A real-time dynamic variable lane and trunk green wave collaborative optimization method

Abstract

The present invention discloses a method for coordinated optimization of real-time dynamic variable lanes and trunk green waves, which relates to traffic green wave control technology; it solves the problem of delayed response and strategy failure in emergency scenarios of variable lane and green wave coordinated optimization based on historical data of target intersections in the prior art. The present invention predicts the flow of the target intersection based on the flow data of the first two cycles of the upstream intersection, and determines the direction of the variable lane according to the saturation of the lane direction corresponding to the target intersection. The signal phase timing within the signal cycle is optimized with the goal of minimizing the total delay of the import lane group of the key intersection, thereby realizing coordinated control of the variable lane and traffic green wave; it is related to the influence of the upstream intersection on the target intersection as a whole, and the control strategy of the target intersection can be quickly adjusted according to the dynamic changes of the upstream intersection, which is more flexible and can cope with data mutations caused by emergency situations, thereby improving the use value of the control strategy.

Description

Technical Field

[0001] The present invention relates to a traffic green wave control technology, in particular to a real-time dynamically variable lane and trunk green wave collaborative optimization method. Background Art

[0002] With the acceleration of urbanization and the continuous growth of traffic demand, urban traffic congestion is becoming more and more serious. The traditional red and green lights with fixed time settings and fixed direction lanes can no longer meet the current traffic conditions.

[0003] Dynamic variable lane technology and trunk green wave technology have been proposed and applied in the prior art, which effectively alleviated traffic congestion; however, the existing traffic arterial control measures are basically based on the long-term historical data of the target intersection, and variable lanes and green wave control are performed on the target intersection. This method can achieve good results for regular traffic conditions, but once extreme weather conditions or traffic accidents occur, causing the traffic flow on some sections to disrupt the regularity, the optimization control scheme based on long-term historical data will be unable to cope with it, and even cause greater traffic congestion; for this reason, we propose a real-time dynamic variable lane and trunk green wave collaborative optimization method. Summary of the invention

[0004] The purpose of the present invention is to provide a real-time dynamic variable lane and trunk green wave collaborative optimization method to solve the problem that the existing variable lane and green wave control cannot make real-time response according to irregular scenes, the reaction is slow, causing traffic congestion and control strategy failure.

[0005] The present invention can be implemented by the following technical solution: A real-time dynamic variable lane and trunk green wave collaborative optimization method, comprising the following steps:

[0006] Step 1: Collect traffic flow data of trunk intersections based on traffic flow detectors, including traffic flow volume, flow direction, vehicle speed and queue length at each intersection;

[0007] Step 2: Take the traffic flow data of the target intersection and its two upstream intersections in the first two cycles as input, and the traffic flow data of the target intersection in the next cycle as output, build a traffic flow prediction model for the target intersection and use support vector regression SVR for training;

[0008] Step 3: Calculate the predicted saturation of different directions of the intersection based on the predicted intersection flow data, and determine the variable lane direction of the next optimization cycle according to the predicted saturation;

[0009] Step 4: Use the Webster model to calculate the phase cycle duration of each intersection in the trunk green wave line, and select the intersection with the largest cycle duration as the key intersection for trunk green wave control;

[0010] Step 5: Optimize the signal phase timing of the green wave key intersections on the trunk line according to the direction of the variable lanes, so as to minimize the total delay of vehicles in all entrance lane groups of the key intersections;

[0011] Step 6: Solve the optimal green wave bandwidth of non-critical intersections based on the MAXBAND model, and obtain the absolute phase difference between non-critical intersections and critical intersections;

[0012] Step 7: Accumulate the optimized signal phase cycle length. If the accumulated value exceeds the optimization cycle length, enter the next optimization cycle; the optimization cycle length is consistent with the cycle length of the data selected in step 2;

[0013] Step 8: Decide whether to maintain the current control strategy based on whether the variable lane directions in adjacent optimization cycles are consistent. At the same time, decide how to correct the green light duration of the terminal signal phase in the current optimization cycle based on whether the terminal signal phase direction in the current optimization cycle is consistent with the variable lane direction in the next optimization cycle.

