Single-intersection vehicle-road cooperative traffic control method for simulating traffic police command

Abstract

According to the single-intersection vehicle-road cooperative traffic control method for simulating traffic police command provided by the invention, a correlation coefficient between intersections is constructed by combining adjacent upstream and downstream intersections, a'traffic pressure index 'is established, a control scene is constrained, and related strategies are adopted to simulate traffic police handling measures to release traffic pressure, so that the control accuracy is ensured; different tolerance saturation adjustment control force can be set for each phase according to experience, road weight and the like, and the implementation controllability is high. The above method is only a preferred embodiment of the invention and is not used for limiting the invention, and for technicians in the field, the invention can have various changes and changes. In the spirit and principle of the invention, any modification, equivalent replacement, improvement and the like are included in the protection range of the invention.

Description

Technical Field

[0001] The invention belongs to the field of intelligent transportation systems and relates to a single-intersection traffic control method that simulates traffic police command in real time. Background Art

[0002] With the development of social economy and the Internet of Things, the number of cars in cities has increased rapidly, and the imbalance between traffic supply and demand has gradually intensified. However, urban traffic optimization has not developed synchronously, and the intensity of traffic maintenance and diversion work has continued to increase.

[0003] Traffic police have limited manpower. In order to meet the needs of traffic management, there is even room for "fake traffic police" (dummy people who look like traffic police) to take up their posts. Although "fake traffic police" cannot control people and vehicles like real traffic police, they can still play a certain deterrent role and help with daily traffic coordination and ensure the smooth progress of traffic maintenance work.

[0004] Although "fake traffic police" can make up for the shortcomings in some special environments and times, in reality, many "fake traffic police" often have problems with rough production technology and simple appearance design, which are very different from the appearance of real traffic police. While giving people warnings, they are more likely to scare people. When people face sudden fear in their hearts, they are likely to increase the probability of operating errors during driving, which in turn brings certain traffic safety hazards. Therefore, they are resisted by the majority of drivers.

[0005] In this situation, the intelligent transportation industry, which has developed with the help of new technologies such as big data and vehicle-road collaboration, has gradually become the focus of attention.

[0006] At present, there are many optimization methods for traffic signal timing in the prior art, and each optimization method has its own advantages and disadvantages. For example, the patent document with application number CN202011497656.0 discloses a real-time traffic signal control method and device based on deep reinforcement learning, which collects a large amount of real-time vehicle data from the road network and combines it with the idea of ​​reinforcement learning to obtain a traffic signal control scheme. This method needs to be trained based on a large amount of real-time vehicle sample environmental information to obtain a traffic signal control model, which does not take into account the particularity of the traffic police command scene, and there is not much data for training.

[0007] For example, the patent document with application number CN202110711932.7 discloses a single-intersection traffic signal timing optimization method based on the Monte Carlo algorithm, which uses the probability distribution characteristics of historical traffic flow to determine the optimal operating cycle, phase configuration, and the value range of the green light time of each phase. This method cannot adjust the plan in real time like the traffic police command, and the desired plan may be significantly different from the actual traffic flow, and it is not a real-time optimal solution.

[0008] For example, the patent document with application number CN202010927680.7 discloses a traffic signal control method and device, which uses the configured traffic signal control strategy to execute the corresponding traffic signal control scheme. This method also cannot adjust the scheme in real time like the traffic police command, and cannot meet the dynamic changes of traffic demand.

[0009] For example, the patent document with application number CN201910629489.1 discloses a single-intersection traffic signal control method, system, and device based on deep reinforcement learning, which trains the model and updates the network parameters through a large amount of data. This method also does not take into account the particularity of traffic police command scenarios, and there is not much special data for training.

