Dynamic optimization method for intersection short variable lane system

By using space-time integrated optimization strategy in the short variable lane system at the intersection, dynamically adjusting lane functions and signal timing solutions, the problems of low traffic efficiency and insufficient resource utilization in the existing technology are solved, and efficient traffic flow is achieved manage.

CN119028129APending Publication Date: 2024-11-26SHANGHAI KUANGTU TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410991466.6
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-23
Publication Date
2024-11-26

AI Technical Summary

Technical Problem

The existing variable lane system cannot fully utilize road resources during high saturation and peak traffic hours, has low traffic efficiency, and relies on manual regulation and historical experience data to adapt to changes in the traffic state of the entire time and space.

Method used

The dynamic optimization method of the short variable lane system at the intersection is adopted, real-time data is obtained through cameras, radar coils, traffic flow detectors and intelligent algorithms, integrated time and space optimization strategies are implemented, and lane functions and signal timing schemes are dynamically adjusted to improve traffic capacity.

Benefits of technology

It has achieved rapid adaptation to traffic needs, improved the flexibility of variable lane switching and road traffic capacity, made full use of road resources, improved the traffic efficiency of traffic flow, and reduced the limitations of manual regulation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119028129A_ABST
    Figure CN119028129A_ABST
Patent Text Reader

Abstract

The invention discloses a dynamic optimization method for an intersection short variable lane system. The method comprises the steps of obtaining an intersection, and judging whether an intersection entrance lane is an entrance lane of a conditionally set variable lane or not; acquiring system data, and preprocessing the system data; judging whether the variable lane implements a space-time integration strategy or not based on the strategy condition; if the strategy condition is met, implementing a space-time integration strategy, and performing space-time integration optimization based on system data; and analyzing the time-space integrated optimization result based on the theoretical vehicle passing through each period and the queuing length. Based on traffic real-time data such as vehicle flow, queuing length and the like and big data calculation capability, a specific time-space integrated optimization algorithm is combined on a road network topological structure, urban road traffic conditions are identified and analyzed, an optimization scheme is proposed, and the method can achieve the real-time traffic optimization based on data with high update rate such as Internet data and the like. And changing traffic demands can be rapidly adapted.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to a dynamic optimization method for a variable lane system, and in particular to a dynamic optimization method for a short variable lane system at an intersection. Background Art

[0002] In recent years, with the rapid development of cities and the economy, workplaces and residences have gradually concentrated, and traffic flow has shown a significant difference in the size of traffic flow in different directions over time. This phenomenon is called the tidal phenomenon. In the current stage of development, the saturation of traffic flow is getting higher and higher, and the tidal phenomenon of road traffic is becoming more and more obvious. The tidal phenomenon of urban traffic refers to the large traffic flow in the direction of progress during the morning peak and the small traffic flow in the direction of leaving the city; the large traffic flow in the direction of dust removal during the evening peak and the small traffic flow in the direction of progress. At some intersections, the tidal phenomenon is manifested as changes in different turn flows. For example, when an intersection turns left to the direction of progress and turns right to the direction of leaving the city, there may be a large left-turn flow and a small right-turn flow during the morning peak; a large right-turn flow and a small left-turn flow during the evening peak.

[0003] The tidal phenomenon not only occurs in commuter traffic during peak hours in the morning and evening, but is also common in various large-scale collective activities and large-scale events, where a large number of people gather in a specific place or area in a short period of time and then disperse in large numbers. This one-way congestion not only affects the passage of a certain section of road or a certain intersection, but also causes congestion diffusion and spillover on the dense urban road network, and then turns from single-point congestion to regional congestion, seriously affecting the smooth operation of urban traffic. Tidal traffic not only makes road resources not fully utilized, but also causes a waste of road space resources.

[0004] At present, in order to ease traffic congestion and improve road capacity, a variety of measures have been taken, including traffic signal optimization, road infrastructure expansion, traffic demand management and intelligent transportation systems; traffic signal optimization is to balance traffic flow in all directions by adjusting the timing of traffic lights; when expanding road infrastructure, the number of lanes is increased, roads are widened, and the physical capacity of roads is increased; traffic demand management is to reduce vehicle travel during peak hours through measures such as traffic restrictions and license plate restrictions; intelligent transportation systems use advanced information technology, such as vehicle communication systems, traffic monitoring equipment, etc., to monitor and manage traffic flows in real time.

[0005] Among the current solutions, variable lanes are a common and popular method, but the existing variable lane schemes still rely on the experience of traffic experts and need to analyze the rationality of variable lane settings based on the flow obtained from field surveys. They fail to fully consider the evolution of traffic conditions at all time and space scales, affecting urban operation efficiency and residents' quality of life. Especially in highly saturated traffic environments and during peak traffic hours, the efficiency of traffic flow is often low. The existing manual control and time-division pre-allocation schemes are no longer applicable to highly saturated tidal traffic, and have the following shortcomings: the traditional variable lane setting section identification method and optimization scheme are more based on historical experience or short-term survey data, and are not applicable to dynamic traffic under all time and space conditions; the control method of variable lanes is still mainly based on manual switching, timing control and pre-time control, which has low flexibility and efficiency. When the saturation is high and the tidal phenomenon of traffic flow direction is obvious, it is impossible to fully utilize road resources, resulting in manpower and physical waste. Summary of the invention

[0006] In order to solve the deficiencies of the above-mentioned technology, the present invention provides a dynamic optimization method for a short variable lane system at an intersection.

[0007] In order to solve the above technical problems, the technical solution adopted by the present invention is: a dynamic optimization method for a short variable lane system at an intersection, comprising the following steps:

[0008] Step S1: Obtain an intersection, and determine whether an entrance road of the intersection is an entrance road with conditional variable lanes based on constraint conditions;

[0009] Step S2: obtaining system data by setting up intersection cameras, radar coils, traffic flow detectors, data analysis software and intelligent algorithms, and preprocessing the system data;

[0010] Step S3: judging whether the time-space integration strategy is implemented for the variable lane based on the strategy conditions; if the strategy conditions are not met, implementing the time-division control strategy for the variable lane;

[0011] Step S4: If the strategy conditions are met, the time-space integration strategy is implemented, and time-space integration optimization is performed based on system data; the steps of the time-space integration algorithm are:

[0012] Step I: min f(L);

[0013] Step II: min g(f(L),S);

[0014] That is, the time-space integration strategy function is:

[0015] min z(L,S);

[0016] Where L is the lane function, L∈Q; S is the signal timing scheme, S∈Q;

[0017] Step S5: Analyze the time-space integration optimization results based on the theoretical vehicles passing through each cycle and the queue length.

