of a Gaussian distribution in whichkAs mean vector, sigmakIs a covariance matrix; wherein the microwave data set D ═ { x ═ is used1,x2,…,xmIn which xmIs a two-dimensional data; pikIs a mixing coefficient, has a value of more than 0, and
when a clustering algorithm is used, each Gaussian distribution can be understood as a congestion situation, pikThe probability that the k-th situation is selected can be regarded as the probability; let zjE {1,2, … K } represents the input traffic data xjOf a mixture of Gaussian components, zjIs a priori probability P (z)jK) pairsShould pikAccording to Bayes' theorem zjThe posterior probability of (2) is abbreviated as gammajkAs shown in the following formula:
the microwave data set D is divided into k congestion situations by Gaussian mixture clustering, the division of the congestion situations is determined by posterior probability, and the cluster mark lambda of each sample is calculated by formula 3jAs shown in the following formula:
the Gaussian mixture model calculates posterior probability by using the current parameters by using an EM algorithm, updates model parameters according to the following formula, and continuously iterates until the most appropriate parameters are obtained;
preferably, the step (4) is specifically as follows:
(i) replacing the occupancy in the ALINEA algorithm by the main line flow of the express way as control input;
(ii) based on the congestion situation obtained in the step (3), the congestion situation divided by the speed interval is used as a selection standard of the control rate in the ALIENA algorithm, and the control rate is selected in a self-adaptive mode according to the congestion situation;
(iii) and corresponding control rates are formulated according to different congestion situations.
Preferably, as for the result of the congestion situation, the signal timing of the ramp is assumed to be corresponding to the control rate, as shown in the following formula:
wherein r (k) represents the turn-regulation rate of the k-th cycle; kFIs the control gain; r ismaxThe maximum regulation rate of the ramp signal lamp control; r isminIs the minimum regulation rate of the ramp signal lamp control;
representing the expected saturation flow of the main line of the express way;
obtaining the average speed of the upstream vehicle in real time in the k-1 th period;
is the flow from the upstream and the ramp into the downstream in the k-1 th cycle.
Preferably, the step (5) is specifically as follows:
(I) counting the number A (k) of vehicles arriving on a ramp in a control period k according to the electric alarm data of a bayonet, calculating the number D (k) of vehicles departing from the ramp according to the regulation rate of the previous period, and obtaining the accumulated number Q (k) of vehicles staying in the regulation period of k:
Q(k)＝Q(k-1)+(A(k)-D(k))
(II) calculating the current ramp queuing length L' according to Q (k):
wherein, lambda is the number of ramp lanes; mu is the distance between the vehicle heads of the queuing vehicles on the ramp; considering that Q (k) may be a negative value in the case of queue clearing, setting δ to be 0 when Q (k) is a negative value, otherwise δ is 1;
(III) assuming that the total length of the ramp is L, setting a threshold value for the queuing length of the vehicles; setting critical queuing length L by sectional control1Maximum queuing length L2(ii) a When the queue length exceeds L1When in use, the green time needs to be properly prolonged with priority,and adjusting the time timing scheme of the green light to enable the main line to release more vehicles.
Preferably, the step (III) further comprises: when the vehicle queue length is in (L)1，L2]When the number of vehicles allowed to be put into the downstream of the next period is calculated, the queuing length of the ramp is increased to exceed L1Part of the vehicles, the control rate is shown as follows:
when the queuing length exceeds L2When the ramp queues to be close to the overflow state, the influence on the ground road is about to occur, and the r is directly used at the momentmaxControlling; the improved ALINEA control rate is obtained by integration when signal control is combined with a ramp queuing length constraint state:
the invention has the beneficial effects that: according to the method, through analysis of historical data, the congestion situation is divided again by using Gaussian mixture model clustering, because the critical occupancy is usually difficult to obtain, and the inaccurate occupancy can directly influence the performance of the ALINEA algorithm; the method takes the traffic flow as a control parameter of an ALINEA algorithm, adaptively selects a control rate according to the congestion situation, adopts a sectional control method to increase the queuing length constraint and transfers a control target to a ramp when the queuing is too long in order to avoid the influence of overflow generated by the too long queuing of the ramp on ground traffic; under the condition of ensuring the main line to pass, the invention optimizes the queuing of the ramps and the waiting time of the vehicles to different degrees, thereby improving the road passing efficiency.
