Green wave design speed calculation method and device based on bayonet data

Through the green wave design speed calculation method based on bayonet data, the license plate matching algorithm and statistical method are used to screen and fit the vehicle travel time, a multi-objective optimization model is established, and the green wave design speed is solved using simulated annealing algorithm, which has solved the existing technology The problem of insufficiently accurate design speed of medium green waves improves the effect of green wave coordination control.

CN118942262APending Publication Date: 2024-11-12SHANDONG JIAOTONG UNIV +3
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411165173.9
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-23
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, the green wave design speed is not accurate enough, and it is difficult to deal with changes in the dynamic driving speed of vehicles on the road, resulting in long delays and more parking times, affecting the green wave coordination control effect.

Method used

The green wave design speed calculation method based on bayonet data is used to determine the vehicle travel time through the license plate matching algorithm, and the data screening and fitting are carried out in combination with box graph method and statistical method, a multi-objective optimization model is established, and a simulated annealing algorithm is used to solve it. Green wave design speed scheme.

Benefits of technology

It improves the accuracy and effect of green wave coordination control, reduces delay time and parking times, and can effectively deal with changes in the dynamic driving speed of vehicles on the road.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118942262A_ABST
    Figure CN118942262A_ABST
Patent Text Reader

Abstract

The invention discloses a green wave design speed calculation method and device based on checkpoint data, belongs to the technical field of intelligent traffic control, and is used for solving the problems that the timeliness of existing vehicle travel time data is poor, the green wave design speed is not accurate enough, the dynamic running speed change of vehicles on a road is difficult to affect, and the traffic safety is poor. And improvement of the green wave coordination control effect is not facilitated. The method comprises the following steps: determining vehicle travel time between two adjacent intersections; performing sample data filtering processing on the vehicle travel time through the statistical time window of each time period to obtain effective vehicle travel time; performing data fitting between related speed values and cumulative frequency on the effective vehicle travel time to determine a fitting curve; establishing a multi-objective optimization model based on the cumulative frequency maximum difference value and the speed value minimum difference value in the fitting curve; and through a preset simulated annealing rule, carrying out iterative calculation of a global optimal solution on the multi-objective optimization model, and outputting a green wave design speed scheme.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of intelligent traffic control technology, and in particular to a method and device for calculating a green wave design speed based on checkpoint data. Background Art

[0002] At present, there are various methods for the study of green wave design speed. Some researchers have proposed a road section driving speed design method for green wave coordinated control. This method, on the premise of meeting the road section driving speed range requirements and ensuring the maximization of the green wave bandwidth of the trunk road, makes the vehicle's driving speed on the entire trunk road the maximum and the vehicle's driving speed on different sections as close as possible, thereby effectively improving the road's traffic efficiency and improving driving stability and comfort; Some researchers have provided a regional multi-directional green wave design and driving speed guidance method based on the Internet of Vehicles. This method adjusts the intersection signal phase scheme and optimizes the phase difference between intersections based on the topological structure of the road network, thereby providing green waves and speed induction signals for traffic in any direction within the service road network; Some researchers have provided a method for determining the green wave speed. This type of method solves the problem that the related technology relies on expert experience to determine the green wave speed, cannot adapt to the dynamic changes of the trunk traffic status, and thus makes the green wave coordinated trunk traffic flow poor.

[0003] Traditional arterial coordinated control models usually set green wave coordinated control schemes based on average speed or design speed of road sections according to driving experience, measured data or sampling collection methods. However, in actual situations, the speed of a road section is affected by factors such as the number of lanes, the length of the road section, and whether there is traffic confluence. The vehicle speed should have a certain fluctuation range. When the vehicle travels at a speed different from the green wave design speed, the green wave coordinated control effect will be weakened. Therefore, a reasonable green wave design speed calculation method is of great significance for inducing and regulating the vehicle speed and improving the operation effect of the green wave. However, some traditional traffic detection equipment, such as coils and geomagnetism, can only obtain the aggregated traffic information of the section, and it is difficult to capture the specific driving information of the vehicle under green wave control (such as road section speed, trajectory, etc.), which is not conducive to the calculation task of the green wave design speed. Summary of the invention

[0004] The embodiments of the present application provide a method and device for calculating the green wave design speed based on checkpoint data, which are used to solve the following technical problems: the timeliness of existing vehicle travel time data is poor, the green wave design speed is not accurate enough, and it is difficult to effectively respond to the dynamic driving speed changes of vehicles on the road, which easily causes long delays and more stops, and is not conducive to improving the effect of green wave coordinated control.