[0018] Furthermore, the constraint condition in step S1 is:

[0019] 1-1: The number of lanes on the entrance road shall not be less than 5;

[0020] 1-2: The overall saturation of the entrance lane is large, and the maximum queue length of each lane is long. It is necessary to set up variable lanes to improve the traffic capacity;

[0021] 1-3: The road grade of the entrance road and the road grade of the intersecting road shall not be lower than that of the secondary road; the secondary road is a general traffic road with a large number in the city;

[0022] 1-4: The entrance lane has no less than 2 function types and meets specific rules; the entrance lane function types include left turn, straight and right turn.

[0023] Furthermore, the specific rule in 1-4 is one of the following conditions:

[0024] 1-41: The entrance road has a left-turn lane and a left-turn waiting area, and the left-turn lane is a left-turn dedicated lane, not a left-turn mixed lane, nor a right-turn left-turn lane;

[0025] 1-42: The entrance road has a right-turn lane, and the right-turn lane is a right-turn lane, not a right-turn bus lane; the right-turn lane includes a dedicated right-turn lane and a right-turn mixed lane;

[0026] 1-43: If the leftmost lane is a mixed lane of U-turn lanes and left-turn lanes, it is considered a dedicated left-turn lane; if the entrance lane contains a U-turn lane at other locations, this specific rule is not met.

[0027] Furthermore, the system data in step S2 includes the number of vehicles, queue length, channelization, lane turn flow, and signal timing plan; the signal timing plan is the traffic light timing plan implemented at the current intersection, including phase sequence and green-to-signal ratio.

[0028] Further, in step S3, the policy condition is 3-1 or 3-2;

[0029] 3-1: If the variable lane is not set to travel in one direction, the traffic flow cannot be dispersed and will cause congestion;

[0030] 3-2: If variable lanes are not set up, vehicles on the entrance lane will queue up and overflow, causing upstream traffic to be unable to pass smoothly.

[0031] Furthermore, a start threshold is set in 3-1. If the vehicle saturation is greater than the start threshold, it is determined that the traffic flow cannot be evacuated and congestion occurs. The formula of the start threshold S is:

[0032] S = Q ÷ C;

[0033] Among them, Q is the single-vehicle road capacity at the intersection of the straight left lane of the main road; C is the single-vehicle road capacity at the default intersection of the main road.

[0034] Furthermore, the queue length L in the entrance lane in 3-2 w , the overflow formula of vehicle queue at the import road is:

[0035]

[0036] Among them, L1 is the fixed length of the intersection entrance road widening; L2 is the variable length of the intersection entrance road widening; L3 is the path length; is the length of the queue overflowing to the road section, is a variable, or p and q are both unknown coefficients.

[0037] Furthermore, in step S4, if the signal timing scheme is unknown, the signal timing scheme, the switching timing of the variable lane and the clearing time of the variable lane are estimated based on the electromagnetic coil data;

[0038] Step S41: Obtain the timestamp of vehicle arrival time_arrive and vehicle departure time_depart returned by the radar coil; obtain the headway time according to the timestamps of the front and rear vehicles arriving obtained by the radar coil

[0039] Step S42: Obtain the headway data of the vehicle. The data at both ends of the headway data are the time when the vehicle stops in the monitoring area and waits for the traffic light to be released; the data in the middle of the headway data are the time when the vehicle discretely passes through the green light time; when the vehicle stops when the green light counts down, stops during the yellow light flashing, and fails to stop at the red light when running through the yellow light, the red light time t in the statistical direction detected by the radar coil red for:

[0040]

[0041] Among them, N big The time that the vehicle stops in the detection area and waits for the traffic light to release; t error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s;

[0042] Green light time t in the statistical direction detected by the radar coilgreen for:

[0043]

[0044] Among them, N small t is the time for vehicles to pass through the green light discretely; error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s;

[0045] Step S43: Based on the red light time and green light time in each direction of the intersection, a signal timing plan is obtained.

[0046] Furthermore, the clearing time t of the variable lane in step S4 is celar The range formula is:

[0047] t min ≤t celar <t max ;

[0048]

[0049] Among them, t max The time required for a vehicle to pass through the variable-width section at an average speed; t min The time required for a vehicle to pass through the widened non-variable section at an average speed;

[0050] Phase time of phase sequence based on clearing time The range formula is:

[0051]

[0052] in, The green light time for a certain phase or lane direction; The shortest green light time for this phase or this lane direction.

[0053] Furthermore, the theoretical number of vehicles passing through each cycle in step S5 is the maximum number of vehicles that can pass through a specific point in a unit time under ideal conditions without considering factors such as acceleration, deceleration, and queue jumping, so as to evaluate the potential traffic capacity of the road; the theoretical number of vehicles passing through each cycle is calculated based on the green light time in the direction and the headway in the saturated flow, and the calculation formula for the theoretical number of vehicles passing through each cycle is:

[0054]

[0055] in, is the average headway time of saturated traffic flow, n is the number of statistical vehicles; is the headway time of the i-th vehicle; t green_j The green light time for the direction of the lane.

[0056] The present invention discloses a dynamic optimization method for a short variable lane system at an intersection. Based on real-time traffic data such as vehicle flow and queue length and big data computing capabilities, a specific time-space integrated optimization algorithm is combined with the road network topology structure to identify and analyze the road traffic conditions in the city, and an optimization plan is proposed. It can be based on data with a high update rate such as Internet data, propose plans suitable for different traffic conditions, and quickly adapt to changing traffic needs; improve the flexibility of variable lane switching and road traffic capacity, and make full use of road resources; through the design of the time-space integrated optimization algorithm process, it breaks through the limitations of manual capabilities, has high efficiency, can quickly identify the road traffic conditions in the entire road network area of ​​a large city, and efficiently output the best variable lane setting plan. BRIEF DESCRIPTION OF THE DRAWINGS

[0057] Figure 1 It is a process diagram of the space-time integrated optimization model in the present invention.