[0014] Step 9: Monitor and provide feedback on the implementation effect of the variable lane and trunk green wave coordinated optimization plan, and dynamically optimize and adjust the plan based on the feedback results.

[0015] A further technical improvement of the present invention is that the calculation process of predicting saturation in step 3 is:

[0016]

[0017] Among them, α s and α l are the saturation of the straight and left turning directions, q s and q l are the straight and left turn flow rates, n s and n l are the number of lanes for straight and left turns, Q s and Q l They are the single lane capacities for straight going and left turning directions respectively.

[0018] A further technical improvement of the present invention is that the specific method for determining the variable lane direction of the next optimization period according to the predicted saturation is:

[0019] If α s ≥α l And α s ≥0.9, or α s ≥αl And α l ≤0.9 The direction of the variable lane is the straight direction;

[0020] If α s <α l And α l >0.9, or α s >α l And α s <0.9 The variable lane direction is left turn.

[0021] A further technical improvement of the present invention is that the process of defining the minimum total delay of all vehicle lane groups of the critical intersection in step 5 is as follows:

[0022]

[0023] d im =d im1 ×PF+d im2 +d im3

[0024]

[0025] d im2 =900T(α im -1)

[0026]

[0027] Among them, d i is the average delay of vehicles at the i-th intersection, d im ,q im are the average vehicle delay and flow rate of the mth lane group at the ith intersection, d im1 d im2 d im3 are the uniform delay, incremental delay, and initial queue delay of vehicles in the mth lane group of the ith intersection, respectively; PF is the adjustment parameter for uniform control delay; im is the green signal ratio of the mth lane group at the ith intersection, α im is the saturation of the mth lane group at the ith intersection, T is the analysis duration, C i That is, the signal phase period of the i-th intersection;

[0028] K is the incremental delay correction coefficient of induction control; I is the incremental delay correction coefficient; C im is the capacity of the mth lane group at the ith intersection; a lane group is a combination of intersection lanes with the same flow direction.

[0029] A further technical improvement of the present invention is that the constraints for the optimization goal of minimizing the total delay of vehicles in all entrance lane groups of the key intersection include:

[0030] Cycle duration constraint: C min ≤C i ≤C max ;

[0031] C min is the minimum cycle length, C max is the maximum cycle duration;

[0032] Green light duration constraint: g min ≤g i ≤g max ;

[0033] g min is the minimum green light duration, g max It is the maximum green light duration.

[0034] The further technical improvement of the present invention is that the specific steps of solving the optimal green wave bandwidth based on the MAXBAND model in step 6 and obtaining the absolute phase difference between the uplink and downlink directions of intersection i and the key intersection include:

[0035] maxb=b'

[0036]

[0037] Among them, r i 、r i '、t i ,t i '、l i and l i ' are known quantities, which need to be set according to the actual traffic conditions at the main intersection; b i and b′ i are the green wave bandwidths in the uplink and downlink directions of intersection i, r i and r i ' are the red light durations for the up and down directions of intersection i, t i and t i ' are the travel time of the vehicle from intersection i to intersection i+1 in the upward and downward directions, respectively, l i and l i ' are the left turn phase durations in the up and down directions of intersection i respectively;

[0038] w i 、w i '、o i , o i '、z i 、z i',δ i , δ i ' are the variables to be solved, w i and w i ' is the time from the end of the red light to the center line of the green wave band in the up and down directions of the intersection i, respectively, i and i ' are the absolute phase differences in the up and down directions of intersection i, respectively, z i and z i ' are integer variables for the up and down directions of intersection i, respectively, δ i , δ i ' is a variable of 0 or 1. When it takes a value of 0, it means that the left turn phase of the up / down direction of intersection i is delayed. When it takes a value of 1, it means that the left turn phase of the up / down direction of intersection i is advanced. Thus (δ i ,δ i ') form four combination relationships (0,0), (0,1), (1,0), (1,1);

[0039] The calculation formula for the absolute phase difference between the up and down directions of intersection i and the key intersection is:

[0040]

[0041] φ i and φ i ' are the absolute phase differences in the up and down directions of intersection i, respectively.