[0010] For another example, the patent document with application number 201610971018.5 discloses a single-point signal control optimization method based on intersection vehicle passing records, which obtains vehicle passing records of the stop lines of the entrance lanes in all directions of the intersection, analyzes the green light utilization index of each phase, and shortens the green light time if there is a green light time surplus in each cycle. This method is suitable for adjusting the timing scheme under unsaturated conditions, but cannot be applied to saturated conditions, and has poor applicability. Summary of the invention

[0011] The purpose of the present invention is to provide a single-intersection vehicle-road cooperative traffic control method that simulates traffic police command. The single-point control method obtains the strong correlation coefficient and imbalance index between adjacent intersections, establishes a "traffic pressure index", constrains the control scene, and uses relevant strategies to simulate traffic police disposal measures to release traffic pressure, ensuring the accuracy of control; and can set different tolerance saturation adjustment control strength for each phase based on experience, road weight, etc. The present invention is further elaborated below.

[0012] A single-intersection vehicle-road cooperative traffic control method simulating traffic police command, the single-intersection vehicle-road cooperative traffic control method comprising:

[0013] S1. Monitoring real-time saturated flow S: By monitoring the real-time saturated flow, we can indirectly monitor the influencing factors that are difficult to measure directly at ordinary times, such as accidents, extreme weather and other special events or influencing factors that are difficult to collect; the real-time saturated flow S is reflected in the efficiency of the queued vehicles passing the stop line, specifically:

[0014]

[0015] in, is the average headway time of vehicles in the queue near the stop line;

[0016] S2. Calculate the traffic demand at a single intersection: the flow rate q at a green light 通过, and the queue flow q at the end of the green light 排队 The actual traffic demand during the period is obtained by adding them up, which is:

[0017] q 需求 =q 排队 +q 通过 ;

[0018] Among them, q 需求 is the traffic demand of the previous cycle;

[0019] S3. According to the correlation between the flow rate of this intersection and the storage space of the downstream section, a strong correlation coefficient δ is established. First, the Q value of the downstream storage space is obtained: Q = n 下 (Ll 排队 ) / l veh ; Then, the value of the strong correlation coefficient is obtained according to the downstream storage space value: δ = Q 转向 / Q;

[0020] Where Q is the downstream storage space, n 下 is the number of lanes in the downstream section, L is the length of the downstream section, l 排队 is the queue length of the downstream section, l veh is the saturated headway of the current intersection, δ is the strong correlation coefficient, Q 转向 Real-time vehicle count for steering;

[0021] When δ>=ε, it indicates that the downstream intersection is strongly associated with the current intersection. ε can be adjusted according to the actual situation. When ε=1, it indicates that there is a possibility of overflow in the downstream section.

[0022] S4. Establishing an imbalance index: Obtain the saturation ρ of this single intersection, establish an imbalance index β, determine whether there is an imbalance, and trigger real-time adjustment of the current phase, as follows:

[0023] S41, if the current saturation ρ 当前 Not the maximum, you can set the imbalance index β = (ρ max -ρ 当前 ) / ρ 当前 , and only in ρ 当前 <1,ρ max >1 and β>α, it is determined to be unbalanced, triggering real-time adjustment of the current phase and implementing early disconnection of the current phase;

[0024] Among them, ρ max is the maximum saturation value in other phases; α is a constant and can be set according to the actual situation; if the upstream section of the turn is particularly long, the storage space is large, and a larger queue length is allowed, α can be set larger accordingly; saturation ρ 当前 The calculation of is based on the ratio of the release time requirement of the queue flow at the current moment to the remaining green light time;

[0025] S42, if the current saturation ρ 当前 is the maximum, continue to release until the green light ends; the imbalance index β can be set to (ρ 当前 -ρ min ) / ρ min , and only in ρ 当前 >1,ρ min <1, β>α means imbalance is determined, triggering real-time adjustment of the current phase to extend the current phase by a green light step length γ; when it runs again to the point close to the end of the green light, it can continue to determine whether it is necessary to extend the green light step length again;

[0026] Among them, ρ min is the minimum saturation value in other phases, and γ is a constant, which can be set according to actual conditions;

[0027] S5. Establishing the "traffic pressure index" θ of the intersection: when θ is within the normal range, the traffic police's command is not required. When θ is close to or exceeds the normal range, the corresponding "traffic police's command" needs to be initiated as soon as possible or in advance to release or suppress the pressure explosion in advance. The value of the "traffic pressure index" θ is as follows: θ = max(maxδ, maxβ);