Drawings
FIG. 1 is a schematic flow diagram of the process of the present invention;
fig. 2 is a schematic view of a light strip based on congestion situation classification according to an embodiment of the present invention;
fig. 3 is a schematic view of ramp control according to an embodiment of the present invention.
Detailed Description
The invention will be further described with reference to specific examples, but the scope of the invention is not limited thereto:
example (b): as shown in fig. 1, a ramp control method based on congestion situation classification includes the following steps:
(1) and collecting microwave data of a microwave detector on a main line of the urban expressway and gate data detected by a ramp gate electric alarm, and counting the number of passing cars according to the number of the license plate by regularly recording the gate electric alarm to obtain the number of passing cars of the gate in order to keep the time collection granularity of the gate data and the microwave data the same.
The flow rate through the section per unit time is obtained in pcu/5min from a microwave detector installed on the main line of the urban expressway, and the average speed through the vehicle is in km/h. According to the method, the traffic flow entering the ramp in unit time is counted by a bayonet electric alarm installed at the entrance of the ramp, wherein the bayonet data mainly comprise license plate numbers, license plate types, vehicle brands, intersection names and passing time, and in order to ensure that the time granularity of the bayonet electric alarm time data is consistent with that of microwave data, the number of vehicles entering the ramp is counted every 5 minutes, and the unit is pcu/5 min.
(2) Preprocessing the acquired microwave data and the acquired checkpoint data, correcting abnormal data by adopting a historical mean value in a period, eliminating missing, repeating and abnormal conditions by adopting a license plate number matching method, and realizing detection and correction of the abnormal data;
microwave data is given according to lanes and needs to be integrated, the flow is measured and summed when the data of each lane is complete, and the speed is averaged; if only 1 lane has data at a certain moment, taking the product of the lane and the number of lanes as the upstream flow, and taking the value measured by the lane as the speed; the data at the previous time is substituted for the data at the previous time when there is no speed flow data at a certain time. The data of the card port is mainly removed aiming at repeated data appearing in 5 minutes, namely the condition that the same license plate number appears in a plurality of records.
(3) Based on the data obtained in the step (2), clustering by using a Gaussian mixture clustering algorithm, and re-dividing the congestion situation of the main line of the express way by the vehicle speed according to a clustering result;
the congestion situation of the urban expressway main line is divided into six grades (speed km/h) according to the public safety industry standard of the people's republic of China: very unobstructed (v)>65) Clear [50,65), light congestion [35,50), moderate congestion [20,35), heavy congestion [5,20), severe congestion [0, 5). The national standard has strong universality, but the influences of other factors such as the number of lanes in an actual road section, the lane width and the like are considered, and the congestion situation grades of different express ways are obtained by analyzing according to actual conditions. In actual control, when the express way is in heavy congestion, the ramp does not allow more vehicles to continuously merge, and the ramp should be closed without ramp control. When the traffic signal is in a smooth and very smooth state, the main line traffic state is good, and the very smooth state can be regarded as a special smooth state that the signal timing is in a green-full release state. Therefore, the invention divides the congestion situation of the urban expressway into four levels according to the speed: unobstructed (v is more than or equal to v)1) Slight congestion [ v ]2,v1) Moderate congestion [ v ]3,v2) Severe congestion (v < v)3)。
The Gaussian Mixture clustering is based on a Gaussian Mixture Model (Gaussian Mixture Model), and is different from a clustering method for calculating a distance from a central point, such as K-menas, etc., the Gaussian Mixture clustering uses a probability Model to represent a clustering prototype, and formula 1 is a probability density function of Gaussian Mixture distribution.