[0042] A further technical improvement of the present invention is that the method of switching the control strategy according to the consistency of the variable lane directions includes:

[0043] (1) If the variable lane direction in the current optimization cycle is the same as that in the next optimization cycle, the current variable lane and trunk green wave control strategy is maintained;

[0044] (2)① If they are not the same, and the end signal phase direction of the current optimization cycle is consistent with the variable lane direction of the next optimization cycle, the green light duration of the end signal phase of the optimization cycle is extended;

[0045] ② If they are not the same, and the end signal phase direction of the current optimization cycle is inconsistent with the variable lane direction of the next optimization cycle, the green light duration of the end signal phase of the current optimization cycle will be shortened.

[0046] A further technical improvement of the present invention is that the extension time is defined as the time to clear the queue length of the variable lane in the current cycle, that is,

[0047] in, is the average saturated headway, L qis the distance between the variable lane guide sign and the entrance lane stop line, is the average headway between vehicles;

[0048] The shortened time is set to a manually set value of 5 or 10 seconds.

[0049] A further technical improvement of the present invention is that the feedback result is obtained based on the vehicle queue length, vehicle delay, and travel time as evaluation indicators, thereby adjusting the signal period, green-to-signal ratio, and phase difference of the trunk green wave intersection.

[0050] Compared with the prior art, the present invention has the following beneficial effects:

[0051] The present invention predicts the flow of the target intersection based on the flow data of the first two cycles of the upstream intersection, and determines the direction of the variable lane according to the saturation of the corresponding lane direction of the target intersection, and then optimizes the signal phase timing within the signal cycle with the goal of minimizing the total delay of the import lane group of the key intersection, thereby realizing the coordinated control of the variable lane and the traffic green wave; and decides whether to maintain the current control strategy based on whether the variable lane directions in adjacent optimization cycles are consistent, and at the same time, decides on the correction method of the green light duration of the terminal signal phase of the current optimization cycle based on the consistency between the terminal signal phase direction of the current optimization cycle and the variable lane direction of the next optimization cycle; it is related to the influence of the upstream intersection on the target intersection as a whole, and can quickly adjust the control strategy of the target intersection according to the dynamic changes of the upstream intersection, which is more flexible and can cope with data mutations caused by emergency situations, thereby improving the use value of the control strategy. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to facilitate understanding by those skilled in the art, the present invention is further described below with reference to the accompanying drawings.

[0053] Figure 1 The present invention is a schematic diagram of the method execution flow. DETAILED DESCRIPTION

[0054] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the specific implementation methods, structures, features and effects of the present invention are described in detail below in conjunction with the accompanying drawings and preferred embodiments.

[0055] See also Figure 1 As shown, a real-time dynamic variable lane and trunk green wave collaborative optimization method, the method specifically includes the following steps:

[0056] Step 1: Collect traffic flow data at arterial intersections

[0057] Based on the radar, video and other traffic flow detectors deployed at each intersection, traffic flow data such as traffic volume, flow direction, vehicle speed and queue length at each intersection of the trunk line are collected;

[0058] Step 2: Predict traffic flow at arterial intersections

[0059] For a target intersection, the traffic data of the previous two cycles and the traffic data of the previous two cycles of the two intersections upstream of the target intersection are selected from the historical traffic data as the input of the prediction model, and the support vector regression SVR is used to predict the traffic flow data of the target intersection in the next cycle. The cycle length can be set to 15 minutes or 30 minutes.