[0028] Among them, maxδ is the maximum value of the correlation coefficient among all turns in the real-time intersection; maxβ is the maximum value of the imbalance index among all phases in the previous cycle;

[0029] When θ>1, it indicates that there is a traffic pressure problem at the intersection;

[0030] S6. Simulate traffic police command:

[0031] S61. Adopting the vehicle-road collaboration method: relying on the corresponding electronic equipment installed at each intersection of the road, obtaining the judgment of cutting in and not driving according to the traffic light instructions, as follows:

[0032] S611: Determination of lane-cutting: Determination is made based on the lane-cutting capture area, lane-cutting speed conditions, and lane change, as follows:

[0033] Lane-cutting capture area: Define key areas of each lane, generally including solid lines, for which strict enforcement is applied; dotted lines and even dashed lines can also be added based on actual conditions;

[0034] Speed ​​threshold determination: The lane-cutting behavior is generally carried out at a slow speed, and the heading angle deviates from the normal driving angle;

[0035] Lane change behavior judgment: being in different lanes before and after the speed change within the area;

[0036] S612: The determination of not following the traffic light instructions is as follows:

[0037] Os <t r ≤5s, send a startup reminder;

[0038] When t g ≤3s, if the vehicle speed v<5km / h and there is no obstacle in front of the vehicle, the vehicle is judged not to be started and a "not driving according to traffic light instructions" warning is sent;

[0039] t g >3s, send a penalty notice for "not driving according to traffic light instructions";

[0040] Among them, t r is the time from the current direction to the next green light on, t g It is the running time after the preceding vehicle runs normally during the green light period in the current direction within the cycle;

[0041] S62: When judging whether a waiting area for a corresponding turn needs to be set up according to the static data and dynamic data of a single intersection, the specific conditions are:

[0042] when ρ 转向 When >1, a waiting area can be set up to improve traffic efficiency;

[0043] Among them, l 对向 is the distance between opposite stop lines, is the distance threshold, ρ 转向 is the saturation of the steering;

[0044] S63, adjusting the phase in real time according to the own imbalance index of the intersection and the storage volume of the downstream intersection, (1) if ρ 当前 If β>α is not the maximum, it is determined to be unbalanced, triggering real-time adjustment of the current phase and ending the current phase; (2) If ρ 当前 is the maximum, and the vehicle is released until the green light ends. If β>α, it is determined to be unbalanced, and the current phase is adjusted in real time, extending the current phase by a green light step length γ. γ should meet the constraint: γ <t max +t more ; When the system runs to the end of the green light again, it can continue to determine whether it is necessary to extend the green light step again until the release time reaches the maximum constraint;

[0045] The specific method is:

[0046] S631, the release time of this intersection cannot exceed the maximum release time t max :t max =t veh Q / n 上 ;

[0047] S632: When a turn is released at this intersection, the downstream related turn is also being released. Assuming that the saturation flow of the upstream and downstream related turns is the same, the release time t can be increased. more :t more =t 下 n 下 / n 上 ;

[0048] S633. In actual use, the situation where the upstream and downstream related steering saturation flows are different can be considered:

[0049] t more =t 下 n 下 t veh,上 / (n 上 t veh,下 );

[0050] Among them, t veh,上 is the upstream saturation headway, t veh,下 is the downstream saturation headway, t more Add value to the release time.

[0051] The present invention provides a single-intersection vehicle-road cooperative traffic control method that simulates traffic police command. It constructs a correlation coefficient between intersections in combination with adjacent upstream and downstream intersections, establishes a "traffic pressure index", constrains control scenarios, and adopts relevant strategies to simulate traffic police disposal measures to release traffic pressure, thereby ensuring the accuracy of control. Different tolerance saturation adjustment control strengths can be set for each phase based on experience, road weights, etc., and the implementation is highly controllable. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] Figure 1 Overall process. DETAILED DESCRIPTION

[0053] The present invention is described in detail below in conjunction with the accompanying drawings. The following embodiments do not constitute a limitation of the present invention.

