WO2022199442A1

WO2022199442A1 - Multi-intersection coordination adaptive control method based on checkpoint police system and video detector

Info

Publication number: WO2022199442A1
Application number: PCT/CN2022/081150
Authority: WO
Inventors: 付本刚; 柴畅
Original assignee: 江苏航天大为科技股份有限公司
Priority date: 2021-03-25
Filing date: 2022-03-16
Publication date: 2022-09-29
Also published as: CN113516854B; LU502405B1; CN113516854A

Abstract

A multi-intersection coordination adaptive control method based on a checkpoint police system and a video detector, comprising: first, observing the release situation of consecutive cycles of a coordinated intersection during off-peak and peak periods, respectively, analyzing release time information in each movement direction during the off-peak and peak periods, formulating a coordinated intersection release rule for corresponding periods to predict the release situation of future release cycles during the off-peak and peak periods, then formulating a vehicle data-based cancel rule to realize accurate statistics about the number of vehicles to provide reliable and credible data information for the movement directions, and finally, comparing with corresponding thresholds the actual vehicle queue length, the actual number of vehicles, and the actual time interval in the next phase obtained in real time, respectively, and using the cancel rule of release in movement directions to achieve reasonable cancel of release in movement directions, thereby maximizing the utilization rate of signal lights.

Description

A Multi-Intersection Coordinated Adaptive Control Method Based on Card Alarm and Video Detector

technical field

The invention relates to the technical field of intelligent traffic, in particular to a multi-intersection coordinated adaptive control method based on a card alarm and a video detector.

Background technique

Traffic signal light is an important display tool in traffic safety products. Its effective release can strengthen the management of vehicles at intersections, improve the utilization rate of road resources, reduce the waiting time of vehicles, and more importantly, stimulate the rapid development of the economy. In the early work, a vision-based traffic light detection algorithm was proposed in order to recognize the behavior of running a red light. Other signal light detection algorithms are based on different technologies, such as Maldives random field, neural network, mathematical morphology, etc., and more empirical models are introduced to improve the performance of the algorithm, but there is still a certain gap in the efficiency in achieving real-time performance.

At present, the video traffic detection technology has been applied to road traffic, but few technicians apply the big data technology of traffic jam police to actual intersections, especially the combination of traffic traffic detection technology and video traffic detection technology in multi-intersection coordination. Rare. The application of these technologies can provide reliable and effective real-time dynamic traffic flow data and improve the efficiency of coordinated intersections. In recent years, the optimization service engineering for road traffic release has gradually become an important field of research, and higher requirements have also been put forward for the precision of traffic flow data, real-time monitoring of traffic flow status, and prediction and release of selected schemes. Among them, the application of card police big data technology and video traffic detector technology has entered the field of vision of researchers, and how to maximize the utilization of these data and make effective global control of single point of intersection, especially multi-intersection coordination has been imminent.

SUMMARY OF THE INVENTION

In view of the above problems and technical requirements, the present inventor proposes a multi-intersection coordinated adaptive control method based on jamming and video detectors. Through the release rules and trend elimination rules, the prediction of the future release mode of the intersection is realized, and the movement release is realized. Reasonable elimination, to achieve the maximum utilization of the intersection release, thereby improving the efficiency of the main road release.

The technical scheme of the present invention is as follows:

A multi-intersection coordinated adaptive control method based on a card alarm and a video detector, comprising the following steps:

The release period of the coordinated intersection during the off-peak period and the peak period is selected as the reference period, and the release mode of each release period is the same;

Obtain the first historical dynamic information of the reference period through the card alarm and video detector, and execute the release rules according to the first historical dynamic information to respectively predict and coordinate the release mode of the future release period of the intersection during the off-peak period and the peak period;

Determine the current release mode of the key intersection according to the release rules, obtain the second historical dynamic information and real-time dynamic information of the key intersection through the card alarm and video detector, and execute the trend elimination rule according to the second historical dynamic information and real-time dynamic information, so as to update Release patterns at key intersections.