[0060] Specifically, the prediction steps are as follows:

[0061] S21. Input the traffic flow sample set and divide it

[0062] Define the sample set as N = {(x1,y1),(x2,y2),...,(x n ,y n )}, where x i = {q i-2 ,q i-1 ,q i} t-2:t ,y i =q i t:t+1 ; and q i t-2:t Represents the flow data of the ith trunk intersection from the t-2th period to the tth period;

[0063] The above sample set is randomly divided into training set and test set in a ratio of 7:3;

[0064] S22. Use support vector regression SVR to train the prediction model

[0065] Find a linear function f(x) = (w, x) + b, so that the error between the function value in the training sample and the expected value is no greater than a given value ε; where w is the weight vector and b is the threshold. Add insensitive loss coefficient ε and slack variable ζ i and i *, then SVR can be summarized as the following regression problem:

[0066]

[0067] Among them, A is the penalty function.

[0068] Introducing Lagrange multipliers a and a*, the above optimization problem is transformed according to the duality theory:

[0069]

[0070]

[0071] Solve the planning problem to get a, and then calculate w:

[0072]

[0073] The deviation b is calculated using the KKT Kuhn-Tucker condition:

[0074]

[0075] The final regression function is:

[0076]

[0077] S23. Input the time-series historical traffic data of the trunk intersection to be predicted, and predict the future traffic of the intersection based on the trained model.

[0078] Step 3: Optimize the configuration of variable lanes

[0079] Based on the predicted traffic flow data of each intersection on the trunk line, the predicted saturation of each intersection in different directions (straight / left turn) in the next optimization period is calculated. The calculation formula is as follows:

[0080]

[0081] α s and α l are the saturation of the straight and left turning directions, q s and q l are the straight and left turn flow rates, n s and n l are the number of lanes for straight and left turns, Q s and Q l They are the single lane capacities for straight going and left turning directions respectively.

[0082] According to the saturation of the straight / left turn direction, the variable lane direction of the next optimization cycle is determined:

[0083] If α s ≥α l And α s ≥0.9, or α s ≥α l And α l ≤0.9 The direction of the variable lane is the straight direction;

[0084] If α s ≤α l And α l ≥0.9, or αs ≥α l And α s ≤0.9 The variable lane direction is left turn.

[0085] Step 4: Determine the key intersections of the trunk green wave

[0086] The Webster model is used to calculate the phase cycle duration C of each intersection in the trunk green wave line, and the intersection with the largest cycle duration is selected as the key intersection for trunk green wave control.

[0087] The calculation formula of the Webster model is:

[0088]

[0089] Where: L i Y is the loss time of the i-th intersection in a single signal cycle, such as the vehicle start loss time; i It is the ratio of the predicted flow value of the next optimization cycle of the i-th intersection to the saturated flow; where the saturated flow is the design flow of the corresponding road intersection in unit time.

[0090] Step 5: Collaborative optimization of variable lanes and green wave signal control at key intersections

[0091] An optimization cycle generally includes multiple signal phase optimization cycles. According to the variable lane direction of the next optimization cycle, the signal phase timing of the green wave key intersection of the trunk line is optimized. The optimization goal is to minimize the total vehicle delay of all entrance lane groups at the key intersection, that is:

[0092]

[0093] d im =d im1 ×PF+d im2 +d im3

[0094]

[0095] d im2 =900T(α im -1)

[0096]

[0097] Among them, d i is the average delay of vehicles at the i-th intersection, d im ,q im are the average vehicle delay and flow rate of the mth lane group at the ith intersection, d im1 d im2 d im3are the uniform delay, incremental delay, and initial queue delay of vehicles in the mth lane group of the ith intersection, respectively; PF is the adjustment parameter for uniform control delay; im is the green signal ratio of the mth lane group at the ith intersection, α im is the saturation of the mth lane group at the ith intersection, and T is the analysis duration, which is generally 15 minutes.

[0098] K is the incremental delay correction coefficient of induction control, which is generally taken as 0.5; I is the incremental delay correction coefficient, which is generally taken as 1.0; C im is the capacity of the mth lane group at the ith intersection. A lane group is a combination of lanes at an intersection with the same flow direction.

[0099] The constraints include signal cycle duration constraints and green light duration constraints.

[0100] Signal cycle duration constraint: C min ≤C i ≤C max ; C min is the minimum cycle length, C max is the maximum cycle duration.