[0054] The present invention proposes a single-intersection vehicle-road cooperative traffic control method simulating traffic police command, the method comprising the following steps:

[0055] S1, monitor the real-time saturation flow S;

[0056] The need to monitor real-time saturated flow is because the factors that affect traffic efficiency are not only timing and flow, but also related to special events such as accidents and extreme weather. By monitoring real-time saturated flow, we can indirectly monitor special events such as accidents and extreme weather or influencing factors that are difficult to collect; the real-time saturated flow S is reflected in the efficiency of queued vehicles when passing the stop line, specifically:

[0057]

[0058] in is the average headway time of vehicles in the queue near the stop line.

[0059] S2. Calculate the traffic demand at a single intersection: the flow rate q at a green light 通过 The queue flow q at the end of the green light 排队 By adding them together, we can get the real flow demand in the period;

[0060] At present, most of the traditional intelligent single-intersection signal control is based on the prediction of historical cycle traffic conditions or the estimation of the current cycle traffic flow to control the allocation of green signal ratio. However, the predicted or estimated traffic flow often has a certain gap with the actual situation, and this gap is often unbearable for some intersections. Even if the prediction results are corrected in combination with the traffic data of the day, they are often extremely inaccurate due to dynamic adjustment and release. This is also the reason why there are still a large number of intersections that need on-site command by traffic police today when traffic technology is constantly developing;

[0061] The on-site command of the traffic police is based on the real-time situation on the scene. At the right time, even if the light in a certain direction is red, the vehicles in that direction can be released. Moreover, when the road congestion is found to be aggravated, the vehicles in the direction with less pressure can be stopped to allow vehicles in the direction with more pressure to pass first. The corresponding behavior of the traffic police is that in the direction that needs to be released, the green light can be turned on in advance or even extended (divided into early green light and extended green light); corresponding to the direction with small traffic volume, the red light needs to be extended (divided into early red light and extended red light). Therefore, the adjustment of signal light timing needs to be realized by means of the vehicle-road cooperative system in a way that calculates and adjusts the signal timing in real time according to the real-time number of vehicles in demand;

[0062] If the green light duration is adjusted completely according to the number of vehicles in real-time demand, that is, it is degraded to the induction control mode, an extremely unreasonable situation will often occur, that is, unlimited release in the direction with many vehicles, which may not only cause too many vehicles to drive into the downstream, causing congestion in the downstream, but also cause too little release time in other directions at the current intersection, causing new congestion. Therefore, based on the similarity of short-term traffic flow in general, the timing plan can be evenly distributed according to the traffic demand of the previous cycle, so as to avoid extremely unreasonable situations caused by releasing completely according to vehicle flow induction;

[0063] This embodiment will avoid extremely unreasonable situations by calculating the traffic demand at the intersection, and can make a balanced distribution based on the traffic demand of the previous cycle, as follows:

[0064] q 需求 =q 排队 +q 通过 ;

[0065] Among them, q需求 is the traffic demand in the previous cycle, q 通过 is the flow rate of green light passing, q 排队 Queue traffic for when the green light ends.

[0066] S3, establish a strong correlation coefficient δ;

[0067] The need to establish a strong correlation coefficient is that the closer the distance between adjacent intersections is, the greater the impact of the flow of the current intersection on the downstream, that is, the greater the correlation; if the flow of the relevant turn at the current intersection exceeds the storage space of the downstream section, it constitutes a strong correlation, and it is necessary to constrain the cycle and start time of the current intersection or constrain the release time of the relevant turn at the current intersection;

[0068] Some ideal strong correlation scenarios can achieve the situation that the vehicle at the current intersection just passes the green light when it arrives at the downstream intersection by constraining the cycle and start time of the current intersection, so that there is no need for traffic police command. This is not what this embodiment needs to discuss. This embodiment mainly solves the scenario of strong correlation requiring on-site command by traffic police, and can only be solved by constraining the relevant turn release time of the current intersection; the details are as follows:

[0069] First, obtain the downstream storage space Q value: Q = n 下 (Ll 排队 ) / l veh ;

[0070] Then, the value of the strong correlation coefficient is obtained according to the downstream storage space value: δ = Q 转向 / Q;

[0071] Among them, n 下 is the number of lanes in the downstream section, L is the length of the downstream section, l 排队 is the queue length of the downstream section, l veh is the saturated headway of the current intersection, δ is the strong correlation coefficient, Q is the downstream storage space, Q 转向 Real-time vehicle count for steering;

[0072] When δ>=ε, it indicates that the downstream intersection is strongly associated with the current intersection. ε can be adjusted according to the actual situation;

[0073] When ε = 1, it indicates that the downstream section is about to overflow;

[0074] S4. Establish an imbalance index;

[0075] The balance index is established because the traffic police will not pay attention to the demand deviation between the previous and next cycles, but will only pay attention to whether the timing in the current cycle matches the current situation. Except for the obvious empty situation, it is necessary to further pursue efficiency and balance.

[0076] In this embodiment, the imbalance index β is established through the saturation ρ of a single intersection to determine whether there is an imbalance, so as to adjust the phase in real time; when the phase runs to half of the green light time, the situations include:

[0077] S41, if the current saturation ρ 当前 Not the maximum, you can set the imbalance index β = (ρ max -ρ 当前 ) / ρ 当前 , and only in ρ 当前 <1,ρ max >1, β>α means imbalance is determined, triggering real-time adjustment of the current phase and implementing early disconnection of the current phase;

[0078] Among them, ρ max is the maximum saturation value in other phases, α is a constant and can be set according to the actual situation; if the upstream section of the turn is particularly long, the storage space is large, and a larger queue length is allowed, α can be set larger accordingly; saturation ρ 当前 The calculation uses the ratio of the release time requirement of the queue flow at the current moment to the remaining green light time. This method needs to pay attention to the fact that the saturation calculation is meaningful only when the queue flow is greater than a certain value, that is, when the queue flow is small, it is not the pressure traffic direction that the traffic police need to pay attention to;

[0079] S42, if the current saturation ρ 当前 is the maximum, and the vehicle will continue to be released until the green light is approaching. The imbalance index β can be set to (ρ 当前 -ρ min ) / ρ min , and only in ρ 当前 >1, and ρ min <1, β>α is considered unbalanced, triggering the current phase to adjust in real time and extend the current phase by a green light step length γ; when it runs again to the point close to the end of the green light, it can continue to determine whether it is necessary to extend the green light step length again;

[0080] Among them, ρ min is the minimum saturation value in other phases, and γ is a constant, which can be set according to actual conditions.

[0081] S5, establishing a "traffic pressure index" θ of the intersection;

[0082] The work of traffic police is mainly to prevent and optimize the safety and efficiency of intersections during morning and evening peak hours. The efficiency is especially reflected in the prevention of tangling and overflow.

[0083] This embodiment establishes a "traffic pressure index" θ at the intersection to achieve early release or suppress the occurrence of pressure explosion;

[0084] When θ is within the normal range, there is no need for traffic police command. When θ is close to or exceeds the normal range, it is necessary to start the corresponding "traffic police command" as soon as possible or in advance to release or suppress the pressure explosion in advance. The "traffic pressure index" θ is as follows:

[0085] θ=max(maxδ,maxβ);

[0086] Among them, maxδ is the maximum value of the correlation coefficient among all turns in the real-time intersection; maxβ is the maximum value of the imbalance index among all phases in the previous cycle;

[0087] When θ>1, it indicates that there is a traffic pressure problem at the intersection.