The distribution consists of K mixture components, each corresponding to a Gaussian distribution, N (x | μ)k,Σk) Is a probability density function of a Gaussian distribution in whichkAs mean vector, sigmakFor covariance matrix, the microwave data set D ═ { x used in the invention1,x2,…,xmIn which xmThe data is two-dimensional data and is derived from speed and flow data obtained by a fast path microwave detector. PikIs a mixing coefficient, has a value of more than 0, and
when the clustering algorithm is used, each Gaussian distribution can be understood as a congestion situation, pikCan be considered as the probability that the k-th situation is selected. Let zjE {1,2, … K } represents the input traffic data xjOf a mixture of Gaussian components, zjIs a priori probability P (z)jK) corresponds to pikAccording to Bayes' theorem zjThe posterior probability of (2) is abbreviated as gammajkShown by formula 2.
The microwave data set D is divided into k congestion situations by Gaussian mixture clustering, the division of the congestion situations is determined by posterior probability, and the cluster mark lambda of each sample is calculated by formula 3j。
From the formula 1, it can be seen thatk、Σk、πkIt needs to be determined in advance. Maximum likelihood estimation is adopted, namely a group of parameters enabling the likelihood function to be maximum are found and are determined as the most appropriate parameters. The Gaussian mixture model utilizes an EM algorithm, firstly uses the current parameters to calculate posterior probability, then updates the model parameters according to formulas 4, 5 and 6, and continuously iterates until the most appropriate parameters are obtained.
And clustering and analyzing historical data of adjacent road sections to divide congestion situations, drawing the congestion situations on the same time axis, and replacing time intervals of different congestion intervals with different colors to obtain a lamp band diagram shown in fig. 2. The time, the place, the spreading condition of the traffic jam on the express way and the jam dissipation process can be reflected from the lamp band diagram, the local jam problem with small range and light degree can be solved by single-point control, the chain reaction caused by the jam generated by a plurality of ramps simultaneously is correspondingly weak, and the drawing of the lamp band diagram plays a guiding role in the formulation of the multi-turn road coordination control strategy.
(4) Performing ramp signal control by utilizing an ALINEA algorithm; the method specifically comprises the following steps: replacing the occupancy in the ALINEA algorithm by the main line flow of the express way as control input; based on the congestion situation obtained in the step (3), the congestion situation divided by the speed interval is used as a selection standard of the control rate in the ALIENA algorithm, and the control rate is selected in a self-adaptive mode according to the congestion situation; and corresponding control rates are formulated according to different congestion situations.
In the present embodiment, the control of the expressway ramps in the city is shown in fig. 3. And judging the congestion situation of the main line of the express way according to the upstream vehicle speed of the ramp measured in each control period, and aiming at the analysis result of the congestion situation, adopting a corresponding control rate in signal timing of the ramp, as shown in a formula 7.
In the formula: r (k) represents the turn-regulation rate of the k-th cycle; kFIs the control gain; r ismaxThe maximum regulation rate of the ramp signal lamp control; r isminIs the minimum regulation rate of the ramp signal lamp control;
representing the expected saturation flow of the main line of the express way;
obtaining the average speed of the upstream vehicle in real time in the k-1 th period;
is the flow from the upstream and the ramp into the downstream in the k-1 th cycle.
The upstream flow of the ramp and the flow of the ramp in the previous period are used as input, the flow of the afflux downstream in the previous period is obtained by calculation, and the sum of the calculated flow and the saturated flow qactCalculating the difference to obtain the flow rate still held at the downstream of the next period, calculating the difference between the two in the early morning or peak period to obtain a larger difference value, and K is used for avoiding the large fluctuation of the regulation rateFValues less than 1 should be taken. Saturated flow qactAccording to historical data analysis, the capacity of a road section can be lower in consideration of the influence of interference factors such as weather and traffic accidents, secondly, because the upstream flow is used as an input, congestion is usually generated from the downstream and spreads to the upstream, and the expected saturated flow is obtained
Should be slightly less than the actual saturation flow qactThereby, control can be performed in advance. Through multiple tests, the test proves that when
KFWhen the value is 0.1, a better control effect can be realized.