[0058] Figure 2 This is a diagram of the spatiotemporal resource configuration results in the numerical calculation of an embodiment of the present invention.

[0059] Figure 3 This is a comparison chart of the average vehicle delay results in the numerical calculation of the embodiments of the present invention.

[0060] Figure 4 The embodiment of the present invention verifies the configuration of intersection space-time resources in the test scene diagram.

[0061] Figure 5 This is a comparison chart of the average vehicle delay of the fixed timing solution and the distribution optimization solution in the embodiment of the present invention.

[0062] Figure 6 This is a diagram of the morning peak timing strategy in an embodiment of the present invention.

[0063] Figure 7 It is a flow chart of the present invention. DETAILED DESCRIPTION

[0064] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0065] A dynamic optimization method for a short variable lane system at an intersection, comprising the following steps:

[0066] Step S1: Obtain an intersection, and determine whether an entrance road of the intersection is an entrance road with conditional variable lanes based on constraint conditions;

[0067] Step S2: obtaining system data by setting up intersection cameras, radar coils, traffic flow detectors, data analysis software and intelligent algorithms, and preprocessing the system data;

[0068] Step S3: judging whether the time-space integration strategy is implemented for the variable lane based on the strategy conditions; if the strategy conditions are not met, implementing the time-division control strategy for the variable lane;

[0069] Step S4: If the strategy conditions are met, the time-space integration strategy is implemented, and time-space integration optimization is performed based on system data; the steps of the time-space integration algorithm are:

[0070] Step I: min f(L);

[0071] Step II: min g(f(L),S);

[0072] That is, the time-space integration strategy function is:

[0073] min z(L,S);

[0074] Where L is the lane function, L∈Q; S is the signal timing scheme, S∈Q;

[0075] Step S5: Analyze the time-space integration optimization results based on the theoretical vehicles passing through each cycle and the queue length.

[0076] Specifically, the constraints in step S1 are:

[0077] 1-1: The number of lanes on the entrance road shall not be less than 5;

[0078] 1-2: The overall saturation of the entrance lane is large, and the maximum queue length of each lane is long. It is necessary to set up variable lanes to improve the traffic capacity;

[0079] 1-3: The road grade of the entrance road and the road grade of the intersecting road shall not be lower than that of the secondary road; the secondary road is a general traffic road with a large number in the city;

[0080] 1-4: The entrance lane has no less than 2 function types and meets specific rules; the entrance lane function types include left turn, straight and right turn.

[0081] The specific rule in 1-4 is one of the following conditions:

[0082] 1-41: The entrance road has a left-turn lane and a left-turn waiting area, and the left-turn lane is a left-turn dedicated lane, not a left-turn mixed lane, nor a right-turn left-turn lane;

[0083] 1-42: The entrance road has a right-turn lane, and the right-turn lane is a right-turn lane, not a right-turn bus lane; the right-turn lane includes a dedicated right-turn lane and a right-turn mixed lane;

[0084] 1-43: If the leftmost lane is a mixed lane of U-turn lanes and left-turn lanes, it is considered a dedicated left-turn lane; if the entrance lane contains a U-turn lane at other locations, this specific rule is not met.

[0085] The system data in step S2 includes the number of vehicles, queue length, channelization, lane turn flow, and signal timing plan; the signal timing plan is the traffic light timing plan implemented at the current intersection, including phase sequence and green-to-signal ratio.

[0086] In step S3, the policy condition is 3-1 or 3-2;

[0087] 3-1: If the variable lane is not set to travel in one direction, the traffic flow cannot be dispersed and will cause congestion;

[0088] 3-2: If variable lanes are not set up, vehicles on the entrance lane will queue up and overflow, causing upstream traffic to be unable to pass smoothly.

[0089] The start threshold is set in 3-1. If the vehicle saturation is greater than the start threshold, it is determined that the traffic flow cannot be evacuated and congestion occurs. The formula for the start threshold S is:

[0090] S = Q ÷ C;

[0091] Among them, Q is the single-vehicle road capacity at the intersection of the straight left lane of the main road; C is the single-vehicle road capacity at the default intersection of the main road.

[0092] 3-2 Length of queue at the middle entrance L w , the overflow formula of vehicle queue at the import road is:

[0093]

[0094] Among them, L1 is the fixed length of the intersection entrance road widening; L2 is the variable length of the intersection entrance road widening; L3 is the path length; is the length of the queue overflowing to the road section, is a variable, or p and q are both unknown coefficients.

[0095] In step S4, if the signal timing scheme is unknown, the signal timing scheme, the switching timing of the variable lane and the clearing time of the variable lane are estimated based on the electromagnetic coil data;

[0096] Step S41: Obtain the timestamp of vehicle arrival time_arrive and vehicle departure time_depart returned by the radar coil; obtain the headway time according to the timestamps of the front and rear vehicles arriving obtained by the radar coil

[0097] Step S42: Obtain the headway data of the vehicle. The data at both ends of the headway data are the time when the vehicle stops in the monitoring area and waits for the traffic light to be released; the data in the middle of the headway data are the time when the vehicle discretely passes through the green light time; when the vehicle stops when the green light counts down, stops during the yellow light flashing, and fails to stop at the red light when running through the yellow light, the red light time t in the statistical direction detected by the radar coil red for:

[0098]

[0099] Among them, N big The time that the vehicle stops in the detection area and waits for the traffic light to release; t error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s;

[0100] Green light time t in the statistical direction detected by the radar coil green for:

[0101]

[0102] Among them, N small t is the time for vehicles to pass through the green light discretely; error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s;

[0103] Step S43: Based on the red light time and green light time in each direction of the intersection, a signal timing plan is obtained.