[0088] S6. Simulate traffic police command:

[0089] When θ>1, it indicates that there is a serious overflow or imbalance problem at the intersection, and the traffic police need to intervene in the management. Some traffic police will guide vehicles into non-existent straight-ahead and turning waiting areas according to the traffic conditions, prolonging the travel time in the direction of large traffic flow and shortening the travel time of small traffic flow; and the traditional intelligent signal system of signal timing often predicts the traffic conditions of the current cycle based on historical traffic data, and this prediction is often not accurate enough;

[0090] This embodiment adopts a single intersection vehicle-road cooperative traffic control method that simulates traffic police command, which can simulate traffic police command. The method is as follows:

[0091] S61. Adopting the vehicle-road collaboration method: relying on the corresponding electronic equipment installed at each intersection of the road, obtaining the judgment of cutting in and not driving according to the traffic light instructions, as follows:

[0092] S611: Determination of lane-cutting: Determination is made based on the lane-cutting capture area, lane-cutting speed conditions, and lane change, as follows:

[0093] Lane-cutting capture area: Define key areas of each lane, generally including solid lines, for which strict enforcement is applied; dotted lines and even dashed lines can also be added based on actual conditions;

[0094] Speed ​​threshold determination: The lane-cutting behavior is generally carried out at a slow speed, and the heading angle deviates from the normal driving angle;

[0095] Lane change behavior judgment: being in different lanes before and after the speed change within the area;

[0096] S612: The determination of not following the traffic light instructions is as follows:

[0097] Os <t r ≤5s, send a startup reminder;

[0098] When t g ≤3s, if the vehicle speed v<5km / h and there is no obstacle in front of the vehicle, the vehicle is judged not to be started and a "not driving according to traffic light instructions" warning is sent;

[0099] t g >3s, send a penalty notice for "not driving according to traffic light instructions";

[0100] Among them, t r is the time from the current direction to the next green light on, t g It is the running time after the preceding vehicle runs normally during the green light in the current direction within the cycle.

[0101] S62. Determine whether it is necessary to set up a waiting area for the corresponding turn based on the static data and dynamic data of the single intersection, so as to effectively improve the road traffic capacity, and the conditions are as follows:

[0102] ρ 转向 >1 hour;

[0103] Among them, l 对向 is the distance between opposite stop lines, is the distance threshold, ρ 转向 is the saturation of the steering;

[0104] S63, adjusting the phase in real time according to the own imbalance index of the intersection and the storage volume of the downstream intersection, (1) if ρ 当前 If β>α is not the maximum, it is determined to be unbalanced, triggering real-time adjustment of the current phase and ending the current phase; (2) If ρ 当前 is the maximum, and the vehicle is released until the green light ends. If β>α, it is determined to be unbalanced, and the current phase is adjusted in real time, extending the current phase by a green light step length γ. γ should meet the constraint: γ <t max +t more ; When the system runs to the end of the green light again, it can continue to determine whether it is necessary to extend the green light step again until the release time reaches the maximum constraint; specifically:

[0105] S631, the release time of this intersection cannot exceed the maximum release time t max :t max =t veh Q / n 上 ;

[0106] S632: When a turn is released at this intersection, the downstream related turn is also being released. Assuming that the saturation flow of the upstream and downstream related turns is the same, the release time t can be increased. more :t more =t 下 n 下 / n 上 ;

[0107] S633. In actual use, the situation where the upstream and downstream related steering saturation flows are different can be considered:

[0108] t more =t 下 n 下 t veh,上 / (n 上 t veh,下 );

[0109] Among them, t veh,上 is the upstream saturation headway, t veh,下 is the downstream saturation headway, t more Add value to the release time.