[0005] The present application embodiment adopts the following technical solutions:

[0006] On the one hand, an embodiment of the present application provides a method for calculating a green wave design speed based on checkpoint data, including: determining a vehicle travel time between two adjacent intersections based on a license plate matching algorithm; filtering sample data of the vehicle travel time through a statistical time window of each time period to obtain an effective vehicle travel time; performing data fitting between relevant speed values ​​and cumulative frequencies on the effective vehicle travel time to determine a fitting curve; establishing a multi-objective optimization model based on the maximum difference between the cumulative frequencies and the minimum difference between the speed values ​​in the fitting curve; iteratively calculating a global optimal solution for the multi-objective optimization model through a preset simulated annealing rule, and outputting a green wave design speed plan.

[0007] The method of obtaining vehicle travel time data through checkpoint data in the embodiment of the present application has the advantages of small investment, large sample size, and high data accuracy, and provides a corresponding checkpoint data processing method, which expands the traditional travel time survey method. By redefining the green wave design speed, a specific green wave design speed calculation process is proposed, which lays the foundation for the research of green wave coordinated control. It is also possible to screen out the delayed vehicle travel time data by analyzing the vehicle delay mechanism and combining the box plot method with the statistical method; and establish a multi-objective optimization model based on the distribution characteristics of the vehicle driving speed; use the simulated annealing algorithm to solve the driving speed range of the vast majority of vehicles without delays; provide a scientific and practical theoretical method for the calculation of the green wave design speed, which can effectively respond to the dynamic driving speed changes of vehicles on the road, reduce delay time and the number of stops, and is of great significance to improving the effect of green wave coordinated control.

[0008] In a feasible implementation, based on the license plate matching algorithm, the vehicle travel time between two adjacent intersections is determined, specifically including: determining a screening time threshold based on the maximum duration of the all-red signal phase of the two adjacent intersections; screening all the acquired driving record data for a single intersection, and grouping all the driving record data by license plate number; performing time series processing on all the grouped driving record data; calculating the data time difference between the same license plate data in the single intersection; removing duplicate driving record data from all the driving record data through the screening time threshold to obtain valid driving record data; marking abnormal vehicle license plate data in the valid driving record data as abnormal, and deleting the abnormal vehicle license plate data; based on the location information of two adjacent intersections and through the license plate matching algorithm, screening the same license plate data for the valid driving record data to obtain the same license plate data at different timestamps; determining the data time difference as the vehicle travel time of the same vehicle based on the same license plate data at different timestamps.

[0009] In a feasible implementation manner, the sample data of the vehicle travel time is filtered through the statistical time window of each time period to obtain the effective vehicle travel time, which specifically includes: segmenting the vehicle travel time period between two adjacent intersections to obtain multiple time periods; wherein the multiple time periods include: morning peak time period, evening peak time period and off-peak time period; based on the box plot, the vehicle travel time in each time period is filtered within the numerical boundary to obtain the first vehicle travel time; wherein the parameters in the box plot include: lower quartile value, upper quartile value, upper and lower quartile difference, upper boundary and lower boundary; the first vehicle travel time is sampled and processed to obtain a first travel sample set; based on the first travel sample set, the mean and standard deviation of the first vehicle travel time are determined; and the vehicle travel samples in the first travel sample set that are beyond two times the standard deviation of the mean are filtered out to obtain a second travel sample set; the second travel sample set is processed with absolute deviation to obtain a third travel sample set; the vehicle travel time of the third travel sample in the statistical time window of each time period is determined as the effective vehicle travel time.

[0010] In a feasible implementation manner, absolute deviation processing is performed on the second trip sample set to obtain a third trip sample set, specifically including: sorting the travel time samples in the second trip sample set to determine the median; calculating the absolute deviation between each travel time sample in the second trip sample set and the median; according to the absolute deviation and with the median as the dividing line, obtaining the first half average absolute deviation and the second half average absolute deviation of the second trip sample set respectively; performing threshold filtering on the second trip sample set by means of the median, the first half average absolute deviation and the second half average absolute deviation to obtain the third trip sample set.

[0011] In a feasible implementation manner, data fitting between relevant speed values ​​and cumulative frequencies is performed on the effective vehicle travel time to determine a fitting curve, which specifically includes: determining the effective vehicle travel speed based on a section length parameter between two adjacent intersections and the effective vehicle travel time; grouping the effective vehicle travel speeds to obtain a travel speed distribution table; and based on the travel speed distribution table, dividing the effective vehicle travel speeds into intervals to obtain a vehicle speed frequency distribution histogram; based on the vehicle speed frequency distribution histogram, determining a speed cumulative frequency distribution curve related to the effective vehicle travel time; performing curve fitting processing on all speed values ​​in the effective vehicle travel speed and the speed cumulative frequency in the speed cumulative frequency distribution curve to obtain the fitting curve.

[0012] In a feasible implementation manner, before establishing a multi-objective optimization model based on the maximum cumulative frequency difference and the minimum speed value difference in the fitting curve, the method further includes: Obtain the maximum cumulative frequency difference maxf1(x); wherein x0 is a specific value in the fitting curve; x1 is a first speed value in the fitting curve; x2 is a second speed value in the fitting curve; ω1 is a first function coefficient related to the maximum cumulative frequency difference; A max is the maximum value in the fitting curve; min is the minimum value in the fitting curve; s and h are the curve fitting parameters of the fitting curve; according to minf2(x)=ω2(x2-x1), the minimum speed value difference minf2(x) is obtained; wherein ω2 is the second function coefficient related to the minimum speed value difference.