[0101] Green light duration constraint: g min ≤g i ≤g max ; g min is the minimum green light duration, usually 15s, g max It is the maximum green light duration, usually 60s.

[0102] Intelligent algorithms such as genetic algorithm or particle swarm algorithm are used to solve the optimal signal cycle and green light duration of key intersections of the main green wave.

[0103] Step 6: Collaborative optimization of variable lanes and green wave signal control at non-critical intersections

[0104] The signal cycle duration of the trunk green wave non-critical intersections is consistent with that of the critical intersections. Based on the MAXBAND model, the maximum trunk green wave bandwidth is taken as the optimization goal to solve the optimal green wave bandwidth of adjacent intersections, and then the phase difference between non-critical intersections and critical intersections is obtained.

[0105] The specific steps for solving the green wave speed and bandwidth based on the MAXBAND model are as follows:

[0106] maxb=b'

[0107]

[0108] Among them, r i 、r i '、t i ,ti '、l i and l i ' are known quantities and need to be set according to the actual traffic conditions at the main intersection. i and b′ i are the green wave bandwidths in the uplink and downlink directions of intersection i, r i and r i ' are the red light durations for the up and down directions of intersection i, t i and t i ' are the travel time of the vehicle from intersection i to intersection i+1 in the upward and downward directions, respectively, l i and l i ' are the left turn phase durations in the upstream and downstream directions of intersection i respectively.

[0109] w i 、w i ',o i , o i '、z i 、z i ',δ i , δ i ' are the variables to be solved, w i and w i ' is the time from the end of the red light to the center line of the green wave band in the up and down directions of the intersection i, respectively, i and i ' are the absolute phase differences in the up and down directions of intersection i, respectively, z i and z i ' are integer variables for the up and down directions of intersection i, respectively, δ i , δ i ' is a 0 or 1 variable, δ i =0,δ i '=0 means that the left turn phase in both the up and down directions of intersection i is delayed, δ i =0,δ i ' = 1 means that the left turn phase in the up direction of intersection i is delayed, and the left turn phase in the down direction is advanced, δ i =1,δ i '=0 means that the left turn phase in the up direction of intersection i is ahead and the left turn phase in the down direction is behind, δ i =1,δ i '=1 means that the left turn phases in both the upstream and downstream directions of intersection i are advanced.

[0110] Substitute the solved variables into the phase difference calculation formula to obtain the absolute phase difference between the up and down directions of intersection i and the key intersection. The calculation formula is:

[0111]

[0112] φ i and φ i ' are the absolute phase differences in the upward and downward directions of intersection i, that is, the minimum time difference between the coordination phase of intersection i and the starting time of the key intersection.

[0113] Step 7: Switch the signal control scheme for adjacent optimization cycles

[0114] When a round of signal cycle optimization strategy within the optimization cycle is completed, the length of the signal cycle of this round is accumulated. When the total duration of the optimized signal cycle is greater than the optimization cycle length, switch to the next optimization cycle and optimize the variable lane direction and trunk green wave strategy of the next optimization cycle; it should be noted that the optimization cycle length is consistent with the cycle length set in the sample set constructed in step 2.

[0115] Step 8: Formulate a control strategy switching strategy for adjacent optimization cycles

[0116] After the coordinated optimization of the variable lanes and trunk green wave schemes in the current optimization cycle is completed, it is necessary to coordinate the variable lanes and trunk green wave control strategies in the adjacent optimization cycles.

[0117] If the variable lane direction of the current optimization cycle is the same as that of the next optimization cycle, the variable lane and trunk green wave control strategy of the current and next optimization cycles will not be modified;

[0118] On the contrary, if they are different, the time it takes to clear existing vehicles in the variable lanes during the current optimization period should be considered.

[0119] If the end signal phase direction of the current optimization cycle is consistent with the variable lane direction of the next optimization cycle, the green light duration of the end signal phase of the optimization cycle is extended by the variable lane queue length clearing time of the current signal cycle, that is:

[0120]

[0121] in, is the average saturated headway, L q is the distance between the variable lane guide sign and the entrance lane stop line, is the average headway.