[0104] The clearing time t of the variable lane in step S4 celar The range formula is:

[0105] t min ≤t celar <t max ;

[0106]

[0107] Among them, t max The time required for a vehicle to pass through the variable-width section at an average speed; t min The time required for a vehicle to pass through the widened non-variable section at an average speed;

[0108] Phase time of phase sequence based on clearing time The range formula is:

[0109]

[0110] in, The green light time for a certain phase or lane direction; The shortest green light time for this phase or this lane direction.

[0111] The theoretical number of vehicles passing through each cycle in step S5 is the maximum number of vehicles that can pass through a specific point in a unit time under ideal conditions without considering factors such as acceleration, deceleration, and queue jumping, so as to evaluate the potential traffic capacity of the road; the theoretical number of vehicles passing through each cycle is calculated based on the green light time in the direction and the headway in the saturated flow. The calculation formula for the theoretical number of vehicles passing through each cycle is:

[0112]

[0113] in, is the average headway time of saturated traffic flow, n is the number of statistical vehicles; is the headway time of the i-th vehicle; t green_j The green light time for the direction of the lane.

[0114] This embodiment is a method for dynamically optimizing a short variable lane system at an intersection. It provides a variable lane dynamic optimization solution using an intersection in Hangzhou as a pilot. Figure 7 The flowchart of a dynamic optimization method for a short variable lane system at an intersection specifically includes the following steps:

[0115] Step A1: Obtain the intersection and determine whether the variable lane is an entrance lane with conditional variable lane setting based on the constraint conditions; the entrance lane is the lane in the road section that enters the intersection according to the vehicle driving direction. The constraint conditions are:

[0116] Step A11: The number of lanes of the entrance road is not less than 5;

[0117] Step A12: The overall saturation of the entrance lane is large, and the maximum queue length of each lane is long. It is necessary to set up variable lanes to improve the traffic capacity;

[0118] Step A13: The road grade of the entrance road and the road grade of the intersecting road shall not be lower than the grade of the secondary trunk road; the secondary trunk road is a general traffic road with a large number in the city;

[0119] Step A14: The lane function types of the entrance road are not less than 2 and meet specific rules; the lane function types of the entrance road include left turn, straight and right turn; the specific rules are:

[0120] Step A141: The entrance road has a left-turn lane and a left-turn waiting area, and the left-turn lane is a left-turn dedicated lane, not a left-turn mixed lane, nor a right-turn left-turn lane;

[0121] Step A142: The entrance road has a right-turn lane, and the right-turn lane is a right-turn lane, not a right-turn bus lane; the right-turn lane includes a dedicated right-turn lane and a right-turn mixed lane;

[0122] Step A143: If the leftmost lane is a mixed lane of a U-turn lane and a left-turn lane, it is regarded as a dedicated left-turn lane; if the entrance lane contains a U-turn lane at other locations, this specific rule is not met.

[0123] Step A2: Obtain system data through intersection cameras, radar coils, traffic flow detectors, data analysis software and intelligent algorithms, and pre-process the system data. The system data includes vehicle passing data, queue length, channelization, lane turn flow, and signal timing plan;

[0124] The vehicle passing data includes the arrival time of each vehicle, the departure time of each vehicle, the lane number of the vehicle, and the turn of the vehicle.

[0125] The queue length is the length of the queue of vehicles waiting to pass through the intersection, mainly the maximum queue length of each lane in each cycle.

[0126] Channelization refers to the driving path and distribution of vehicles at the intersection, including the length of the variable widening section, the length of the non-shortened widening section, and the length of the upstream section.

[0127] The signal timing plan is the traffic light timing plan currently implemented at the intersection, including phase sequence, greening ratio, phase time and other information.

[0128] Intersection cameras are high-definition cameras installed above or near intersections to capture vehicle images and identify vehicle types. Radar coils are buried under the road surface to detect real-time data when vehicles pass through, including vehicle speed and vehicle quantity, i.e., vehicle passing data. Traffic flow detectors are sensors deployed at the entrance and exit of intersections to measure vehicle turning flow and queue length. Data analysis software Instant messaging Professional data analysis software processes the collected system data to extract useful traffic information. Intelligent algorithms are used to intelligently process the developed algorithms to obtain the required system data.

[0129] The default values ​​of road grade and capacity are shown in Table 1;

[0130] Table 1

[0131]

[0132] Step A3: Determine whether to implement the spatiotemporal integration strategy based on the strategy conditions; the strategy conditions are to satisfy step S31 or step S32;

[0133] Step A31: If the variable lane is not set to go in the same direction, the traffic flow cannot be evacuated and congestion will occur; the saturation S> start threshold 0.6 is satisfied; in the Hangzhou pilot, the road section is 3 lanes and the widening section is 5 lanes. The lane functions of each lane are: lane 1 is a left turn lane, lane 2 is a straight lane with left turn, lane 3 is a straight lane, lane 4 is a straight lane, and lane 5 is a straight lane. Since the radar data has a large fluctuation in the vehicle detection data beyond the range of 150m, it has caused a large error. The start threshold is calculated based on the vehicle flow data. The calculation formula of the start threshold S is:

[0134] S = Q ÷ C;

[0135] Among them, Q is the single-vehicle road capacity at the intersection of the straight left lane of the main road; C is the single-vehicle road capacity at the default intersection of the main road.

[0136] After theoretical calculations and actual implementation results, the starting threshold of the Hangzhou pilot is 0.6. If the saturation of the Hangzhou pilot is greater than the starting threshold, the time-space integration strategy will be activated, otherwise it will cause the intersection to be locked, and then the congestion will overflow.

[0137] Step A32: If variable lanes are not set, vehicles on the entrance lane will queue up and overflow, causing the upstream traffic flow to be unable to pass smoothly; the queue length of the entrance lane is L w , the import vehicle queue overflow formula is:

[0138]

[0139] Among them, L1 is the fixed length of the intersection entrance road widening; L2 is the variable length of the intersection entrance road widening; L3 is the path length; is the length of the queue overflowing to the road section, is a variable, or p and q are both unknown coefficients.