[0013] In a feasible implementation, a multi-objective optimization model is established based on the maximum cumulative frequency difference and the minimum speed value difference in the fitting curve, specifically including: performing data mapping processing on the maximum cumulative frequency difference and the minimum speed value difference to obtain a normalized sequence; weighting optimization the maximum cumulative frequency difference and the minimum speed value difference in the normalized sequence through a preset scoring function to obtain a quantitative score; and taking the maximum cumulative frequency difference, the minimum speed value difference and the value range of the cumulative frequency and the speed value as target constraints to establish the multi-objective optimization model.

[0014] In a feasible implementation manner, before iteratively calculating the global optimal solution of the multi-objective optimization model through a preset simulated annealing rule and outputting the green wave design speed scheme, it specifically includes: determining the initial temperature parameters, the temperature attenuation function and the Markov chain length based on the Metropolis criterion of the simulated degradation algorithm; randomly selecting two speed values ​​through a preset initial vehicle speed value alternative speed range and determining them as the initial green wave design speed scheme; and calculating the corresponding first scoring function value based on the initial green wave design speed scheme.

[0015] In a feasible implementation manner, the multi-objective optimization model is iteratively calculated for the global optimal solution through a preset simulated annealing rule, and a green wave design speed scheme is output, specifically including: based on the initial green wave design speed scheme, a new green wave design speed scheme is generated and a corresponding second scoring function value is obtained; if the second scoring function value is greater than the first scoring function value, the new green wave design speed scheme is received; according to the ratio of the initial temperature parameter to the changed temperature parameter, the new green wave design speed scheme is converged with respect to the global optimal solution, and the green wave design speed scheme is output to complete the coordinated control of the green wave.

[0016] On the other hand, an embodiment of the present application also provides a green wave design speed calculation device based on checkpoint data, comprising: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute a green wave design speed calculation method based on checkpoint data as described in any of the above embodiments.

[0017] Compared with the prior art, the present application provides a method and device for calculating the green wave design speed based on bayonet data. Compared with the prior art, the present application has the following beneficial technical effects:

[0018] 1. Expanded the method of obtaining vehicle travel time data. The traditional way to obtain travel time data is to obtain relevant parameters through field surveys. However, as the scale of cities continues to expand, it is increasingly difficult to adapt to the needs of traffic analysis due to poor timeliness and small sample size. The method of using checkpoint data to obtain vehicle travel time data has the advantages of low investment, large sample size, and high data accuracy. It also provides corresponding checkpoint data processing methods, which expands the traditional travel time survey method.

[0019] 2. Improve the research related to green wave design speed. At present, most of the research on green waves is focused on the green wave coordinated control scheme, and there is less research on the green wave design speed parameter. By redefining the green wave design speed, a set of specific green wave design speed calculation processes is proposed, which lays the foundation for the research on green wave coordinated control.

[0020] 3. The green wave design speed is calculated based on vehicle delay analysis. By analyzing the mechanism of vehicle delay, the box plot method is combined with the statistical method to screen out the travel time data of vehicle delays; and a multi-objective optimization model is established based on the distribution characteristics of vehicle speeds. The simulated annealing algorithm is used to solve the speed range of the vast majority of vehicles without delays. It provides a scientific and practical theoretical method for the calculation of the green wave design speed, which can effectively respond to the dynamic speed changes of vehicles on the road, reduce delay time and the number of stops, and is of great significance to improving the effect of green wave coordinated control. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0022] Figure 1A flow chart of a method for calculating green wave design speed based on bayonet data provided in an embodiment of the present application;

[0023] Figure 2 A schematic diagram of the structure of a green wave design speed calculation device based on bayonet data provided in an embodiment of the present application. DETAILED DESCRIPTION

[0024] In order to enable those skilled in the art to better understand the technical solutions in the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments of this specification, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of this application.

[0025] It should be noted that the rapid development of modern big data technology has made it possible to collect, store, process and analyze massive traffic data. Various types of traffic detection equipment, such as checkpoints, floating vehicles, bus cards, etc., can collect individual traffic information data in real time. The diversification of these data sources provides a rich information resource for the intelligent transportation system. Among them, the checkpoint data is captured by the checkpoint cameras set on the road. The specific time, driving direction, lane usage position of different vehicles passing through each checkpoint, so as to accurately obtain the personalized and refined data such as the speed, driving time, and driving path of each vehicle on the road. Therefore, it has the characteristics of real-time, accuracy, and high granularity. It can well reflect the actual traffic conditions, which is conducive to more accurate calculation of the green wave design speed and improve the stability of the green wave. Therefore, the calculation method of the green wave design speed based on the checkpoint data has high practical value and application prospects.