[0122] If the end signal phase direction of the current optimization cycle is inconsistent with the variable lane direction of the next optimization cycle, the green light duration of the end signal phase of the current optimization cycle will be shortened, and the shortening time is generally 5s or 10s.

[0123] Step 9: Optimize and adjust the trunk green wave signal control scheme

[0124] Under variable lane conditions, the operation status and efficiency of trunk green wave traffic are continuously monitored, and the trunk green wave scheme is dynamically optimized and adjusted based on the optimized output of vehicle queue length, vehicle delay, travel time and other operation efficiency evaluation indicators; for example: under the current variable lane setting scenario, if the total queue length, total delay or travel time of vehicles at all intersections of several green wave lines increases, the signal period, green-to-signal ratio and phase difference of the trunk green wave intersection will be adjusted.

[0125] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Although the present invention has been disclosed as a preferred embodiment as above, it is not used to limit the present invention. Any technical personnel in this field can make some changes or modify the technical contents disclosed above into equivalent embodiments without departing from the scope of the technical solution of the present invention. However, any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present invention without departing from the content of the technical solution of the present invention still fall within the scope of the technical solution of the present invention.

Claims

1. A real-time dynamically variable lane and trunk green wave collaborative optimization method, characterized in that: The steps include: Step 1: Collect traffic flow data of trunk intersections based on traffic flow detectors, including traffic flow volume, flow direction, vehicle speed and queue length at each intersection; Step 2: Take the traffic flow data of the target intersection and its two upstream intersections in the first two cycles as input, and the traffic flow data of the target intersection in the next cycle as output, build a traffic flow prediction model for the target intersection and use support vector regression SVR for training; Step 3: Calculate the predicted saturation of different directions of the intersection based on the predicted intersection flow data, and determine the variable lane direction of the next optimization cycle according to the predicted saturation; Step 4: Use the Webster model to calculate the phase cycle duration of each intersection in the trunk green wave line, and select the intersection with the largest cycle duration as the key intersection for trunk green wave control; Step 5: Optimize the signal phase timing of the green wave key intersections on the trunk line according to the direction of the variable lanes, so as to minimize the total delay of vehicles in all entrance lane groups of the key intersections; Step 6: Solve the optimal green wave bandwidth of non-critical intersections based on the MAXBAND model, and obtain the absolute phase difference between non-critical intersections and critical intersections; Step 7: Accumulate the optimized signal phase cycle length. If the accumulated value exceeds the optimization cycle length, enter the next optimization cycle; the optimization cycle length is consistent with the cycle length of the data selected in step 2; Step 8: Decide whether to maintain the current control strategy based on whether the variable lane directions in adjacent optimization cycles are consistent. At the same time, decide how to correct the green light duration of the terminal signal phase in the current optimization cycle based on whether the terminal signal phase direction in the current optimization cycle is consistent with the variable lane direction in the next optimization cycle. Step 9: Monitor and provide feedback on the implementation effect of the variable lane and trunk green wave coordinated optimization plan, and dynamically optimize and adjust the plan based on the feedback results.

2. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 1 is characterized in that: The calculation process of the predicted saturation in step 3 is: Among them, α s and α l are the saturation of the straight and left turning directions, q s and q l are the straight and left turn flow rates, n s and n l are the number of lanes for straight and left turns, Q s and Q l They are the single lane capacities for straight going and left turning directions respectively.

3. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 2 is characterized in that: The specific method for determining the variable lane direction of the next optimization cycle according to the predicted saturation is: If α s ≥α l And α s ≥0.9, or α s ≥α l And α l ≤0.9 The direction of the variable lane is the straight direction; If α s <α l And α l >0.9, or α s >α l And α s <0.9 The variable lane direction is left turn.

4. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 1, characterized in that: The process of defining the minimum total vehicle delay of all entrance lane groups at the key intersection described in step 5 is as follows: d im =d im1 ×PF+d im2 +d im3 d im2 =900T(a im -1) Among them, d i is the average delay of vehicles at the i-th intersection, d im ,q im are the average vehicle delay and flow rate of the mth lane group at the ith intersection, d im1 d im2 d im3 are the uniform delay, incremental delay, and initial queue delay of vehicles in the mth lane group of the ith intersection, respectively; PF is the adjustment parameter for uniform control delay; im is the green signal ratio of the mth lane group at the ith intersection, α im is the saturation of the mth lane group at the ith intersection, T is the analysis duration, C i That is, the signal phase period of the i-th intersection; K is the incremental delay correction coefficient of induction control; I is the incremental delay correction coefficient; C im is the capacity of the mth lane group at the ith intersection; a lane group is a combination of intersection lanes with the same flow direction.

5. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 4 is characterized in that: The constraints for the optimization goal of minimizing the total vehicle delay of all entrance lane groups at the critical intersection include: Cycle duration constraint: C min ≤C i ≤C max ; C min is the minimum cycle length, C max is the maximum cycle duration; Green light duration constraint: g min ≤g i ≤g max ; g min is the minimum green light duration, g max It is the maximum green light duration.

6. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 1, characterized in that: In step 6, the specific steps of solving the optimal green wave bandwidth based on the MAXBAND model and obtaining the absolute phase difference between the uplink and downlink directions of intersection i and the key intersection include: maxb=b' Among them, r i 、r i '、t i ,t i '、l i and l i ' are known quantities, which need to be set according to the actual traffic conditions at the main intersection; b i and b i ' are the green wave bandwidths in the up and down directions of intersection i, r i and r i ' are the red light durations for the up and down directions of intersection i, t i and t i ' are the travel time of the vehicle from intersection i to intersection i+1 in the upward and downward directions, respectively, l i and l i ' are the left turn phase durations in the up and down directions of intersection i respectively; w i 、w i '、o i , o i '、z i 、z i ',δ i , δ i ' are the variables to be solved, w i and w i ' is the time from the end of the red light to the center line of the green wave band in the up and down directions of the intersection i, respectively, i and i ' are the absolute phase differences in the up and down directions of intersection i, respectively, z i and z i ' are integer variables for the up and down directions of intersection i, respectively, δ i , δ i ' is a variable of 0 or 1. When it takes a value of 0, it means that the left turn phase of the up / down direction of intersection i is delayed. When it takes a value of 1, it means that the left turn phase of the up / down direction of intersection i is advanced. Thus (δ i ,δ i ') form four combination relationships (0,0), (0,1), (1,0), (1,1); The calculation formula for the absolute phase difference between the up and down directions of intersection i and the key intersection is: φ i and φ i ' are the absolute phase differences in the up and down directions of intersection i, respectively.

7. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 1, characterized in that: The method of switching the control strategy according to the consistency of the variable lane direction includes: (1) If the variable lane direction in the current optimization cycle is the same as that in the next optimization cycle, the current variable lane and trunk green wave control strategy is maintained; (2)① If they are not the same, and the end signal phase direction of the current optimization cycle is consistent with the variable lane direction of the next optimization cycle, the green light duration of the end signal phase of the optimization cycle is extended; ② If they are not the same, and the end signal phase direction of the current optimization cycle is inconsistent with the variable lane direction of the next optimization cycle, the green light duration of the end signal phase of the current optimization cycle will be shortened.

8. The method for collaboratively optimizing real-time dynamically variable lanes and trunk green waves according to claim 7 is characterized in that: The extension time is defined as the time to clear the queue length of the variable lane in the current cycle, i.e. in, is the average saturated headway, L q is the distance between the variable lane guide sign and the entrance lane stop line, is the average headway between vehicles; The shortened time is set to a manually set value of 5 or 10 seconds.

9. The method for collaborative optimization of real-time dynamically variable lanes and trunk green waves according to claim 1, characterized in that: The feedback result is obtained based on the vehicle queue length, vehicle delay and travel time as evaluation indicators, so as to adjust the signal period, green-to-signal ratio and phase difference of the trunk green wave intersection.

Images