Claims

1. A single-intersection vehicle-road cooperative traffic control method simulating traffic police command, characterized in that: The single intersection traffic control method comprises: S1. Monitor real-time saturated flow S: By monitoring real-time saturated flow, we can indirectly monitor special events such as accidents and extreme weather or influencing factors that are difficult to collect; real-time saturated flow S is reflected in the efficiency of queued vehicles passing through the stop line, and the calculation formula is: in, is the average headway time of vehicles in the queue near the stop line; S2. Calculate the traffic demand at a single intersection: q 需求 =q 排队 +q 通过 , where q 需求 is the traffic demand in the previous cycle, q 通过 is the flow rate of green light passing, q 排队 To queue traffic when the green light ends; S3. According to the relationship between the flow rate of this intersection and the storage space of the downstream section, a strong correlation coefficient δ is established. First, the downstream storage space Q value is obtained: Q = n 下 (Ll 排队 ) / l veh ; Then, the value of the strong correlation coefficient δ is obtained according to the downstream storage space value: δ = Q 转向 / Q; Among them, n 下 is the number of lanes in the downstream section, L is the length of the downstream section, l 排队 is the queue length of the downstream section, l veh is the saturated headway of the current intersection, δ is the strong correlation coefficient, Q is the downstream storage space, Q 转向 Real-time vehicle count for steering; S4, establishing an imbalance index β: obtaining the saturation ρ of each phase of the intersection, establishing an imbalance index β, and thus determining whether there is an imbalance; S5. Establish the "traffic pressure index" θ of the intersection: θ=max(maxδ, maxβ); Among them, maxδ is the maximum value of the correlation coefficient among all turns in the real-time intersection; maxβ is the maximum value of the imbalance index among all phases in the previous cycle; When θ is within the normal range, there is no need for traffic police command. When θ is close to or exceeds the normal range, it is necessary to start the corresponding "traffic police command" as soon as possible or in advance to release the pressure in advance or suppress the occurrence of pressure explosion. When θ>1, it indicates that there is a traffic pressure problem at the intersection; S6. Simulate traffic police command: S61. Adopting the vehicle-road collaboration method: relying on the corresponding electronic equipment installed at each intersection of the road, obtaining the judgment of cutting in and not driving according to the traffic light instructions, as follows: S611, judging whether lane-cutting occurs: judging based on the lane-cutting capture area, lane-cutting speed conditions, and whether lane change occurs; S612: The determination of not following the traffic light instructions is as follows: 0s<t r ≤5s, send a startup reminder; When t g ≤3s, if the vehicle speed v is less than 5km / h and there is no obstacle in front of the vehicle, the vehicle is judged not to be started and a "not driving according to traffic light instructions" warning is sent; t g >3s, send a penalty notice for "not driving according to traffic light instructions"; Among them, t r is the time from the current direction to the next green light on, t g It is the running time after the preceding vehicle runs normally during the green light period in the current direction within the cycle; S62: Determine whether to set a waiting area for a corresponding turn based on the static data and dynamic data of the single intersection. The specific conditions are: When l 对向 >δ,ρ 转向 When the ratio is greater than 1, a waiting area can be set up to improve traffic efficiency; Among them, l 对向 is the distance between opposite stop lines, is the distance threshold, ρ 转向 is the saturation of the steering; S63, adjusting the phase in real time according to the imbalance index of the intersection and the storage volume of the downstream intersection: (1) If ρ 当前 If it is not the maximum, β>α is considered unbalanced, triggering real-time adjustment of the current phase and ending the current phase; (2) If ρ 当前 is the maximum, and the vehicle is released until the green light ends. If β>α, it is determined to be unbalanced, triggering the real-time adjustment of the current phase, and extending the current phase by a green light step length γ; γ should meet the constraint: γ<t max +t more ; When the system runs to the end of the green light again, it can continue to determine whether it is necessary to extend the green light step again until the release time reaches the maximum constraint; Among them, t max The release time for this intersection cannot exceed the maximum release time: t max =t veh Q / n 上 ;t more To increase the release time (when there is a turn release at this intersection, the downstream related turn is also being released, and it is assumed that the upstream and downstream saturation flow is the same): t more =t 下 n 下 / n 上 ; In actual use, the situation where the upstream and downstream related steering saturation flows are different can be considered: t more =t 下 n 下 t veh，上 / (n 上 t veh，下 );t veh is the saturated headway time at the current intersection, t veh，上 is the upstream saturation headway, t veh，下 is the downstream saturation headway, t more is the increase in release time, ρ is the saturation of the intersection, β is the imbalance index of the intersection, and γ is a constant.

2. The single-intersection vehicle-road cooperative traffic control method according to claim 1, characterized in that: The real-time saturation volume S of the single intersection is obtained. The real-time saturation flow S reflects the efficiency of the queued vehicles when they pass the stop line. The calculation formula is: in is the average headway time of vehicles in the queue near the stop line.