Its further technical solution is that the first historical dynamic information includes traffic flow information of each direction, and the release duration of each direction in each release cycle; according to the first historical dynamic information, the release rule is executed to predict and coordinate the release mode of the future release cycle during the peak period of the intersection, include:

In a release direction, the transition release time of two phases is set respectively according to the traffic flow information, denoted as _Tri , the release direction includes north-south direction and east-west direction, and the phase includes north-south left turn and north-south straight, east-west left turn and east-west straight;

Calculate the average release duration of each movement in the reference cycle according to the release duration of each movement in each release cycle;

Calculate the difference between the average release durations of the two movements relative to the release direction, denoted as u _i ;

Determine the comparison value of the two phases by comparing the magnitudes of _u _i and Tri;

According to the magnitudes of the two comparison values, the release mode of the future release cycle during the peak period of the coordinated intersection is determined.

Its further technical solution is to determine the comparison value of the two phases by comparing the magnitudes of _ui and _Tri , including:

Assuming that the release direction is the north-south direction, the corresponding phases are the north-south turn left and the north-south straight, and the comparison values of the two phases are recorded as M _L and M _S respectively. The comparison results are as follows:

Among them, u ₁ and u ₂ are the difference between the average release durations of south/north turn left and south/north straight, and u ₁ =NL _t -SL _t , NL _t is the average release duration of north left turn, and SL _t is south left Turn average release time, u ₂ =NS _t -SS _t , NS _t is the average release time of north going straight, SS _t is the average release time of south going straight; Tr ₁ is the transitional release time of north-south left turn, Tr ₂ is the transition of north-south straight release time.

Its further technical solution is to determine the release mode of the future release cycle of the coordinated intersection during the peak period according to the magnitude of the two comparison values, including:

If M _L × M _S >0, it includes two release modes, each of which contains three phases, including north-south straight release, left-turn and straight-through release in the main direction, and north-south left-turn release, where the main direction is south or North;

If M _L × M _S ≤ 0, the release mode contains two phases, including north-south straight release and north-south left-turn release.

Its further technical solution is to obtain the second historical dynamic information from several consecutive release cycles randomly selected, including the traffic flow information of each movement in each release cycle; obtain real-time dynamic information from the current release cycle, including the current release mode. The actual release time of each trend in the

If the movement elimination rules of each phase of the key intersection are the same, the movement elimination rules are executed according to the second historical dynamic information and real-time dynamic information, so as to update the release mode of the key intersection, including:

Set the maximum release duration and the minimum release duration of the first movement in the first phase according to the traffic flow information, denoted as T _max , T _min ;

Obtain the actual release duration of the first movement in real time and record it as T, and then compare it with the maximum release duration and the minimum release duration respectively. When T _min < T < T _max and the movement elimination rule is not satisfied, continue to release the first phase, and Re-execute the step of obtaining the actual release duration of the first movement in real time; when T _min < T < T _max and the movement elimination rule is satisfied, or, when T ≥ T _max , the first movement is eliminated and the first phase is set. In the steps of the maximum release duration and the minimum release duration of the next movement, all movements of the first phase are judged in turn. If all movements are eliminated, the first phase will be eliminated, and the movement elimination rule will be implemented for the next phase.

A further technical solution is that the second historical dynamic information further includes the number of vehicles in each release cycle at each stage, the movement release duration, and the interval between two adjacent vehicles in each movement, and the real-time dynamic information also includes each movement in the next phase. The total number of vehicles and the actual interval time T _n1 between two adjacent vehicles in the current phase;

Methods for judging whether the trend elimination rule is met include:

For each selected release cycle, divide the movement release duration of all phases of the release cycle into three phases, count the number of vehicles in each phase, and implement the vehicle number error elimination rule to accurately count the number of vehicles in each phase and each channel. number of vehicles;

Calculate the average number of vehicles at each stage of all selected release cycles, denoted as

Determine the number of channels n of each movement and the length l of the space occupied by each vehicle, then the threshold of the queue length of vehicles in each movement is:

Select the maximum value in the average value of the number of vehicles at each stage of the release cycle, divide it by the corresponding moving release duration, and the obtained quotient value is used as the threshold value of the number of vehicles, denoted as N _max ;

Calculate the average value of the interval time between two adjacent vehicles in each movement of all the selected release cycles as the interval time threshold between the two adjacent vehicles, denoted as T _n ;

According to the real-time acquisition of the total number of vehicles N' in each direction in the next phase, the actual vehicle queue length of the next phase is obtained as: L=N'×n×l;