[0026] The present application embodiment provides a method for calculating the green wave design speed based on bayonet data, such as Figure 1 As shown, the green wave design speed calculation method based on the bayonet data specifically includes steps S101-S105:

[0027] S101. Determine a vehicle travel time between two adjacent intersections based on a license plate matching algorithm.

[0028] Specifically, it is necessary to first determine the screening time threshold based on the maximum duration of the full red signal phase of two adjacent intersections. Then, all the acquired driving record data are screened for a single intersection, and all the driving record data are grouped by license plate number. All the grouped driving record data are processed in time series.

[0029] Furthermore, the data time difference between the same license plate data in a single intersection is calculated. The duplicate driving record data in all driving record data is eliminated by using the screening time threshold to obtain valid driving record data. Then, the abnormal vehicle license plate data in the valid driving record data is marked as abnormal and the abnormal vehicle license plate data is deleted.

[0030] Furthermore, it is also necessary to screen the same license plate data of the valid driving record data based on the location information of the two adjacent intersections and through the license plate matching algorithm to obtain the same license plate data at different timestamps. Based on the same license plate data at different timestamps, the vehicle travel time of the same vehicle is determined.

[0031] In one embodiment, a time threshold B is used to determine whether the record generated by a vehicle at a specific intersection is redundant data, and the formula is as follows: B = max{T a The value of time threshold B is the maximum all-red time, T a is the all-red time of the a-th signal intersection. Select all the vehicle records at each shooting location from the vehicle record data, group all the vehicle record data by license plate number, sort these records in chronological order, and calculate the time difference of the record data with the same license plate number at the intersection. If the time difference is less than this time threshold, the latter record is considered to be repeated and removed; otherwise, it is considered normal data.

[0032] As a feasible implementation method, since the vehicle license plate information is missing or exceeds the specified length and cannot be accurately repaired, it is necessary to identify these abnormal license plate information and remove them from the system. Filter the valid data extracted from the traffic operation parameters such as license plate number, collection location, collection timestamp, lane function, etc. According to the license plate matching algorithm, filter the vehicle information of the same license plate number in two adjacent intersections, calculate the time difference between the vehicles, that is, the vehicle travel time of the same vehicle.

[0033] S102: Filter the sample data of the vehicle travel time through the statistical time window of each time period to obtain the effective vehicle travel time.

[0034] Specifically, the vehicle traffic period between two adjacent intersections is first divided into sections to obtain multiple time periods, wherein the multiple time periods include: a morning peak period, an evening peak period, and an off-peak period.

[0035] Furthermore, the box plot is used to filter the vehicle travel time in each time period within the numerical boundary to obtain the first vehicle travel time. The parameters in the box plot include: lower quartile value, upper quartile value, upper and lower quartile difference, upper boundary and lower boundary.

[0036] Furthermore, the first vehicle travel time is processed into a sample set to obtain a first travel sample set. Based on the first travel sample set, the mean and standard deviation of the first vehicle travel time are determined. The vehicle travel samples outside the two standard deviations of the mean in the first travel sample set are filtered out to obtain a second travel sample set.

[0037] Furthermore, the second trip sample set is processed with absolute deviation to obtain the third trip sample set. That is, the travel time samples in the second trip sample set are sorted and the median value is determined. Then, the absolute deviation between each travel time sample in the second trip sample set and the median value is calculated. Then, based on the absolute deviation and taking the median value as the dividing line, the first half average absolute deviation and the second half average absolute deviation of the second trip sample set are obtained respectively.

[0038] Furthermore, it is also necessary to perform threshold filtering on the second trip sample set through the median value, the first half average absolute deviation, and the second half average absolute deviation to obtain a third trip sample set. The vehicle travel time of the third trip sample in the statistical time window of each period is determined as the effective vehicle travel time. That is, it is necessary to determine the median value of the travel time sample set, calculate the absolute deviation of the travel time samples in the first half and the second half of the median value and the median value, and filter out the travel time that is greater than the sum of the median value and three times the absolute deviation of the first half, and less than the sum of the median value and three times the absolute deviation of the second half, so as to obtain the third trip sample set.

[0039] In one embodiment, based on the data set collected during the statistical period, the lower quartile value Q1, the upper quartile value Q3, the upper and lower quartile difference IQR, the upper boundary Q max and the lower boundary Q min These values ​​are used to filter out min ,Q max ] and remove abnormal data outside this range, and filter the travel time for the first time:

[0040] IQR=Q3-Q1

[0041] Q max =Q3+1.5IQR

[0042] Q min =Q1-1.5IQR

[0043] Next, the travel time in the jth time window constitutes the first travel time sample set TDS j1 , calculate TDS j1 Mean travel time and standard deviation Filter out the travel time data beyond 2 standard deviations from the mean, and the remaining travel time constitutes the second travel time sample set TDS j2 , the formula is as follows:

[0044]

[0045] Where: i is the serial number of the travel time; n is the number of samples.