[0140] Step A4: Perform time-space integrated optimization; determine the lane function L, use the integrated optimization function to perform time-space integrated optimization on the lane function L and the signal timing S, and output the optimal solution of the lane function and signal timing at one time; Figure 1 The process diagram of the time-space integrated optimization model in the dynamic optimization method of the short variable lane system at an intersection is shown in the figure. The calculation steps of the time-space integrated optimization algorithm are:

[0141] Step I: min f(L);

[0142] Step II: min g(f(L),S);

[0143] It is concluded that the space-time integrated optimization function is:

[0144] min z(L,S);

[0145] Where L is the lane function, L∈Q; S is the signal timing scheme, S∈Q;

[0146] Step A41: Determine whether the signal timing scheme is known; if the signal timing scheme is known, perform time-space integrated optimization based on the existing system data;

[0147] Step A42: If the signal timing scheme is unknown, estimate the signal timing based on the electromagnetic coil data;

[0148] The vehicle arrival time returned by the radar coil is time_arrive; the departure timestamp is time_depart; the headway time is obtained based on the timestamps of the front and rear vehicles obtained by the radar coil It can be used in the case of unknown intersection cycle to ensure high accuracy;

[0149] Once the green light on time, phase sequence, clearing time and phase time constraints are determined, the plan will be issued. The time-space integration strategy can be switched by directly controlling the lane function instructions through the Internet, or it can be directed by simple operations by on-site personnel;

[0150] The time distance data of the two groups of vehicles are:

[0151] Group A: 80-3-2-1-1-2-2-2-2-3-3-4-1-1-2-5-2-2-2-2-85;

[0152] Group B: 90-2-3-1-2-3-1-2-2-2-2-2-3-3-3-3-4-5-5-6-84;

[0153] The time interval data at both ends of Group A and Group B is the time when the vehicle stops in the detection area and waits for the traffic light to be released; the remaining time interval data is the time when the vehicle passes through the green light. According to the vehicle stopping when the green light counts down, stopping during the yellow light flashing, and stopping at the initial red light when the yellow light is unsuccessful; the red light time t in the statistical direction detected by the radar coil red for:

[0154]

[0155] Among them, N big The time that the vehicle stops in the detection area and waits for the traffic light to release; t error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s;

[0156] Green light time t in the statistical direction detected by the radar coil green for:

[0157]

[0158] Among them, N small t is the time for vehicles to pass through the green light discretely; error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s;

[0159] Based on the red light time and green light time in each direction of the intersection, the signal timing plan is calculated; for the time-space integration optimization strategy, it is necessary to determine the switching timing of the variable lane and the clearing time of the variable lane according to the passing vehicles; since the signboard of the variable lane is usually set on the gantry, and the gantry is located between the widened variable lanes, based on the vehicles that have entered the variable lane at the time of switching and the vehicles that enter the variable lane due to negligence, the switching of the variable lane should actually meet the requirements that the vehicle can safely leave the variable vehicle without blocking the following vehicles. The formula for satisfying the clearing time is:

[0160] t min ≤t celar <t max ;

[0161]

[0162] Among them, t max The time required for a vehicle to pass through the variable-width section at an average speed; t min The time required for a vehicle to pass through the widened non-variable section at an average speed;

[0163] The phase time formula is:

[0164]

[0165] in, The green light time for a certain phase or lane direction; The shortest green light time for this phase or this lane direction is calculated based on saturation, green-to-signal ratio, vehicle passing time and control strategy.

[0166] A variable lane in a certain direction at an intersection has two constraints and clearing time. In the pilot "straight-left variable lane" in Hangzhou, the clearing time is calculated. The clearing time for switching from straight to left turn is 20s, and the clearing time for switching from left turn is 15s; the length of the clearing time represents the radicalism and robustness of the control strategy. The smaller the clearing time, the more radical the control strategy and the higher the requirements for the driver.

[0167] Step A5: Perform time-space integrated optimization result analysis;

[0168] According to the influence of driver acceptance, driving distraction, traffic flow propagation characteristics and other realistic factors, the length of the queue at the entrance can be effectively reduced and balanced; the indicator of the number of passing vehicles needs to be calculated according to the ideal number of passages of saturated traffic flow, that is, theoretical calculation, ignoring the influence of factors such as traffic gaps and driver driving; the queue length is based on the actual data actually detected by the radar coil.

[0169] Step A51: The theoretical calculation is the maximum number of vehicles that can pass a specific point in a unit time under ideal conditions without considering realistic factors such as acceleration, deceleration and queue jumping, so as to evaluate the potential traffic capacity of the road; the theoretical number of vehicles passing each cycle is calculated based on the green light time in that direction and the headway under saturated flow;

[0170] In the Hangzhou pilot, Group A: 80-3-2-1-1-2-2-2-2-3-3-4-1-1-2-5-2-2-2-2-85; the front and rear vehicles with large headway due to starting loss are discarded, and then 5 to 10 vehicles are selected as the headway of saturated traffic, for example, "2-1-1-2-2-2", then the theoretical number of vehicles passing each cycle is n j for:

[0171]

[0172] in, is the average headway time of saturated traffic flow, n is the number of statistical vehicles; is the headway time of the i-th vehicle; t green_j It refers to the green light time of a certain phase or lane direction.

[0173] Step A52: Calculate the queue length; the queue length is based on actual data, and the maximum queue length of each entrance is selected to calculate the optimization efficiency.

[0174] A comparative analysis was conducted using three schemes:

[0175] The first scheme is a fixed timing scheme, which uses a two-stage method to optimize the signal when the lane function is fixed;

[0176] The second solution is the distribution optimization solution, which is to optimize lane function first and then optimize signal timing;

[0177] The third solution is the collaborative optimization solution, that is, the time-space integrated optimization solution.

