Arrangement method of trunk line intersection CAV special lane in mixed driving environment and self-adaptive control system

Abstract

The invention discloses a main line intersection CAV special lane setting method and a self-adaptive control system in a mixed driving environment. The method comprises the following steps: collecting information in real time; according to the current CAV permeability, the traffic capacity in the trunk line direction in three different states is calculated, the maximum net income of the traffic capacity is calculated, and then lane function switching judgment conditions are set in a mode of comparing the maximum net income of the traffic capacity; the lane function switching judgment condition is used for judging whether to open or close a CAV special lane, and determining that the CAV special lane is set as a CAV full-steering special lane or a CAV straight special lane; and iteratively optimizing the generated signal control scheme by adopting a PS-PPO-based multi-agent deep reinforcement learning algorithm, and generating an optimal signal control scheme in real time. According to the invention, the hybrid lane and CAV lane functions are dynamically adjusted according to the change of the CAV permeability, the cooperative optimization of the lane functions and the signal control scheme is realized, and the traffic efficiency of the trunk line intersection is improved to the greatest extent.

Description

Technical Field

[0001] This invention belongs to the field of intelligent traffic management and control technology, specifically referring to a method for setting up CAV dedicated lanes at arterial intersections in mixed traffic environments and an adaptive control system. Background Technology

[0002] The development of intelligent connected and autonomous driving technologies has given rise to intelligent connected mixed traffic flows composed of human-driven vehicles (HVs) and connected and automated vehicles (CAVs). In the mixed traffic environment of CAVs and HVs, due to the uncertainty and uncontrollability of HV operation, the different operating characteristics of CAVs compared to HVs, and the constantly changing CAV penetration rate in the mixed traffic flow, the operating patterns of the mixed traffic flow differ from those of traditional traffic flows composed of human-driven vehicles, and many uncertainties exist.

[0003] From a traffic management perspective, the significant characteristics of urban road traffic congestion in my country are its short duration and localized nature. Furthermore, congestion on local road sections or intersections can trigger congestion on related road sections and intersections. Therefore, it is crucial to comprehensively analyze the overall operational status of urban road traffic flow at the trunk line level and implement corresponding control measures. Intelligent connected vehicle (ICV) systems provide more and more accurate real-time traffic data for the coordinated control of urban main traffic arteries. With the gradual promotion and popularization of CAVs (Continuously Operated Vehicles), the establishment of new traffic infrastructure such as dedicated CAV lanes is changing the traditional traffic operating environment, and the corresponding road traffic flow characteristics will also change accordingly.

[0004] To fully leverage the advantages of CAV technology while addressing the complexities of mixed traffic flow, deploying dedicated CAV lanes (lanes for connected autonomous vehicles) on roads where HVs and CAVs share the road has become an important technological trend and management choice.

[0005] However, existing technologies still have some limitations in the setup of dedicated CAV lanes.

[0006] For example, the first method is to set up a dedicated CAV lane shared by left-turning CAVs and straight-going CAVs. This method usually places the dedicated CAV lane in the leftmost lane. However, the straight-going CAV traffic and the left-turning HV traffic have intersecting and conflicting travel trajectories within the intersection, which poses a safety hazard. The second approach is to reuse the CAV lane and the bus lane. Since the bus lane is usually located on the far right of the road, the CAV lane is also located on the far right lane. However, the straight-through traffic of the CAV and the right-turning traffic of the HV conflict in their driving trajectories within the intersection, posing a safety hazard. The third approach is to set up dedicated lanes for all-turning CAVs. This design has the following problems: it requires a separate blue phase, which increases the cycle time and introduces additional phase transition time, reducing the effective green light time at the intersection; in addition, the dedicated phase will increase vehicle delays in other regular phases, reduce the overall traffic efficiency of the intersection, and affect traffic flow.

[0007] The fourth approach is to set up shared CAV dedicated lanes, with virtual parking lines set up behind the entrance lanes to dynamically adjust the flow of traffic entering the shared dedicated lanes from different directions. However, this approach increases the number of times vehicles stop and does not take into account the waste of road resources caused by setting up dedicated CAV lanes when CAV penetration is low, making it difficult to adapt to changes in CAV penetration rate. Summary of the Invention

[0008] To address the shortcomings of existing technologies, this invention proposes a method for setting up dedicated CAV lanes at arterial intersections in mixed-traffic environments and an adaptive control system. This method not only rationally divides lane functions based on changes in CAV penetration rate, enabling real-time dynamic adjustment of mixed-traffic lane and CAV lane functions, but also achieves coordinated optimization of lane functions and signal control schemes, maximizing the traffic efficiency of arterial intersections and realizing efficient management and control of traffic flow.

[0009] To achieve the above objectives, this invention designs a method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment, which is characterized by including the following steps: S1) Collect traffic flow information, traffic signal status information, and CAV speed and quantity information in real time, and share the collected information to the following steps S2) and S3). S2) Based on the current CAV penetration rate, calculate the trunk line traffic capacity when the CAV dedicated lane is closed, the trunk line traffic capacity when the CAV dedicated lane is open and is a CAV straight-through dedicated lane, and the trunk line traffic capacity when the CAV dedicated lane is open and is a CAV all-turn dedicated lane. Calculate the maximum net traffic capacity benefit from the trunk line traffic capacity under the above three different states, and then set lane function switching judgment conditions by comparing the maximum net traffic capacity benefit to generate a signal control scheme. The lane function switching judgment condition is used to determine whether to open or close the CAV dedicated lane, and to determine whether the CAV dedicated lane is set as a CAV all-turn dedicated lane or a CAV straight-through dedicated lane. The maximum net benefit of the traffic capacity is calculated using the following formula. In the formula, This represents the maximum net gain in accessibility among the three different states. C off (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is closed. C onT (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV all-turn lane. Step S3) Employs a multi-agent deep reinforcement learning algorithm based on PS-PPO to iteratively optimize the signal control scheme generated in step S2), generate the optimal signal control scheme in real time, and output signal control commands.