When the expressway mainline is in a smooth state, the downstream flow does not reach saturation, the regulation rate rises, the green light phase is continuously prolonged, the control rate can know that the regulation rate can continuously rise under the condition of less traffic flow in the early morning or late night without setting an upper bound, and r is setmaxLimiting the increase in modulation rate. The condition that the maximum value of the adjusting rate appears is that the ramp signal lamp is completely green and is released, and the situation is equivalent to no control.
When the main line of the express way is in a light congestion state, the control rate should be reduced, and the flow rate of the ramp merging into the main line is limited. Due to the normal distribution of flow, as the velocity continues to decrease, the flow gradually decreases after increasing to the saturation flow. Q is known from the control rateactAnd
the difference of (a) is still a positive number, if the control rate in the unobstructed state is continued,the adjustment rate value is still continuously rising, contrary to the control target at this time. Therefore, when the regulation rate is calculated, the flow rate which can still be accommodated in the next downstream period when the main line is in the open state does not represent the flow rate which is still allowed to be put in the ramp in the congestion state, and the flow rate should be subtracted to reduce the regulation rate. If the vehicle is released by adopting a small adjustment rate all the time, the time that the vehicle stays on the ramp is too long, the emotion of a driver is negatively influenced, and safety accidents are easily caused. In order to avoid potential safety hazards. In order to prevent the situation that the regulation rate is continuously reduced to 0 due to congestion along with the arrival of the peak period, r is setminA restriction is made. In the case of moderate congestion, the congestion degree is deepened, and the regulation rate is directly set to rmin. When the main line enters a heavy congestion state, the ramp can be directly closed, and signal control is not needed at the moment, namely the regulation rate is equal to 0.
(5) And (4) increasing the restriction of the queuing length of the ramp on the basis of the step (4), setting a threshold value for the queuing length of the vehicles, and adopting a sectional control to adjust a long time timing scheme of a green light.
In the embodiment, the number of arriving and exiting vehicles in each period is calculated by using the bayonet passing data and the regulation rate, and the accumulated queuing length is obtained by combining the method. The number of vehicles arriving on the ramp A (k) in the set regulation period k can be obtained by analyzing the electric alarm data of the gate, the number of vehicles leaving D (k) is obtained by calculating the regulation rate in the previous period, and then the cumulative number of vehicles staying in the regulation period k is as follows:
Q(k)＝Q(k-1)+(A(k)-D(k)) (8)
the cumulative queue length L' is:
in the formula: lambda is the number of ramp lanes; mu is the distance between the vehicle heads of the queuing vehicles on the ramp; considering that q (k) may be negative in the case of queue clearing, it is set that δ takes 0 when q (k) is negative, and δ takes 1 otherwise.
Assuming that the total ramp length is L, a threshold value is set for the queuing length of the vehicles. Setting the critical queuing length by sectional controlL1Maximum queuing length L2. The queuing length exceeds L1In time, it is necessary to give priority to properly extending the green time so that the main line allows more vehicles to pass. When the vehicle queue length is in (L)1，L2]When the number of vehicles allowed to be put into the downstream of the next period is calculated, the queuing length of the ramp is increased to exceed L1The control rate of some vehicles is shown in equation 10.
When the queuing length exceeds L2When the ramp queues to be close to the overflow state, the influence on the ground road is about to occur, and the r is directly used at the momentmaxAnd (5) controlling. The improved ALINEA control rate is obtained by integration when signal control is combined with a ramp queuing length constraint state:
the SUMO simulation software is used for carrying out simulation experiments, and results show that the method has different degrees of optimization on ramp queuing and waiting time of vehicles under the condition of ensuring main line traffic, and improves road traffic efficiency.