[0178] Programming was done using Python, and three vehicle flows were randomly selected. Three schemes were used for cross-combination, forming a total of nine experiments. Figure 2 The spatiotemporal resource allocation result diagram of the numerical calculation of an embodiment of a dynamic optimization method for a short variable lane system at an intersection is shown in FIG. Figure 2 The horizontal axis in the middle is the proportion of vehicles going straight, and the vertical axis is the proportion of vehicles turning left. The black area is divided into the optimal flow interval for different lane functions. The color changes from light to dark, indicating that the straight green signal ratio changes from small to large. Figure 2 The coordinates of M1 to M9 are all (20,40). Figure 2 -(a), 2-(b), and 2-(c) represent the fixed timing scheme, the distributed optimization scheme, and the collaborative optimization scheme for high traffic, respectively; Figure 2 -(d), 2-(e), and 2-(f) represent the fixed timing scheme, the distribution optimization scheme, and the collaborative optimization scheme for medium traffic, respectively; Figure 2 -(g), 2-(h), and 2-(i) represent the fixed timing scheme, distributed optimization scheme, and collaborative optimization scheme for low traffic, respectively; in the spatial resource allocation, the different lane function schemes of the distributed optimization scheme and the collaborative optimization scheme are discretely distributed in different traffic intervals. As the straight-through ratio or the left-turn ratio increases, the number of straight-through lanes or the number of left-turn lanes gradually increases, that is, when the supply and demand relationship at the intersection changes greatly, it is necessary to switch the lane function for adjustment. In the time allocation of resources, within the unified lane function interval of the distributed optimization scheme and the collaborative optimization scheme, as the straight-through ratio increases, the straight-through green-signal ratio gradually increases, that is, when the supply and demand relationship at the intersection does not change much, only the green-signal ratio needs to be adjusted. For the fixed timing scheme, since the lane function remains unchanged, the straight-through green-signal ratio is adjusted as the traffic turning ratio changes.

[0179] In order to further compare the traffic efficiency improvement of DLG, the average vehicle delay is used as the evaluation index; Figure 3 The comparison results of the average vehicle delay in the numerical calculation of an embodiment of a dynamic optimization method for a short variable lane system at an intersection are shown in the figure. In the figure, the horizontal axis is the straight driving ratio, and the vertical axis is the left turning ratio; the color of the squares from light to dark indicates that the average vehicle delay reduction percentage of the collaborative optimization scheme, that is, the time-space integration optimization scheme, is from low to high. Figure 3 -(a) represents the percentage of average vehicle delay reduction of the collaborative optimization scheme compared with the fixed timing scheme under high traffic conditions. Figure 3 -(b) shows the percentage of average vehicle delay reduction of the collaborative optimization scheme compared with the distributed optimization scheme under high traffic conditions. Figure 3 -(c) indicates the percentage of average delay reduction of the coordinated optimization scheme under medium traffic compared with the fixed timing scheme. Figure 3 -(d) indicates the percentage of average vehicle delay reduction of the collaborative optimization scheme compared with the distributed optimization scheme under medium traffic. Figure 3-(e) indicates the percentage of average vehicle delay reduction of the collaborative optimization scheme compared with the fixed timing scheme under low traffic conditions. Figure 3 -(f) represents the percentage of average vehicle delay reduction of the collaborative optimization scheme compared with the distributed optimization scheme under low traffic conditions; Figure 3 -(a), Figure 3 -(c) Figure 3 -(e) It is found that the time-space integrated optimization scheme is generally helpful in reducing intersection delays compared with the fixed timing scheme. Figure 3 -(b), Figure 3 -(d), Figure 3 -(f) shows that the time-space integration optimization scheme reduces the delay range compared with the distribution optimization scheme mainly when the straight-through ratio is less than 30%; when the straight-through ratio is greater than 30%, the optimization results of the time-space integration optimization scheme and the distribution optimization scheme are consistent. Figure 3 -(b), 3-(d), and 3-(f) when the straight driving ratio is greater than 30%, there are a few points where the average vehicle error increases. The reason is that the distribution optimization scheme cannot be consistent with the objective function of the space-time integration optimization scheme in the first step of optimization, resulting in a better solution. However, on the whole, the average vehicle delay of the space-time integration optimization scheme is less than or equal to that of the distribution optimization scheme, which proves that the space-time integration optimization can improve the traffic efficiency of the intersection.

[0180] like Figure 4 An example of a dynamic optimization method for a short variable lane system at an intersection demonstrates the spatial and temporal resource configuration of the intersection as shown in the test scenario diagram. Figure 4 As shown in the intersection signal timing scheme diagram of -b, phase 1 and phase 3 are released symmetrically, phase 2 and phase 4 are released symmetrically, and the total cycle of the four phases is 130s, the cycle of phase 1 is 25s, the cycle of phase 2 is 35s, the cycle of cycle 3 is 30s, and the cycle of phase 4 is 40s; Figure 4 -a, the empirical location is the intersection of Binsheng Road and Airport City Avenue in Hangzhou. The experimental data comes from the traffic flow data collected by the video detector from 09:00 to 10:00 on May 14, 2020. The intersection is a four-phase signal-controlled intersection. According to the current situation of the intersection, two schemes are proposed for the north entrance of Airport Avenue for empirical analysis:

[0181] Solution I: Fixed limited spatial and temporal resource allocation and conducted comparative analysis;

[0182] Scheme II: Optimize the spatial and temporal resource allocation according to the real-time traffic flow and realize the dynamic distribution optimization method; the north import traffic data, lane function configuration and signal timing scheme during the test are shown in Table 2.

[0183] Table 2

[0184]

[0185] like Figure 5 As shown in the comparison diagram of the average vehicle delay of the fixed timing scheme and the distributed optimization scheme in an embodiment of a dynamic optimization method for a short variable lane system at an intersection, the average vehicle delay for left turns using the fixed timing scheme is 65s, and the average vehicle delay for left turns using the distributed optimization scheme is 52s; the average vehicle delay for straight going using the fixed timing scheme is 53s, and the average vehicle delay for straight going using the distributed optimization scheme is 58s; the average vehicle delay for left turn-straight going using the fixed timing scheme is 59s, and the average vehicle delay for left turn-straight going using the distributed optimization scheme is 55s; after adopting the distributed optimization scheme, although the average vehicle delay of the straight entrance lane is increased by 9.4%, the average vehicle delay of the left turn entrance lane is reduced by 20%, and the average vehicle delay of left turns and straight going at the north entrance is reduced by 6.7%, which can effectively reduce the average vehicle delay at the intersection and improve the operation efficiency of the intersection.