[0046] Then TDS j2 The travel time is arranged in ascending order, and the median value TT of the travel time in the statistical time window is calculated. jmedian , the formula is as follows:

[0047] TT jmedian =TT jm

[0048] and

[0049] Where: max(i) is the maximum sequence number of the travel time in the j statistical time window, that is, TDS j2 The number of samples of internal travel time;

[0050] Then calculate the absolute deviation of each travel time sample from the median value, denoted as D ji The median value is used as the dividing line to calculate the average absolute deviation of the travel time of the first and second halves of the second travel time sample set from the median value, which are recorded as D ifmedian , D jamedian , and the calculation formulas for the two are as follows:

[0051] D ji =TT ji -TT jmedian

[0052]

[0053] The travel time is filtered based on the median value and the mean absolute deviation. The remaining third travel time sample set after filtering is TDS. j3 ,Right now

[0054] TDS j3 =(TT ji |TT jmedian +3D ifmedian ≤TT ji ≤TT jmedian +3D jamedian )

[0055] Finally, the above filtering process is repeated until the travel time sample data no longer changes. The effective travel time sample set of statistical time window j is TDS j, which is determined as the effective vehicle travel time.

[0056] S103, performing data fitting between relevant speed values ​​and cumulative frequencies on the effective vehicle travel time to determine a fitting curve.

[0057] Specifically, firstly, the effective vehicle speed is determined based on the road section length parameter between two adjacent intersections and the effective vehicle travel time. Then, the effective vehicle speed is grouped to obtain a speed distribution table. And based on the speed distribution table, the effective vehicle speed is divided into intervals to obtain a speed frequency distribution histogram.

[0058] Furthermore, the speed frequency distribution histogram is used to determine the speed cumulative frequency distribution curve related to the effective vehicle travel time. All speed values ​​in the effective vehicle travel speed are subjected to curve fitting processing with the speed cumulative frequency in the speed cumulative frequency distribution curve to obtain a fitting curve.

[0059] In one implementation, the effective vehicle travel time is obtained, and the road section length parameter is introduced to obtain the driving speed of the vehicle on the road section during the period. The formula is as follows:

[0060]

[0061] Where: v is the effective vehicle speed of the vehicle within the sampling time interval of the road section, m / s; t b is the travel time of the bth vehicle; S is the length of the road section.

[0062] Next, after obtaining the effective vehicle speed without delay, determine the number of groups and group distances, group the vehicle speed data, obtain the distribution table of the driving speed, and obtain the vehicle speed frequency distribution histogram and the speed cumulative frequency distribution curve.

[0063] Then, using the speed cumulative frequency distribution curve, the 15% vehicle speed value, 50% vehicle speed value and 85% vehicle speed value on the cumulative frequency distribution curve are roughly obtained as the minimum speed limit, median speed and maximum speed limit indicators on the road respectively.

[0064] A variety of fitting functions are used to fit the velocity value and the cumulative frequency. 2 , we get a better fitting function equation between speed and cumulative frequency, that is, the fitting curve is:

[0065]

[0066] Among them: A min , A max , x0, h, s are the curve fitting parameters of speed and cumulative frequency.

[0067] S104: Establish a multi-objective optimization model based on the maximum difference in cumulative frequency and the minimum difference in speed value in the fitting curve.

[0068] Specifically, first according to The maximum cumulative frequency difference maxf1(x) is obtained. Where, x0 is a specific value in the fitting curve; x1 is the first speed value in the fitting curve; x2 is the second speed value in the fitting curve; ω1 is the first function coefficient related to the maximum cumulative frequency difference; A max is the maximum value in the fitting curve; A min is the minimum value in the fitting curve; s and h are the curve fitting parameters of the fitting curve. According to minf2(x)=ω2(x2-x1), the minimum speed difference minf2(x) is obtained. Among them, ω2 is the second function coefficient related to the minimum speed difference.

[0069] Furthermore, the maximum difference of the cumulative frequency and the minimum difference of the speed value are subjected to data mapping processing to obtain a normalized sequence. Then, the maximum difference of the cumulative frequency and the minimum difference of the speed value in the normalized sequence are weighted optimized through a preset scoring function to obtain a quantitative score. The maximum difference of the cumulative frequency, the minimum difference of the speed value, and the value range of the cumulative frequency and the speed value are used as target constraints to establish a multi-objective optimization model.