While the invention has been described in connection with specific embodiments thereof, it will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (9)
1. A ramp control method based on congestion situation grading is characterized by comprising the following steps:
(1) collecting microwave data of a microwave detector on a main line of the urban expressway and gate data detected by a ramp gate electric alarm, and counting the number of passing vehicles at the gates;
(2) preprocessing the acquired microwave data and the acquired checkpoint data, correcting abnormal data by adopting a historical mean value in a period, eliminating missing, repeating and abnormal conditions by adopting a license plate number matching method, and realizing detection and correction of the abnormal data;
(3) based on the data obtained in the step (2), clustering by using a Gaussian mixture clustering algorithm, and re-dividing the congestion situation of the main line of the express way by the vehicle speed according to a clustering result;
(4) performing ramp signal control by utilizing an ALINEA algorithm;
(5) and (4) increasing the restriction of the queuing length of the ramp on the basis of the step (4), setting a threshold value for the queuing length of the vehicles, and adopting a sectional control to adjust a long time timing scheme of a green light.
2. The method for controlling a ramp based on congestion situation classification as claimed in claim 1, wherein: in the step (1), in order to keep the time collection granularity of the bayonet data and the microwave data to be the same, the number of passing cars is counted according to the number of the license plate on the bayonet electric alarm record at regular time, and the number of passing cars of the bayonet is obtained.
3. The method for controlling a ramp based on congestion situation classification as claimed in claim 1, wherein: in the step (2), microwave data are given according to lanes and need to be integrated, the flow is measured and summed when the data of each lane are complete, and the speed is averaged; if only 1 lane has data at a certain moment, taking the product of the lane and the number of lanes as the upstream flow, and taking the measured value of the lane as the speed; if the speed flow data does not exist at a certain moment, replacing the speed flow data with the data at the previous moment; in the data of the card port, repeated data appearing within 5 minutes are removed, namely the condition that the same license plate number appears in a plurality of records is removed.
4. The method for controlling a ramp based on congestion situation classification as claimed in claim 1, wherein: in the step (3), firstly, the number K of the clustering clusters is determined, then Gaussian mixed clustering is carried out, and the current parameter pi is used by utilizing an EM algorithmk、μk、ΣkCalculating posterior probability, and continuously iterating until the most appropriate parameter is obtained; tong (Chinese character of 'tong')The microwave data set D is divided into k congestion situations by Gaussian mixture clustering, and the congestion situations are divided according to the posterior probability gammajkDetermining cluster labels of each sample
And finally, dividing the congestion situation into four levels by taking the speed as a dividing condition: unobstructed (v is more than or equal to v)1) Slight congestion [ v ]2,v1) Moderate congestion [ v ]3,v2) Severe congestion (v < v)3)。
5. The method for controlling a ramp based on congestion situation classification as claimed in claim 4, wherein: the Gaussian mixture clustering algorithm specifically comprises the following steps: gaussian mixture clustering uses a probability model to represent a clustering prototype based on a gaussian mixture model, and the probability density function of the gaussian mixture distribution is shown in the following formula (1):
the distribution consists of K mixture components, each corresponding to a Gaussian distribution, N (x | μ)k,Σk) Is a probability density function of a Gaussian distribution in whichkAs mean vector, sigmakIs a covariance matrix; wherein the microwave data set D ═ { x ═ is used1,x2,…,xmIn which xmIs a two-dimensional data; pikIs a mixing coefficient, has a value of more than 0, and
when a clustering algorithm is used, each Gaussian distribution can be understood as a congestion situation, pikThe probability that the k-th situation is selected can be regarded as the probability; let zjE {1,2, … K } represents the input traffic data xjOf a mixture of Gaussian components, zjIs a priori probability P (z)jK) corresponds to pikAccording to Bayes' theorem zjThe posterior probability of (2) is abbreviated as gammajkAs followsRepresented by formula (2):
the microwave data set D is divided into k congestion situations by Gaussian mixture clustering, the division of the congestion situations is determined by posterior probability, and the cluster mark lambda of each sample is calculated by formula 3jAs shown in the following formula (3):
the Gaussian mixture model calculates posterior probability by using the current parameters by using an EM algorithm, updates model parameters according to the formulas (4), (5) and (6), and continuously iterates until the most appropriate parameters are obtained;
6. the method for controlling a ramp based on congestion situation classification as claimed in claim 1, wherein: the step (4) is specifically as follows:
(i) replacing the occupancy in the ALINEA algorithm by the main line flow of the express way as control input;
(ii) based on the congestion situation obtained in the step (3), the congestion situation divided by the speed interval is used as a selection standard of the control rate in the ALIENA algorithm, and the control rate is selected in a self-adaptive mode according to the congestion situation;
(iii) and corresponding control rates are formulated according to different congestion situations.