Calculate the actual number of vehicles in each movement unit count period as:

Among them, T ₁ , T ₂ , and T ₃ are the corresponding movement release durations of each stage;

Compare the actual vehicle queuing length, actual number of vehicles and actual interval time obtained in real time with the corresponding thresholds to determine the elimination standard value, denoted as P, and the comparison results are as follows:

Among them, if P=1 indicates that the current movement is eliminated, if P=0 indicates that the current movement has not been eliminated, if at least one movement of the current phase has not been eliminated, the current phase has not been eliminated, and the current phase will continue to be released;

Condition [1] indicates that there are fewer vehicles in the current movement, and the current movement needs to be eliminated; otherwise, Condition [4] indicates that the current movement needs to continue to be released;

Condition [2] indicates that the distance between two adjacent vehicles in the current movement is large, and the traffic flow is sparse, and the current movement needs to be eliminated. On the contrary, condition [5] indicates that the current movement needs to continue to be released;

Condition [3] indicates that the moving traffic flow in the next phase is relatively large and there are fewer vehicles moving in the current phase, and the current movement needs to be eliminated. On the contrary, Condition [6] indicates that the moving traffic flow in the next phase is denser, and the current movement needs to be released. .

Its further technical solution is that the second historical dynamic information also includes the total time TA of the total vehicles passing through the stop line in each movement and the total number of vehicles N _A passing through the stop line in each movement in the previous release cycle, and the real _- time dynamic information also includes the current release. The actual passing time _TU of the unit vehicle of each trend in the mode;

Execute the wrong elimination rule for the number of vehicles to count the number of vehicles in each channel in each movement at each stage, including:

According to the traffic flow information, the minimum passing time and the maximum passing time of a unit vehicle in a random period of movement are set, which are respectively recorded as T _{U min} and T _{U max} ;

Compare the actual passing time per unit vehicle with the minimum passing time and the maximum passing time respectively, and determine the number of vehicles in each channel in the movement. The comparison results are as follows:

When T _{U min} <T _U <T _{U max} , the unit vehicle passing time for calculating the movement is:

Set the estimated time for the vehicle to pass the stop line according to the speed of the vehicle at the key intersection, denoted as T _Q ;

Compare the unit vehicle travel time with the estimated time to determine the number of vehicles in each channel in the movement. The comparison results are as follows:

Its further technical solution is, when the first phase is eliminated, the phase balance time of the first phase is calculated and added to the movement release duration of the next phase, until the phase balance time of the last phase is obtained after the end of the current release cycle. As the first cycle balance time, the first cycle balance time is added to the duration of the green wave phase of the coordinated intersection until the last phase of the last coordinated intersection, and the phase balance time of the last phase is obtained as the second cycle balance time, The second cycle balance time is the total phase balance time of the green wave phase under the control of the green wave, and the coordinated intersection is the downstream intersection of the key intersection.

Its further technical solution is, the expression of the phase balance time is: T _P =T _max -T;

The expression of cycle balance time is: T _C = _TP ', where _TP ' represents the last phase balance time in a release cycle.

The beneficial technical effects of the present invention are:

The control method provided by the present application effectively avoids the limitations of localized release at a single intersection, incomplete access to traffic flow information, erroneous elimination of intersection movements, and the phenomenon of congestion at the entrance and exit overflow during peak hours. Under the coordinated control of multiple intersections, the card police and video traffic detectors are used to obtain real-time vehicle data, and at the same time, the release conditions in several cycles selected during the peak and peak periods of the coordinated intersection are fully considered, and the predicted release and real-time dynamic release modes are implemented, which is reasonable and effective. Coordinate the implementation of the single-point movement prediction and release mode at the intersection, the movement elimination rule, the wrong number of vehicles elimination rule, and the coordinated downstream intersection movement elimination rule and other methods. Thereby ensuring smooth traffic flow on the main line, accurate collection and statistics of vehicle volume information, maximum utilization of signal lights, and coordinated control and release efficiency from a single intersection to multiple intersections. Use the existing vehicle data to carry out predictive release and dynamic elimination release in the release, effectively reduce the traffic time of vehicles at intersections and improve the efficiency of coordinated intersections. The signal system is effectively and reasonably coordinated with the vehicle data elimination and release criteria based on the card police and video traffic detectors to realize intelligent traffic release at the intersection.