[0070] As a feasible implementation method, the Min-Max method is selected to normalize the data, and the maximum cumulative frequency difference f1(x) and the minimum speed difference f2(x) data are mapped to [0,1]. The calculation formula is as follows:

[0071]

[0072] Where y o is the new sequence after standardization, x r is the original sequence, min 1≤r≤m {x r} is the minimum value in the original sequence, max 1≤r≤m {x r} is the maximum value in the original sequence;

[0073] In addition, in order to measure the effect of different objective function values ​​on the green wave design speed, it is necessary to establish a scoring function F(x) to unify the quantitative scores, that is, to use the quantitative scores; according to the optimization effects brought by different objective functions, the positive and negative weights are weighed to make the optimization direction consistent. The formula is as follows:

[0074] maxF(x)=ω1f1(x)+ω2f2(x)

[0075] Next, in order to make the driving speed of most vehicles concentrated in a smaller speed range, the following constraints are added to the multi-objective optimization model: the difference between the cumulative frequencies of two speed values ​​is at least greater than 50%, the difference between the two speed values ​​cannot exceed 5 km / h, the cumulative frequency value corresponding to a certain speed value must be between 15% and 85%, and a certain speed value must be between the 15% speed value and the 85% speed value:

[0076]

[0077]

[0078] S105. Perform iterative calculation of the global optimal solution of the multi-objective optimization model through a preset simulated annealing rule, and output a green wave design speed solution.

[0079] Specifically, the initial temperature parameters, temperature attenuation function and Markov chain length are determined by using the Metropolis criterion of the simulated degradation algorithm. Then, two speed values ​​are randomly selected and determined as the initial green wave design speed scheme through the preset initial vehicle speed value alternative speed interval. Based on the initial green wave design speed scheme, the corresponding first scoring function value is calculated.

[0080] Furthermore, based on the initial green wave design speed scheme, a new green wave design speed scheme is generated and the corresponding second scoring function value is obtained. If the second scoring function value is greater than the first scoring function value, the new green wave design speed scheme is received. Finally, according to the ratio of the initial temperature parameter to the changed temperature parameter, the new green wave design speed scheme is subjected to convergence processing related to the global optimal solution, and the green wave design speed scheme is output to complete the coordinated control of the green wave.

[0081] As a feasible implementation method, 1): first determine the Metropolis criterion of the simulated degradation algorithm, the treatment of the initial temperature parameters, the temperature attenuation function, and the Markov chain length and other parameters, the initial temperature T0 = 1000 ° C, the temperature is limited to the lower limit Temperature decay rate α = 0.98, number of iterations L = 100N, number of iterations of initial temperature K = 1;

[0082] 2): Randomly select speed values ​​x1 and x2 from the established initial vehicle speed value alternative speed interval as the initial green wave design speed scheme X, and calculate the scoring function value F(x) corresponding to the green wave design speed scheme at this time;

[0083] 3): At this temperature T, from k = 1 to L;

[0084] 4): For the new green wave design speed scheme Xnew generated by scheme X, if the relevant constraints are met, output the scoring function value F(x) of the new scheme at this timenew , otherwise continue to generate domain solutions;

[0085] 5): Get the increased difference ΔF = F(x) new -F(x), where F(x) is the scoring function;

[0086] 6): If ΔF>0, accept the new green wave design speed scheme Xnew, otherwise randomly generate a random number ε in the interval [0,1) to satisfy ε <exp((F(x) new -F(x)) / T0) is used as the condition to determine whether to accept the new green wave design speed plan Xnew;

[0087] 7): If K≤L, update the number of iterations K=K+1 and execute 4), otherwise jump to 8);

[0088] 8): The temperature T slowly decreases, T = αT0, return to 2) and repeat the above operation;

[0089] 9): If T0≤Tf (variable temperature parameter), the termination requirement is met, and the green wave design speed plan at this time is output as the global optimal solution, that is, the green wave design speed plan is output, and the algorithm ends. Otherwise, it jumps to 4) for loop calculation until the convergence condition is met to complete the coordinated control of the green wave.

[0090] In one implementation, an experiment was used to prove that the data sets used in this experiment are the checkpoint vehicle passing data and intersection geometry data from west to east at the intersection of Road ① and Road ②, and Road ① and Road ③. The calculated green wave design speed is selected for comparison with the actual green wave speed value currently implemented.

[0091] Through data preprocessing and screening of travel time without delay, it is found that vehicle travel time is basically concentrated between 20s and 50s.

[0092] The speed distribution characteristics of the vehicle are analyzed through the speed frequency distribution curve and the speed cumulative frequency distribution curve, and the fitting function equation of speed and cumulative frequency is obtained. As shown in the following table:

[0093]

[0094] The target optimization model of vehicle speed is constructed as follows:

[0095]

[0096] minf2(x)=0.5(x2-x1)

[0097] maxF(x)=0.5f1(x)-0.5f2(x)

[0098] Calibration algorithm initial parameter value: initial temperature T0 = 1000℃, temperature cut-off lower limit The temperature decay rate α=0.98, and the number of iterations L=1500.