7. The method for controlling a ramp based on congestion situation classification as claimed in claim 6, wherein: aiming at the congestion situation result, the signal timing of the ramp adopts a corresponding control rate, as shown in the following formula (7):
wherein r (k) represents the turn-regulation rate of the k-th cycle; kFIs the control gain; r ismaxThe maximum regulation rate of the ramp signal lamp control; r isminIs the minimum regulation rate of the ramp signal lamp control;
representing the expected saturation flow of the main line of the express way;
obtaining the average speed of the upstream vehicle in real time in the k-1 th period;
is the flow from the upstream and the ramp into the downstream in the k-1 th cycle.
8. The method for controlling a ramp based on congestion situation classification as claimed in claim 1, wherein: the step (5) is specifically as follows:
(I) counting the number A (k) of vehicles arriving on a ramp in a control period k according to the electric alarm data of a bayonet, calculating the number D (k) of vehicles departing from the ramp according to the regulation rate of the previous period, and obtaining the accumulated number Q (k) of vehicles staying in the regulation period of k:
Q(k)＝Q(k-1)+(A(k)-D(k)) (8)
(II) calculating the current ramp queuing length L' according to Q (k):
wherein λ isNumber of ramp lanes; mu is the distance between the vehicle heads of the queuing vehicles on the ramp; considering that Q (k) may be a negative value in the case of queue clearing, setting δ to be 0 when Q (k) is a negative value, otherwise δ is 1; (III) assuming that the total length of the ramp is L, setting a threshold value for the queuing length of the vehicles; setting critical queuing length L by sectional control1Maximum queuing length L2(ii) a When the queue length exceeds L1In time, it is necessary to give priority to properly extending the green time and adjusting the green time timing scheme, so that more vehicles are released by the main line.
9. The method for controlling a ramp based on congestion situation classification as claimed in claim 8, wherein: the step (III) further comprises: when the vehicle queue length is in (L)1，L2]When the number of vehicles allowed to be put into the downstream of the next period is calculated, the queuing length of the ramp is increased to exceed L1In some vehicles, the control rate is as shown in equation (10):
when the queuing length exceeds L2When the ramp queues to be close to the overflow state, the influence on the ground road is about to occur, and the r is directly used at the momentmaxControlling; the improved ALINEA control rate is obtained by integration when signal control is combined with a ramp queuing length constraint state:
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010064336.XA CN111260922B (en) | 2020-01-20 | 2020-01-20 | Ramp control method based on congestion situation classification |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010064336.XA CN111260922B (en) | 2020-01-20 | 2020-01-20 | Ramp control method based on congestion situation classification |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111260922A true CN111260922A (en) | 2020-06-09 |
| CN111260922B CN111260922B (en) | 2021-01-29 |
Family
ID=70950857
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010064336.XA Active CN111260922B (en) | 2020-01-20 | 2020-01-20 | Ramp control method based on congestion situation classification |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111260922B (en) |
Cited By (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111951555A (en) * | 2020-08-20 | 2020-11-17 | 重庆大学 | Traffic flow speed critical phase change interval estimation method based on RFID data |
| CN112447051A (en) * | 2020-11-16 | 2021-03-05 | 公安部交通管理科学研究所 | Multi-mode cooperative control method and system for frequent congestion point segments of express way |