Description of drawings

FIG. 1 is a flow chart of the predictive release mode provided by the present application.

FIG. 2 is a flow chart of the judgment of the trend elimination rule provided by the present application.

FIG. 3 is a flow chart of the judgment of the rule for erroneous elimination of the number of vehicles provided by the present application.

Detailed ways

The specific embodiments of the present invention will be further described below with reference to the accompanying drawings.

The following is a brief introduction to the professional terms involved in this method: the release cycle refers to the total duration of one release in each direction or movement, such as the start of the release of things, and a release cycle until the start of the next release of things. The release direction includes north-south direction and east-west direction, and the phases corresponding to the release direction include north-south left turn and north-south straight, east-west left turn and east-west straight. The eight major trends include going straight in the south, turning left in the south, going straight in the north, turning left in the north, going straight in the east, turning left in the east, going straight in the west, and turning left in the west. The phase is also the current release mode. For example, the combination of the north-south direction and the release indicates a phase, and the release of the south-left turn and the north-left turn together is also a phase.

The present application discloses a multi-intersection coordinated adaptive control method based on a card alarm and a video detector. Since the method for predicting the release mode during the off-peak period and the peak period is the same, this embodiment will be described in detail with the method for predicting the release mode during the off-peak period. The method includes:

Step 1: Taking the north-south direction as the release direction, the corresponding phases are left-turn from north-south and straight-through from north-south, and randomly select three consecutive release periods during the peak period of the coordinated intersection as the reference period.

Step 2: Obtain the first historical dynamic information of the reference period through the card alarm and video detector, including the traffic flow information of each movement and the release duration of each movement in each release cycle. According to the first historical dynamic information, the release rule is executed to predict and coordinate the release mode of the future release cycle during the peak period of the intersection.

As shown in Figure 1, it includes the following steps:

Step 21: Set the transition release time of the two phases respectively according to the traffic flow information, denoted as _Tri , and the transition release time refers to the sum of the green light flashing time and the yellow light lighting time.

Step 22: Calculate the average release duration of each movement in the reference period (ie, three release periods) according to the release duration of each movement in each release cycle.

Step 23: Calculate the difference between the average release durations of the two movements relative to the release direction, and denote it as _ui .

Step 24: Determine the comparison value of the two phases by comparing the magnitudes of _ui and _Tri .

The comparison values of the two phases are recorded as _ML and MS _respectively , and the comparison results are as follows:

Step 25: Determine the release mode of the future release cycle in the peak period of the coordinated intersection according to the magnitude of the two comparison values.

If M _L × M _S > 0, it includes two release modes, the first release mode is mainly in the north direction, including three phases. The first phase is the north-south straight release, the second phase is the north-left turn and the north-straight release, and the third phase is the north-south left-turn release. It should be noted that in this case, there are more vehicles going straight in the north direction and less vehicles turning left in the south direction. The second release mode is dominated by the southward release and includes three phases. The first phase is the north-south straight release, the second phase is the south-left turn and the south-straight release, and the third phase is the north-south left-turn release. It should be noted that in this case, there are more vehicles going straight in the south direction and fewer vehicles turning left in the north direction.

Step 3: Determine the current release mode of the key intersection according to the release rules, obtain the second historical dynamic information and real-time dynamic information of the key intersection through the card alarm and video detector, and execute the trend elimination rule according to the second historical dynamic information and real-time dynamic information, This updates the release mode of key intersections.

Obtain the second historical dynamic information from randomly selected three consecutive release cycles, including the number of vehicles N _i at each stage in each release cycle, the movement release duration T _i , the traffic flow information of each movement, and the number of adjacent vehicles in each movement. The interval time, the total time TA of the total vehicles passing through the stop line in each movement, and the total number of vehicles N _A passing through the stop line in each movement in the previous release cycle; real _- time dynamic information is obtained from the current release cycle, including the current release mode of each movement. The actual release duration T, the actual passing duration T _U of a unit vehicle, the total number of vehicles N' in each direction in the next phase, and the actual interval time T _n1 between two adjacent vehicles in the current phase.