[0186] like Figure 6 In the embodiment of a dynamic optimization method for a short variable lane system at an intersection, the morning peak timing strategy diagram is shown, where the ten-point intersection is a T-shaped intersection. Figure 6 -(a) indicates that the east-west straight travel period of A is 40s-45s, Figure 6 -(b) indicates that the cycle of B going straight and turning left in the west is 30s-35s, Figure 6 -(c) indicates that the period of C turning left to the west is 15s-20s. Figure 6 -(d) indicates that the left turn cycle of D north is 45s-50s, and the total cycle length is 150s; as shown in Table 3, the real-time function suggestion table of the Hangzhou variable lane pilot is shown, the issued lane function 1 indicates left turn, and the issued lane function 2 indicates straight ahead, as shown in Table 3:

[0187] Table 3

[0188]

[0189]

[0190] Implementation process: Automatic control through integrated signal control machine. First, the system is structured. The time-space integrated optimization algorithm will be closely integrated with the existing traffic signal control system and input into the intelligent traffic management system. Use sensor technology, such as video surveillance, geomagnetic sensors, etc., to collect real-time traffic data of each lane to realize data collection. Based on the collected implementation traffic data, theoretical calculations and evaluations are performed. If the boundary conditions for the switching of the variable lane function are met, the phase and duration of the signal light are automatically adjusted to optimize the lane function. During peak hours or special circumstances, the time-space integrated optimization algorithm will automatically trigger the conversion of the variable lane, such as converting the straight lane to the left turn lane to meet traffic needs; set up a feedback mechanism, the system will monitor the control effect in real time, and adjust the algorithm parameters according to the actual traffic to achieve optimal traffic flow management.

[0191] The command center will choose to conduct online control based on the real-time suggestions of the system. Develop an intuitive user interface that displays real-time traffic data and the suggestions of the spatiotemporal optimization algorithm for use by the staff of the traffic police command center. The spatiotemporal optimization algorithm will provide decision support based on data analysis, such as the function of suggesting changes to certain lanes within a cycle or at a specific time period. Establish a fast-response communication mechanism to ensure that the traffic police command center can receive the suggestions of the algorithm in a timely manner and make decisions quickly. On the basis of automatic control, retain the option of manual intervention so that in special circumstances, the traffic police can make adjustments based on experience and real-time suggestions.

[0192] Traffic police obtain system information to facilitate on-site command. Develop a mobile application that enables traffic police to receive system information and algorithm suggestions in real time on the spot. Traffic police make quick judgments and conduct on-site command based on the information received, such as temporarily changing lane functions or guiding vehicles to detour. Train and guide traffic police to ensure that they can understand the algorithm's suggestions and effectively use system information for traffic management. Specify emergency response measures to ensure that traffic police can take quick action in situations that the algorithm cannot cover or predict.

[0193] At present, variable lanes with high lane function switching frequency are rarely used in practice. In the initial trial and promotion stage, drivers may fail to notice the change of lane function and hesitate when entering the variable lane due to factors such as unfamiliarity and negligence. Therefore, by introducing the clearing time, the clearing time can be adjusted according to the actual situation and driver acceptance in the subsequent application and promotion process; public publicity and education activities can be used to improve drivers' awareness and understanding of the variable lane function. In the actual application of the variable lane pilot in Hangzhou, it was found that no matter how the clearing time of the variable lane was set, there was a small probability that vehicles would not have time to drive out of the variable lane and block the vehicles passing in the next phase. It was often the case that left-turning vehicles blocked straight-moving vehicles, while straight-moving vehicles rarely blocked left-turning vehicles. Considering the actual driving conditions of left-turning vehicles, three solutions are set up; Method 1 is to draw a left-turn waiting area in the adjacent lane according to the situation of the lane, but this method will visually confuse the drivers of straight-moving vehicles, causing the drivers to hesitate, which is a potential safety hazard, and this hazard may reduce the efficiency of the entrance lane; Method 2 is to draw one or more waiting spaces for vehicles that have not left in time before the stop line of the waiting area of ​​the lane. This method may cause some vehicles to maliciously enter the variable lane or waiting space in order to rush through the intersection, resulting in queue cutting and concurrent violations. Electric alarms can be added at the position of the gantry of the variable lane. With the promotion of variable lanes, corresponding regulations can be added to ensure the safety and order of intersections; Method 3 is to add signs in the visible area on the opposite side of the road or the gantry on the upstream section of the variable lane to remind left-turning vehicles to give way in time, exit the waiting space, and pass the straight-moving vehicles. The effect of the algorithm is evaluated regularly, and optimized and upgraded based on the effect feedback.

[0194] By using real-time traffic data such as flow and queue length and big data computing capabilities, and combining specific professional algorithms with the universal road network topology, the city's road traffic conditions can be identified and analyzed; based on Internet data, such as flow and queue length, which have high update rates, it is possible to propose solutions suitable for different traffic conditions and quickly adapt to changing traffic needs; it is possible to improve the flexibility of variable lane switching and road capacity, and make full use of road resources; through the design of algorithm processes, it is possible to get rid of the limitations of manual capabilities, be highly efficient, and be able to quickly identify road traffic conditions in the entire road network area of ​​large cities, and efficiently output the best variable lane setting solution. Develop technologies and methods that can effectively improve the capacity in saturated tidal traffic environments, use the increasingly rich accumulation of urban traffic Internet data, and on the basis of traffic engineering and traffic design theory, identify the sections where variable lanes are set more accurately and quickly, and give the best solution based on all-time and space Internet data, so as to achieve the setting of most variable lanes in the city.

[0195] The time-space integration optimization solution, namely the time-space integration strategy, dynamically adjusts the switching strategy of variable lanes based on real-time traffic information such as vehicle flow, queue length and signal timing plan. It can be more accurate and flexible, and can dynamically capture traffic demand, reasonably allocate traffic resources, maximize traffic efficiency and improve traffic capacity.

[0196] The above implementation modes are not limitations of the present invention, and the present invention is not limited to the above examples. Changes, modifications, additions or substitutions made by technicians in this technical field within the scope of the technical solution of the present invention also belong to the protection scope of the present invention.