| CN113257013A (en) * | 2021-07-16 | 2021-08-13 | 深圳市安达信通讯设备有限公司 | Traffic control system for ramp |
| CN114023084A (en) * | 2021-10-29 | 2022-02-08 | 青岛海信网络科技股份有限公司 | Traffic jam processing method, device, equipment and medium |
| CN114329342A (en) * | 2021-12-20 | 2022-04-12 | 中交第二公路勘察设计研究院有限公司 | Method for calculating traffic capacity of expressway based on extreme value-Gaussian mixed distribution model |
| CN114387778A (en) * | 2020-10-22 | 2022-04-22 | 济南市公安局交通警察支队 | Urban expressway congestion reason analysis method |
| CN114882712A (en) * | 2021-09-10 | 2022-08-09 | 无锡华通智能交通技术开发有限公司 | Multi-turn-road control express way closure method based on lane space occupancy |
| CN114882696A (en) * | 2020-10-28 | 2022-08-09 | 华为技术有限公司 | Method and device for determining road capacity and storage medium |
| CN115035734A (en) * | 2022-05-18 | 2022-09-09 | 北京航空航天大学 | Urban expressway entrance ramp signal control method based on congestion state classification |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108122418A (en) * | 2017-12-19 | 2018-06-05 | 公安部交通管理科学研究所 | City expressway main line and ring road cooperative control method and system |
| CN108269401A (en) * | 2018-01-30 | 2018-07-10 | 银江股份有限公司 | A kind of overpass traffic congestion Forecasting Methodology based on data-driven |
| CN108648469A (en) * | 2018-06-13 | 2018-10-12 | 西华大学 | A kind of coordinated ramp metering method based on more bottleneck conditions |
| CN110503833A (en) * | 2019-08-29 | 2019-11-26 | 桂林电子科技大学 | A kind of Entrance ramp inter-linked controlling method based on depth residual error network model |
-
2020
- 2020-01-20 CN CN202010064336.XA patent/CN111260922B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN108122418A (en) * | 2017-12-19 | 2018-06-05 | 公安部交通管理科学研究所 | City expressway main line and ring road cooperative control method and system |
| CN108269401A (en) * | 2018-01-30 | 2018-07-10 | 银江股份有限公司 | A kind of overpass traffic congestion Forecasting Methodology based on data-driven |
| CN108648469A (en) * | 2018-06-13 | 2018-10-12 | 西华大学 | A kind of coordinated ramp metering method based on more bottleneck conditions |
| CN110503833A (en) * | 2019-08-29 | 2019-11-26 | 桂林电子科技大学 | A kind of Entrance ramp inter-linked controlling method based on depth residual error network model |
Non-Patent Citations (3)
| Title |
|---|
| 乔彦甫 等: "基于ALINEA算法的城市快速路", 《西南交通大学学报》 * |
| 刘家锋 等: "《模式识别》", 31 August 2014, 哈尔滨工业大学出版社 * |
| 沈国江 等: "基于排队长度的单点自适应信号控制方法", 《第十一届中国智能交通年会大会论文集》 * |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111951555A (en) * | 2020-08-20 | 2020-11-17 | 重庆大学 | Traffic flow speed critical phase change interval estimation method based on RFID data |
| CN111951555B (en) * | 2020-08-20 | 2022-04-22 | 重庆大学 | Traffic flow speed critical phase change interval estimation method based on RFID data |
| CN114387778B (en) * | 2020-10-22 | 2023-08-15 | 济南市公安局交通警察支队 | Urban expressway congestion cause analysis method |
| CN114387778A (en) * | 2020-10-22 | 2022-04-22 | 济南市公安局交通警察支队 | Urban expressway congestion reason analysis method |
| CN114882696A (en) * | 2020-10-28 | 2022-08-09 | 华为技术有限公司 | Method and device for determining road capacity and storage medium |
| CN114882696B (en) * | 2020-10-28 | 2023-11-03 | 华为技术有限公司 | Road capacity determination method, device and storage medium |
| CN112447051A (en) * | 2020-11-16 | 2021-03-05 | 公安部交通管理科学研究所 | Multi-mode cooperative control method and system for frequent congestion point segments of express way |
| CN113257013A (en) * | 2021-07-16 | 2021-08-13 | 深圳市安达信通讯设备有限公司 | Traffic control system for ramp |
| CN113257013B (en) * | 2021-07-16 | 2021-09-21 | 深圳市安达信通讯设备有限公司 | Traffic control system for ramp |
| CN114882712B (en) * | 2021-09-10 | 2023-12-29 | 无锡华通智能交通技术开发有限公司 | Multi-turn control expressway interception method based on lane space occupancy |