Since the trend elimination rules of each phase at the key intersection are the same, this embodiment takes the first phase as an example to describe in detail, as shown in FIG. 2 , which specifically includes the following sub-steps:

Step 31: Set the maximum release duration and the minimum release duration of the first movement in the first phase according to the traffic flow information, denoted as T _max and T _min .

Step 32: Obtain the actual release duration T of the first movement in real time, and then compare it with the maximum release duration and the minimum release duration respectively. When T _min < T < T _max and the movement elimination rule is not satisfied, continue to release the first phase, and Re-execute the step of obtaining the actual release duration of the first movement in real time. When T _min < T < T _max and the movement elimination rule is satisfied, or, when T ≥ T _max , the first movement is eliminated, and the steps of setting the maximum release duration and the minimum release duration of the next movement in the first phase are performed , judge all the movements of the first phase in turn, if all the movements are eliminated, the first phase will be eliminated, and the movement elimination rules will be implemented for the next phase.

Methods for judging whether the trend elimination rule is met include:

Step 321: Set the vehicle queue length threshold, and obtain the actual vehicle queue length of the next phase in real time.

1) Considering the difference in the traffic density (that is, the number of vehicles passing through per unit time) at the beginning, during and at the end of the phase and each movement in a release cycle of the actual intersection, for each selected release cycle, the vehicle will be released. The movement release duration of all phases of the cycle is divided into three phases. The first stage is the starting stage, and the number of vehicles counted in this stage is recorded as N ₁ ; the second stage is 1/3 to 2/3 of the duration of the moving release. That is, in the release stage, the number of vehicles counted in this stage is recorded as N ₂ ; the third stage is the end stage from 2/3 of the moving release time to the end, and the number of vehicles counted in this stage is recorded as N ₃ .

2) Considering that the actual intersection will be extremely congested or the size of the vehicle is different, implement the wrong elimination rule of the number of vehicles to accurately count the number of vehicles in each channel in each phase and each movement.

3) Calculate the average number of vehicles in each stage of the three release cycles, denoted as

Optionally, if the number of lanes for a certain movement is greater than 1, the number of vehicles counted in this stage should be multiplied by the number of lanes as the number of vehicles in this stage. The length of the space occupied by each vehicle can be taken as an empirical value, that is, l=7m.

4) According to the real-time acquisition of the total number N' of vehicles in each direction in the next phase, the actual vehicle queuing length in the next phase is obtained as: L=N'×n×l.

Step 322 : Set a threshold for the number of vehicles, and calculate the actual number of vehicles in each movement unit count period.

1) Select the maximum value in the average value of the number of vehicles in each stage of the release cycle, divide it by the corresponding moving release duration, and obtain the quotient as the threshold of the number of vehicles, denoted as N _max .

2) Calculate the actual number of vehicles in each movement unit count cycle as:

Among them, T ₁ , T ₂ , and T ₃ are the corresponding movement release durations of each stage, respectively.

Step 323 : Set the interval time threshold between two adjacent vehicles, and obtain the actual interval time T _n1 of the two adjacent vehicles in the current phase in real time.

The average value of the interval time between two adjacent vehicles in each direction of the three release cycles is calculated as the threshold value of the interval time between the two adjacent vehicles, which is denoted as T _n .

Step 324: Calculate the balance time.

The balance time is divided into two types, one is the phase balance time, denoted as T _P . The phase balance time is the difference between the maximum green time of the phase and the actual release green light duration of the phase. It can be seen from step 32 that the expression of the phase balance time is: T _P =T _max -T, it should be noted that T _P can be a positive balance , which can also be a negative balance;

The other is the cycle balance time, denoted as T _C . The period balance time is the last phase balance time _TP ' in this release period, and its expression is: T _C = _TP '.

Step 325: Compare the actual vehicle queuing length L of the next phase, the actual number of vehicles N _t and the actual interval time T _n1 obtained in real time with the corresponding thresholds, respectively, to determine the elimination standard value, denoted as P, and the comparison results are as follows:

Among them, if P=1 indicates that the current movement is eliminated, if P=0 indicates that the current movement is not eliminated, if at least one movement of the current phase is not eliminated, the current phase is not eliminated, and the current phase continues to be released.