3. The single-intersection vehicle-road cooperative traffic control method according to claim 1, characterized in that: The traffic demand of the previous cycle is first obtained, and then balanced distribution is performed; the traffic demand of the previous cycle q 需求 , is the green light flow rate q 通过 , and the queue flow q at the end of the green light 排队 Add up to get the actual traffic demand within the period, specifically: q 需求 =q 排队 +q 通过 .

4. The single-intersection vehicle-road cooperative traffic control method according to claim 1, characterized in that: The closer the distance between the adjacent intersections is, the more the flow of the current intersection will be associated with the storage space of the downstream section. If the flow of the relevant turn of the current intersection exceeds the storage space of the downstream section, a strong association is formed, and it is necessary to constrain the cycle and start time of the current intersection or constrain the release time of the relevant turn of the current intersection, specifically: First obtain the downstream storage space value: Q = n 下 (Ll 排队 ) / l veh ; Then, the value of the strong correlation coefficient is obtained according to the downstream storage space value: δ = Q 转向 / Q; Among them, n 下 is the number of lanes in the downstream section, L is the length of the downstream section, l 排队 is the queue length of the downstream section, l veh is the saturated headway of the current intersection, δ is the strong correlation coefficient, Q is the downstream storage space, Q 转向 Real-time vehicle count for steering; When δ>=ε, it indicates that the downstream intersection is strongly associated with the current intersection, and ε can be adjusted according to the actual situation. When ε=1, it indicates that there is a possibility of overflow in the downstream section.

5. The single-intersection vehicle-road cooperative traffic control method according to claim 1, characterized in that: The step of obtaining the saturation ρ of the single intersection, establishing the imbalance index β, determining whether there is an imbalance, and triggering real-time adjustment of the current phase includes:

51. If the current saturation ρ 当前 Not the maximum, you can set the imbalance index β = (ρ max -ρ 当前 ) / ρ 当前 , and only in ρ 当前 <1,ρ max >1 and β>α, it is determined to be unbalanced, triggering real-time adjustment of the current phase and implementing early disconnection of the current phase; Among them, ρ max is the maximum saturation value in other phases; α is a constant and can be set according to the actual situation. If the upstream section of the turn is particularly long, the storage space is large, and a larger queue length is allowed, α can be set larger accordingly; saturation ρ 当前 The calculation of is based on the ratio of the release time requirement of the queue flow at the current moment to the remaining green light time; 52. If the current saturation ρ 当前 is the maximum, and the vehicle is released until the green light is approaching. The imbalance index β can be set to (ρ 当前 -ρ min ) / ρ min , and only in ρ 当前 >1,ρ min <1, β>α, it is determined to be unbalanced, triggering the current phase to adjust in real time to extend the current phase by a green light step length γ; when it runs again to the point close to the end of the green light, it can continue to determine whether it is necessary to extend the green light step length again; Among them, ρ min is the minimum saturation value in other phases, and γ is a constant that can be set according to actual conditions.

6. The single-intersection vehicle-road cooperative traffic control method according to claim 1, characterized in that: The release time of the intersection cannot exceed the maximum release time t max Otherwise, it is easy to cause excessive vehicles to enter the downstream, causing congestion in the downstream and even spreading to the current intersection; if there is a situation where the downstream related turns are being released when the current intersection is turning, the release time t can be increased. more ;t max ,t more Specifically:

61. The release time of this intersection cannot exceed the maximum release time t max :t max =t veh Q / n 上 ; 62. When there is a turn release at this intersection, the downstream related turn is also being released. Assuming that the saturation flow of the upstream and downstream related turns is the same, the release time t can be increased. more :t more =t 下 n 下 / n 上 ; 63. In actual use, the situation where the upstream and downstream related steering saturation flows are different can be considered: t mor e=t 下 n 下 t veh，上 / (n 上 t veh，下 ); Among them, t veh，上 is the upstream saturation headway, t veh，下 is the downstream saturation headway, t more Add value to the release time.

Images