Claims

1. A dynamic optimization method for a short variable lane system at an intersection, characterized in that: The following steps are involved: Step S1: Obtain an intersection, and determine whether an entrance road of the intersection is an entrance road with conditional variable lanes based on constraint conditions; Step S2: obtaining system data by setting up intersection cameras, radar coils, traffic flow detectors, data analysis software and intelligent algorithms, and preprocessing the system data; Step S3: judging whether the variable lane implements the time-space integration strategy based on the strategy conditions; If the strategy conditions are not met, the time-division control strategy for the variable lanes is implemented; Step S4: If the strategy conditions are met, the time-space integration strategy is implemented, and time-space integration optimization is performed based on system data; The steps of the space-time integration algorithm are: Step I: min f(L); Step II: min g(f(L),S); That is, the time-space integration strategy function is: min z(L,S); Where L is the lane function, L∈Q; S is the signal timing scheme, S∈Q; Step S5: Analyze the time-space integration optimization results based on the theoretical vehicles passing through each cycle and the queue length.

2. The dynamic optimization method for a short variable lane system at an intersection according to claim 1, characterized in that: The constraint conditions in step S1 are: 1-1: The number of lanes on the entrance road shall not be less than 5; 1-2: The overall saturation of the entrance lane is large, and the maximum queue length of each lane is long. It is necessary to set up variable lanes to improve the traffic capacity; 1-3: The road grade of the entrance road and the road grade of the intersecting road shall not be lower than that of the secondary road; the secondary road is a general traffic road with a large number in the city; 1-4: The entrance lane has no less than 2 function types and meets specific rules; the entrance lane function types include left turn, straight and right turn.

3. The dynamic optimization method for a short variable lane system at an intersection according to claim 2, characterized in that: The specific rule in 1-4 is one of the following conditions: 1-41: The entrance road has a left-turn lane and a left-turn waiting area, and the left-turn lane is a left-turn dedicated lane, not a left-turn mixed lane, nor a right-turn left-turn lane; 1-42: The entrance road has a right-turn lane, and the right-turn lane is a right-turn lane, not a right-turn bus lane; The right-turn lanes on the right include dedicated right-turn lanes and mixed right-turn lanes; 1-43: If the leftmost lane is a mixed lane of U-turn lanes and left-turn lanes, it is considered a dedicated left-turn lane; if the entrance lane contains a U-turn lane at other locations, this specific rule is not met.

4. The dynamic optimization method for a short variable lane system at an intersection according to claim 1, characterized in that: The system data in step S2 includes the number of vehicles, queue length, channelization, lane turn flow, and signal timing plan; the signal timing plan is the traffic light timing plan implemented at the current intersection, including phase sequence and green-to-signal ratio.

5. The dynamic optimization method for a short variable lane system at an intersection according to claim 1, characterized in that: The strategy condition in step S3 is 3-1 or 3-2; 3-1: If the variable lane is not set to travel in one direction, the traffic flow cannot be dispersed and will cause congestion; 3-2: If variable lanes are not set up, vehicles on the entrance lane will queue up and overflow, causing upstream traffic to be unable to pass smoothly.

6. The method for dynamic optimization of short variable lane system at intersection according to claim 5, characterized in that: In 3-1, a start threshold is set. If the vehicle saturation is greater than the start threshold, it is determined that the traffic flow cannot be evacuated and congestion occurs. The formula of the start threshold S is: S = Q ÷ C; Among them, Q is the single-vehicle road capacity at the intersection of the straight left lane of the main road; C is the single-vehicle road capacity at the default intersection of the main road.

7. The method for dynamic optimization of short variable lane system at intersection according to claim 6, characterized in that: The queue length L of the entrance in 3-2 w , the overflow formula of vehicle queue at the import road is: Among them, L1 is the fixed length of the intersection entrance road widening; L2 is the variable length of the intersection entrance road widening; L3 is the path length; is the length of the queue overflowing to the road section, is a variable, or p and q are both unknown coefficients.

8. The method for dynamic optimization of short variable lane system at intersection according to claim 1, characterized in that: In step S4, if the signal timing scheme is unknown, the signal timing scheme, the switching timing of the variable lane and the clearing time of the variable lane are estimated based on the electromagnetic coil data; Step S41: Obtain the timestamp of vehicle arrival time_arrive and vehicle departure time_depart returned by the radar coil; obtain the headway time according to the timestamps of the front and rear vehicles arriving obtained by the radar coil Step S42: obtaining the headway data of the vehicle, the data at both ends of the headway data are the time when the vehicle stops in the monitoring area and waits for the traffic light to be released; the data in the middle of the headway data are the time when the vehicle passes discretely during the green light time; When the vehicle stops during the green light countdown, stops during the yellow light flashing, and fails to run the yellow light and stops at the red light, the red light time t in the statistical direction detected by the radar coil red for: Among them, N big The time that the vehicle stops in the detection area and waits for the traffic light to release; t error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s; Green light time t in the statistical direction detected by the radar coil green for: Among them, N small t is the time for vehicles to pass through the green light discretely; error is the fluctuation range, i.e. the yellow light time, t error ≈t yellow =3~5s; Step S43: Based on the red light time and green light time in each direction of the intersection, a signal timing plan is obtained.

9. The method for dynamic optimization of short variable lane system at intersection according to claim 8, characterized in that: The clearing time t of the variable lane in step S4 celar The range formula is: t min ≤t celar <t max ; Among them, t max The time required for a vehicle to pass through the variable-width section at an average speed; t min The time required for a vehicle to pass through the widened non-variable section at an average speed; Phase time of phase sequence based on clearing time The range formula is: in, The green light time for a certain phase or lane direction; The shortest green light time for this phase or this lane direction.

10. The method for dynamic optimization of short variable lane system at intersection according to claim 1, characterized in that: The theoretical number of vehicles passing through each cycle in step S5 is the maximum number of vehicles that can pass through a specific point in a unit time under ideal conditions without considering factors such as acceleration, deceleration, and queue jumping, so as to evaluate the potential traffic capacity of the road; the theoretical number of vehicles passing through each cycle is calculated based on the green light time in the direction and the headway in the saturated flow, and the calculation formula for the theoretical number of vehicles passing through each cycle is: in, is the average headway time of saturated traffic flow, n is the number of statistical vehicles; is the headway time of the i-th vehicle; t green_j The green light time for the direction of the lane.