Condition [1] indicates that there are fewer vehicles in the current movement, and the current movement needs to be eliminated. On the contrary, the condition [4] indicates that there are many vehicles in the current movement, and the current movement needs to continue to be released.

Condition [2] indicates that the distance between two adjacent vehicles in the current trend is large, and the traffic flow is sparse, so the current trend needs to be eliminated. Conversely, Condition [5] indicates that the distance between two adjacent vehicles in the current movement is small, and the traffic flow is denser, and the current movement needs to be released.

Condition [3] indicates that the moving traffic flow in the next phase is large and the number of vehicles moving in the current phase is less, and the current moving needs to be eliminated. Conversely, Condition [6] indicates that the traffic flow in the next phase is denser, and the current direction needs to be released.

Using the same method as above, the maximum release duration T _max ' and the minimum release duration T _min ' of each movement at the coordinated intersection, the vehicle queue length threshold L _max ', the vehicle number threshold N _max ' and the two adjacent vehicles for each movement can be obtained. Interval time threshold _Tn '. Among them, the coordination intersection is the downstream intersection of the key intersection.

When the first phase is eliminated, calculate the phase balance time of the first phase and add it to the movement release duration of the next phase, until the end of the current release cycle to obtain the phase balance time of the last phase as the balance time of the first cycle, Add the balance time of the first cycle to the duration of the green wave phase of the coordinated intersection until the last phase of the last coordinated intersection, and obtain the phase balance time of the last phase as the balance time of the second cycle. The balance time of the second cycle is green. The total phase balance time of the green wave phase under wave control.

In step 321, the number of vehicles by mistake elimination rule is executed to accurately count the number of vehicles in each channel in each movement in each stage, as shown in Figure 3, which specifically includes the following steps:

1) According to the traffic flow information, the minimum passing time and the maximum passing time of a unit vehicle in the first random period of movement are set, which are respectively recorded as T _{U min} and T _{U max} .

2) Compare the actual passing time T _U of the unit vehicle with the minimum passing time T _{U min} and the maximum passing time T _{U max} , respectively, to determine the number of vehicles N _U ′ in each channel in the movement, and the comparison results are as follows:

3) When T _{U min} <T _U <T _{U max} , further judgment is required: the unit vehicle passing time T _U ′ for calculating the trend is:

4) Set the estimated time for the vehicle to pass the stop line according to the speed of the vehicle at the key intersection, denoted as T _Q .

5) Compare the unit vehicle passing time T _U ' with the estimated time T _Q , and further determine the number of vehicles N _U ' in each channel in the movement, and the comparison results are as follows:

Under the coordinated control of multiple intersections, the card police and video traffic detectors are used to obtain real-time vehicle data, and at the same time, the release conditions in several cycles selected during the peak and peak periods of the coordinated intersection are fully considered, and the predicted release and real-time dynamic release modes are implemented, which is reasonable and effective. Coordinate the implementation of the single-point movement prediction and release mode at the intersection, the movement elimination rule, the wrong number of vehicles elimination rule, and the coordinated downstream intersection movement elimination rule and other methods. Thereby ensuring smooth traffic flow on the main line, accurate collection and statistics of vehicle volume information, maximum utilization of signal lights, and coordinated control and release efficiency from a single intersection to multiple intersections. Use the existing vehicle data to carry out predictive release and dynamic elimination release in the release, effectively reduce the traffic time of vehicles at intersections and improve the efficiency of coordinated intersections. The signal system is effectively and reasonably coordinated with the vehicle data elimination and release criteria based on the card police and video traffic detectors to realize intelligent traffic release at the intersection.

The above descriptions are only preferred embodiments of the present application, and the present invention is not limited to the above embodiments. It can be understood that other improvements and changes directly derived or thought of by those skilled in the art without departing from the spirit and concept of the present invention should be considered to be included within the protection scope of the present invention.