[0099] The solution results are shown in the following table:

[0100]

[0101]

[0102] The green wave design speed of the section ① (① to ②) from west to east during the flat peak (14:00-16:00) is 52.2km / h to 54.4km / h. The actual green wave speed currently implemented is 53km / h, which is within the range of the green wave design speed calculated by this application, indicating that the research on the green wave design speed by the target optimization model is in line with reality, thus verifying the accuracy and feasibility of the multi-objective optimization model.

[0103] In addition, the present application also provides a green wave design speed calculation device based on bayonet data, such as Figure 2 As shown, the green wave design speed calculation device based on the bayonet data specifically includes:

[0104] At least one processor 201. And a memory 202 in communication with the at least one processor 201. The memory 202 stores instructions that can be executed by the at least one processor 201, so that the at least one processor 201 can execute:

[0105] Based on the license plate matching algorithm, the vehicle travel time between two adjacent intersections is determined;

[0106] Through the statistical time window of each period, the sample data of vehicle travel time is filtered to obtain the effective vehicle travel time;

[0107] Fitting the data between the relevant speed value and the cumulative frequency for the effective vehicle travel time to determine a fitting curve;

[0108] A multi-objective optimization model is established based on the maximum difference of the cumulative frequency and the minimum difference of the speed value in the fitting curve;

[0109] Through the preset simulated annealing rule, the global optimal solution of the multi-objective optimization model is iteratively calculated to output the green wave design speed plan.

[0110] The method of obtaining vehicle travel time data through checkpoint data in this application has the advantages of small investment, large sample size, and high data accuracy, and provides a corresponding checkpoint data processing method, which expands the traditional travel time survey method. By redefining the green wave design speed, a specific green wave design speed calculation process is proposed, which lays the foundation for the research of green wave coordinated control. It is also possible to screen out the delayed vehicle travel time data by analyzing the mechanism of vehicle delays, combining the box plot method with the statistical method; and establish a multi-objective optimization model based on the distribution characteristics of vehicle driving speeds; use the simulated annealing algorithm to solve the driving speed range of the vast majority of vehicles without delays; it provides a scientific and practical theoretical method for the calculation of the green wave design speed, which can effectively respond to the dynamic driving speed changes of vehicles on the road, reduce delay time and the number of stops, and is of great significance to improving the effect of green wave coordinated control.

[0111] Each embodiment in this application is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the device, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can be referred to the partial description of the method embodiment.

[0112] The above describes specific embodiments of the present application. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recorded in the claims can be performed in an order different from that in the embodiments and still achieve the desired results. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0113] The above is only an embodiment of the present application and is not intended to limit the present application. For those skilled in the art, the embodiments of the present application may have various changes and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the embodiments of the present application should be included in the scope of the claims of the present application.

Claims

1. A method for calculating green wave design speed based on bayonet data, characterized in that: The method comprises: Based on the license plate matching algorithm, the vehicle travel time between two adjacent intersections is determined; Through the statistical time window of each time period, the sample data of the vehicle travel time is filtered to obtain the effective vehicle travel time; Performing data fitting between relevant speed values ​​and cumulative frequency on the effective vehicle travel time to determine a fitting curve; A multi-objective optimization model is established based on the maximum difference of the cumulative frequency and the minimum difference of the speed value in the fitting curve; The multi-objective optimization model is iteratively calculated for the global optimal solution through a preset simulated annealing rule, and a green wave design speed solution is output.

2. The method for calculating green wave design speed based on bayonet data according to claim 1 is characterized in that: Based on the license plate matching algorithm, the vehicle travel time between two adjacent intersections is determined, including: Based on the maximum duration of the full red signal phase of two adjacent intersections, a screening time threshold is determined; Performing screening processing on all acquired driving record data at a single intersection, and grouping all driving record data by license plate number; performing time series processing on all grouped driving record data; Calculating the data time difference between the same license plate data in the single intersection; By using the screening time threshold, duplicate driving record data in all the driving record data are eliminated to obtain valid driving record data; Marking abnormal vehicle license plate data in the valid driving record data as abnormal, and deleting the abnormal vehicle license plate data; Based on the location information of two adjacent intersections, and through the license plate matching algorithm, the valid driving record data is subjected to screening processing for the same license plate data to obtain the same license plate data under different time stamps; Based on the same license plate data at different time stamps, the data time difference is determined as the vehicle travel time of the same vehicle.

3. The method for calculating green wave design speed based on bayonet data according to claim 1 is characterized in that: Through the statistical time window of each time period, the sample data of the vehicle travel time is filtered to obtain the effective vehicle travel time, which specifically includes: The vehicle traffic period between two adjacent intersections is divided into sections to obtain multiple time periods; wherein the multiple time periods include: a morning peak period, an evening peak period, and an off-peak period; Based on the box plot, the vehicle travel time in each time period is screened within the numerical boundary to obtain the first vehicle travel time; wherein the parameters in the box plot include: lower quartile value, upper quartile value, upper and lower quartile difference, upper boundary and lower boundary; Performing sample aggregation processing on the first vehicle travel time to obtain a first travel sample set; Based on the first trip sample set, determine the mean and standard deviation of the first vehicle travel time; and filter out vehicle travel samples in the first trip sample set that are beyond two times the standard deviation of the mean to obtain a second trip sample set; performing absolute deviation processing on the second trip sample set to obtain a third trip sample set; The vehicle travel time of the third travel sample in the statistical time window of each time period is determined as the effective vehicle travel time.