| CN114882712A (en) * | 2021-09-10 | 2022-08-09 | 无锡华通智能交通技术开发有限公司 | Multi-turn-road control express way closure method based on lane space occupancy |
| CN114023084A (en) * | 2021-10-29 | 2022-02-08 | 青岛海信网络科技股份有限公司 | Traffic jam processing method, device, equipment and medium |
| CN114023084B (en) * | 2021-10-29 | 2023-01-17 | 青岛海信网络科技股份有限公司 | Traffic jam processing method, device, equipment and medium |
| CN114329342A (en) * | 2021-12-20 | 2022-04-12 | 中交第二公路勘察设计研究院有限公司 | Method for calculating traffic capacity of expressway based on extreme value-Gaussian mixed distribution model |
| CN114329342B (en) * | 2021-12-20 | 2024-02-06 | 中交第二公路勘察设计研究院有限公司 | Expressway traffic capacity calculation method based on extremum-Gaussian mixture distribution model |
| CN115035734B (en) * | 2022-05-18 | 2023-05-23 | 北京航空航天大学 | Urban expressway entrance ramp signal control method based on congestion state classification |
| CN115035734A (en) * | 2022-05-18 | 2022-09-09 | 北京航空航天大学 | Urban expressway entrance ramp signal control method based on congestion state classification |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111260922B (en) | 2021-01-29 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111260922B (en) | Ramp control method based on congestion situation classification | |
| CN106355907B (en) | Signalized crossing queue length real-time estimation method based on track of vehicle | |
| CN112885088B (en) | Multi-turn road coordination control method based on dynamic traffic flow | |
| CN108510758B (en) | Urban expressway entrance ramp signal control method based on video real-time information | |
| CN113327416A (en) | Urban area traffic signal control method based on short-term traffic flow prediction | |
| CN111341095B (en) | Traffic signal control system and method based on edge side online calculation | |
| CN105551250B (en) | A kind of urban road intersection operating status method of discrimination based on interval clustering | |
| CN106803347B (en) | Urban intersection traffic state judging method based on RFID data | |
| CN115035734B (en) | Urban expressway entrance ramp signal control method based on congestion state classification | |
| CN111882858B (en) | Multi-source data-based method for predicting queuing length of highway abnormal event | |
| CN111583628B (en) | Road network heavy truck traffic flow prediction method based on data quality control | |
| CN109993981A (en) | The self tuning control method of traffic signals based on Holographic test | |
| CN113947929B (en) | Variable speed limit control method for highway reconstruction and extension continuous construction area | |
| CN115063990A (en) | Dynamic speed limit control method for bottleneck section of highway in mixed traffic flow environment | |
| CN106023610B (en) | A kind of bus and the green wave Synchronization of private car main line considering motorcade dispersion feature | |
| Liu et al. | A ramp metering method based on congestion status in the urban freeway | |
| Binglei et al. | Fuzzy-logic-based traffic incident detection algorithm for freeway | |
| CN113112789A (en) | Method for predicting and controlling influence of urban expressway emergency | |
| CN104182633B (en) | Hierarchical traffic operation evaluation method | |
| CN111524345A (en) | Induction control method for multi-objective optimization under constraint of real-time queuing length of vehicle | |
| CN111754790B (en) | Ramp entrance traffic control system and method based on radar | |
| Lin et al. | Adaptive signal control at isolated intersections | |
| CN114613125B (en) | Expressway multi-turn layered cooperative control method | |
| CN113593222B (en) | Multi-source data supported traffic control diagnosis method | |
| Fu et al. | An adaptive model for real-time estimation of overflow queues on congested arterials |
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 |