Claims

A multi-intersection coordinated adaptive control method based on a card alarm and a video detector, characterized in that the method comprises:

Respectively select the release period of the coordinated intersection during the flat-peak period and the peak period as the reference period, and the release mode of each said release period is the same;

Obtain the first historical dynamic information of the reference period through the card alarm and video detector, and execute the release rule according to the first historical dynamic information to predict the release mode of the future release period of the coordinated intersection during the off-peak period and the peak period, respectively;

Determine the current release mode of the key intersection according to the release rule, obtain the second historical dynamic information and real-time dynamic information of the key intersection through the card alarm and video detector, and execute the movement according to the second historical dynamic information and real-time dynamic information. Eliminate the rule to update the release pattern of the critical intersection.
The multi-intersection coordinated adaptive control method based on a card alarm and a video detector according to claim 1, wherein the first historical dynamic information includes traffic flow information of each direction, and the release duration of each direction in each release cycle; Executing the release rule according to the first historical dynamic information to predict the release mode of the future release cycle during the peak period of the coordinated intersection, including:

In one release direction, the transition release time of two phases is respectively set according to the traffic flow information, denoted as Tri , the release direction includes north-south direction and east-west direction, and the phase includes north-south left turn and north-south straight, east-west left turn and things go straight;

Calculate the average release duration of each movement in the reference period according to the release duration of each movement in each release cycle;

Calculate the difference between the average release durations of the two movements relative to the release direction, denoted as u i ;

Determine the comparison value of the two phases by comparing the magnitudes of u i and Tri;

According to the magnitude of the two comparison values, the release mode of the future release period of the coordinated intersection during the peak period is determined.
The multi-intersection coordination adaptive control method based on a card alarm and a video detector according to claim 2, characterized in that, determining the comparison value of two phases by comparing the size of u i and Tri, comprising:

Assuming that the release direction is the north-south direction, the corresponding phases are the north-south turn left and the north-south straight, and the comparison values of the two phases are recorded as M L and M S respectively, and the comparison results are as follows:

Among them, u 1 and u 2 are the difference between the average release durations of south/north turn left and south/north straight, and u 1 =NL t -SL t , NL t is the average release duration of north left turn, and SL t is south left Turn average release time, u 2 =NS t -SS t , NS t is the average release time of north going straight, SS t is the average release time of south going straight; Tr 1 is the transitional release time of north-south left turn, Tr 2 is the transition of north-south straight release time.
The multi-intersection coordination adaptive control method based on a card alarm and a video detector according to claim 3, wherein the determination of the future release cycle of the coordinated intersection during the peak period is determined according to the magnitude of the two comparison values. Release modes, including:

If M L × M S >0, it includes two release modes, each of which contains three phases, including north-south straight release, left-turn and straight-through release in the main direction, and north-south left-turn release, where the main direction is south or North;

If M L × M S ≤ 0, the release mode contains two phases, including north-south straight release and north-south left-turn release.
The multi-intersection coordination adaptive control method based on card alarm and video detector according to claim 1, wherein the second historical dynamic information is obtained from several consecutive release cycles randomly selected, including each release cycle. Traffic flow information of each movement in the cycle; obtain the real-time dynamic information from the current release cycle, including the actual release duration of each movement in the current release mode;

The movement elimination rules of each phase of the key intersection are the same, then the movement elimination rules are executed according to the second historical dynamic information and real-time dynamic information, so as to update the release mode of the key intersection, including:

Set the maximum release duration and the minimum release duration of the first movement in the first phase according to the traffic flow information, denoted as T max , T min ;

Obtain the actual release duration of the first movement in real time and denote it as T, then compare with the maximum release duration and the minimum release duration respectively, when T min <T < T max and the movement elimination rule is not satisfied, continue Release the first phase, and re-execute the step of obtaining the actual release duration of the first movement in real time; when T min <T < T max and the movement elimination rule is satisfied, or, when T ≥ T max , Eliminate the first movement, and perform the steps of setting the maximum release duration and the minimum release duration of the next movement in the first phase, and judge all movements of the first phase in turn. If all movements are eliminated, then eliminate all the movements. The first phase is described, and the dynamic elimination rule is started to be executed for the next phase.
The multi-intersection coordination and adaptive control method based on a card alarm and a video detector according to claim 5, wherein the second historical dynamic information further includes the number of vehicles at each stage in each release cycle, and the movement release duration. , the interval time between two adjacent vehicles in each movement, and the real-time dynamic information also includes the total number of vehicles in each movement in the next phase and the actual interval time T n1 between the two adjacent vehicles in the current phase;

The method for judging whether the trend elimination rule is satisfied includes:

For each selected release cycle, divide the movement release duration of all phases of the release cycle into three stages and count the number of vehicles in each stage, and implement the vehicle number error elimination rule to accurately count each movement in each stage. the number of vehicles in the channel;

Calculate the average number of vehicles at each stage of all selected release cycles, denoted as
Determine the number of channels n of each movement and the length l of the space occupied by each vehicle, then the threshold of the queue length of vehicles in each movement is:

Select the maximum value in the average value of the number of vehicles in each stage of the release cycle, divide it by the corresponding moving release duration, and the obtained quotient value is used as the threshold value of the number of vehicles, denoted as N max ;

Calculate the average value of the interval time between two adjacent vehicles in each movement of all the selected release cycles as the interval time threshold between the two adjacent vehicles, denoted as T n ;

According to the real-time acquisition of the total number of vehicles N' in each direction in the next phase, the actual vehicle queue length of the next phase is obtained as: L=N'×n×l;

Calculate the actual number of vehicles in each movement unit count period as:

Among them, T 1 , T 2 , and T 3 are the corresponding movement release durations of each stage;

The actual vehicle queuing length, actual number of vehicles and actual interval time of the next phase obtained in real time are compared with the corresponding thresholds, respectively, to determine the elimination standard value, which is denoted as P, and the comparison results are as follows:

Among them, if P=1 indicates that the current movement is eliminated, if P=0 indicates that the current movement has not been eliminated, if at least one movement of the current phase has not been eliminated, the current phase has not been eliminated, and the current phase continues to be released;

Condition [1] indicates that there are fewer vehicles in the current movement, and the current movement needs to be eliminated; otherwise, condition [4] indicates that the current movement needs to continue to be released;

Condition [2] indicates that the distance between two adjacent vehicles in the current movement is relatively large, and the traffic flow is sparse, and the current movement needs to be eliminated; otherwise, the condition [5] indicates that the current movement needs to continue to be released;

Condition [3] indicates that the moving traffic flow in the next phase is large and the number of vehicles moving in the current phase is less, and the current movement needs to be eliminated. On the contrary, Condition [6] indicates that the moving traffic flow in the next phase is dense, and it is necessary to continue to release the current trend.
The multi-intersection coordinated adaptive control method based on a card alarm and a video detector according to claim 6, wherein the second historical dynamic information further includes the total time T A for the total vehicles in each direction to pass the stop line and the The total number of vehicles NA passing through the stop line in each movement of each release cycle, the real - time dynamic information further includes the actual passing time TU of the unit vehicle of each movement in the current release mode;

The number of vehicles in the execution of the wrong elimination rule for the number of vehicles counts the number of vehicles in each channel in each movement at each stage, including:

According to the traffic flow information, the minimum passing time and the maximum passing time of a unit vehicle in a random period of movement are set, which are respectively recorded as T U min and T U max ;

The actual passing time of the unit vehicle is compared with the minimum passing time and the maximum passing time, respectively, to determine the number of vehicles in each channel in the movement, and the comparison results are as follows:

When T U min <T U <T U max , the unit vehicle passing time for calculating the movement is:

Set the estimated time for the vehicle to pass the stop line according to the speed at which the vehicle travels at the key intersection, denoted as T Q ;

Compare the passing time of the unit vehicle with the estimated time to determine the number of vehicles in each channel in the movement, and the comparison results are as follows:
The multi-intersection coordination adaptive control method based on a card alarm and a video detector according to any one of claims 5-7, wherein when the first phase is eliminated, the phase balance of the first phase is calculated time, and add it to the movement release duration of the next phase, until the end of the current release cycle, the phase balance time of the last phase is obtained as the first cycle balance time, and the first cycle balance time is added to the green wave at the coordinated intersection. In the duration of the phase, until the last phase of the last coordinated intersection, the phase balance time of the last phase is obtained as the second cycle balance time, and the second cycle balance time is the phase balance of the green wave phase under the control of the green wave The total time, the coordinated intersection is the downstream intersection of the key intersection.
The multi-intersection coordination adaptive control method based on a card alarm and a video detector according to claim 8, wherein the expression of the phase balance time is: T P =T max -T;

The expression of cycle balance time is: T C = TP ', where TP ' represents the last phase balance time in a release cycle.