4. The method for calculating green wave design speed based on bayonet data according to claim 3 is characterized in that: The second trip sample set is subjected to absolute deviation processing to obtain a third trip sample set, which specifically includes: Sorting the travel time samples in the second travel sample set to determine a median value; Calculate the absolute deviation between each travel time sample in the second travel sample set and the median value; According to the absolute deviation, and taking the median as the dividing line, respectively obtaining the first half average absolute deviation and the second half average absolute deviation of the second trip sample set; The second trip sample set is threshold filtered by using the median value, the first half mean absolute deviation, and the second half mean absolute deviation to obtain the third trip sample set.

5. The method for calculating green wave design speed based on bayonet data according to claim 1 is characterized in that: Performing data fitting between the relevant speed value and the cumulative frequency on the effective vehicle travel time to determine a fitting curve specifically includes: Determining an effective vehicle travel speed based on a road section length parameter between two adjacent intersections and the effective vehicle travel time; The effective vehicle driving speeds are grouped to obtain a driving speed distribution table; and based on the driving speed distribution table, the effective vehicle driving speeds are divided into intervals to obtain a vehicle speed frequency distribution histogram; Based on the vehicle speed frequency distribution histogram, determining a speed cumulative frequency distribution curve related to the effective vehicle travel time; All speed values ​​in the effective vehicle running speed are subjected to curve fitting processing with the speed cumulative frequency in the speed cumulative frequency distribution curve to obtain the fitting curve.

6. The method for calculating green wave design speed based on bayonet data according to claim 1 is characterized in that: Before establishing a multi-objective optimization model based on the maximum difference in cumulative frequency and the minimum difference in speed value in the fitting curve, the method further includes: according to Obtain the maximum cumulative frequency difference maxf1(x); wherein x0 is a specific value in the fitting curve; x1 is a first speed value in the fitting curve; x2 is a second speed value in the fitting curve; ω1 is a first function coefficient related to the maximum cumulative frequency difference; A max is the maximum value in the fitting curve; min is the minimum value in the fitting curve; s and h are the curve fitting parameters of the fitting curve; According to minf2(x)=ω2(x2-x1), the minimum speed value difference minf2(x) is obtained; wherein ω2 is the second function coefficient related to the minimum speed value difference.

7. The method for calculating green wave design speed based on bayonet data according to claim 6 is characterized in that: Based on the maximum difference of the cumulative frequency and the minimum difference of the speed value in the fitting curve, a multi-objective optimization model is established, which specifically includes: Performing data mapping processing on the maximum difference of the accumulated frequency and the minimum difference of the speed value to obtain a normalized sequence; Through the preset scoring function, the maximum difference of the cumulative frequency and the minimum difference of the speed value in the normalized sequence are weighted and optimized to obtain a quantitative score; The maximum difference of the cumulative frequency, the minimum difference of the speed value, and the value range of the cumulative frequency and the speed value are used as target constraints to establish the multi-objective optimization model.

8. The method for calculating green wave design speed based on bayonet data according to claim 1 is characterized in that: Before the multi-objective optimization model is iteratively calculated for the global optimal solution by using the preset simulated annealing rule and the green wave design speed scheme is output, the following steps are specifically included: Based on the Metropolis criterion of the simulated degradation algorithm, the initial temperature parameters, temperature decay function and Markov chain length are determined; Through the preset initial vehicle speed value alternative speed interval, two speed values ​​are randomly selected and determined as the initial green wave design speed scheme; Based on the initial green wave design speed scheme, a corresponding first scoring function value is calculated.

9. The method for calculating green wave design speed based on bayonet data according to claim 8, characterized in that: Through the preset simulated annealing rule, the multi-objective optimization model is iteratively calculated for the global optimal solution, and the green wave design speed scheme is output, which specifically includes: Based on the initial green wave design speed scheme, a new green wave design speed scheme is generated and a corresponding second scoring function value is obtained; If the second scoring function value is greater than the first scoring function value, receiving the new green wave design speed solution; According to the ratio of the initial temperature parameter to the changed temperature parameter, the new green wave design speed scheme is subjected to convergence processing related to the global optimal solution, and the green wave design speed scheme is output to complete the coordinated control of the green wave.

10. A green wave design speed calculation device based on bayonet data, characterized in that: The device comprises: at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores instructions that can be executed by the at least one processor, so that the at least one processor can execute the green wave design speed calculation method based on bayonet data according to any one of claims 1 to 9.