[0010] Further, in S1), the traffic flow information includes the number of vehicles arriving at and leaving the intersection, the number of vehicles queuing in each direction at the intersection, the average vehicle delay time, and the CAV penetration rate; the traffic signal status information includes the current phase, phase duration, phase sequence, and the execution time of the current phase.

[0011] Furthermore, in S2), the CAV dedicated lane includes an entrance lane and a driving lane, and is located in the first lane to the right of the left-turn dedicated lane; when the CAV dedicated lane is open, only autonomous vehicles are allowed to drive in this lane; when the CAV dedicated lane is closed, all vehicles are allowed to drive in this lane.

[0012] Furthermore, in S2), the lane function switching judgment condition is: If the current lane is a dedicated CAV lane, the traffic capacity in the main direction is... If so, the CAV lane will be closed; If the current lane is a regular straight-ahead lane, the main road's traffic capacity... If the current state continues for more than 3 cycles, the CAV dedicated lane will be activated; in, In the formula, This indicates the hysteresis throughput capacity corresponding to the activated penetration rate. Indicates the activation of penetration rate. This indicates that the CAV dedicated lane is closed, and the trunk-direction traffic capacity is as shown by the opening penetration rate. Indicates the hysteresis width. Equation The solution, This indicates the hysteresis throughput capacity corresponding to the penetration rate that is closed. Indicates the closure of penetration rate. This indicates that the CAV dedicated lane is closed, and the trunk-direction traffic capacity is at the closure penetration rate.

[0013] Furthermore, in S2), when the CAV dedicated lane is activated, the traffic capacity will be adjusted accordingly. The CAV dedicated lane shall be designated as either a dedicated lane for CAV full-turning or a dedicated lane for CAV straight-through traffic. like Then the CAV dedicated lane is set as a CAV straight-through dedicated lane; like Then the CAV dedicated lane is set as a dedicated lane for CAV full steering; The change in traffic capacity Calculated using the following formula In the formula, Indicates changes in traffic capacity. p This represents CAV penetration rate, ranging from [0,1]. This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated lane for CAV full steering.

[0014] Furthermore, in S2), when the lanes in several directions are four lanes, with the middle two lanes being a dedicated CAV lane and a regular through lane respectively, the trunk-direction traffic capacity when the dedicated CAV lane is closed is calculated using the following formula. In the formula, C off (p)This indicates the trunk-direction traffic capacity when the CAV dedicated lane is closed. p This represents CAV penetration rate, ranging from [0,1]. s 1 This represents the saturation flow rate of two straight-ahead mixed traffic lanes, in units of... pcu / h / ln， g / C Indicates the green credit ratio. This represents the average headway of mixed traffic flow, in seconds. This represents the average headway of a single CAV under the current control strategy, in seconds. This represents the average headway of manually driven vehicles, expressed in seconds.

[0015] Furthermore, in S2), when several lanes are four lanes in a certain direction, with the middle two lanes being a dedicated CAV lane and a regular through lane respectively, the trunk-direction traffic capacity when the dedicated CAV lane is open and is a dedicated CAV through lane is calculated by the following formula. C onT (p) This indicates the trunk road traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. p This represents CAV penetration rate, ranging from [0,1]. s 2 This represents the sum of the saturation flow rates of the dedicated CAV straight-ahead lane and the regular straight-ahead lane. This represents the average headway of a single CAV under the current control strategy, in seconds. This represents the average headway of manually driven vehicles, expressed in seconds.

[0016] g / C This indicates the green credit ratio.

[0017] Furthermore, in S2), when the lanes in several directions are four lanes, with the middle two lanes being a dedicated CAV lane and a regular straight-ahead lane respectively, the trunk-direction traffic capacity when the dedicated CAV lane is open and is a dedicated CAV all-turn lane is calculated using the following formula. In the formula, C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV all-turn lane. p This represents CAV penetration rate, ranging from [0,1]. s3 represents the sum of the saturation flow rates of the CAV all-turn lane and the regular straight lane. This indicates the traffic volume of the CAV (Caravan All-Turning) lane in the straight-ahead direction. It is the total traffic volume of the dedicated lane for all turns. g / C This indicates the green credit ratio.

[0018] Furthermore, in S3), the PS-PPO-based multi-agent deep reinforcement learning algorithm defines the algorithm's state space, action space, and reward function, and iteratively optimizes the reinforcement learning algorithm to generate the optimal signal control scheme in real time.

[0019] This invention also designs an adaptive control system for CAV dedicated lanes at arterial intersections in mixed traffic environments, which is characterized by including a central control area, an adjustment area, and a sensing area. The central control area is connected to the intersection at one end and to the adjustment area at the other end. The central control area includes a central control unit, traffic lights, and LED variable message signs. The central control unit sends signal control plans and suggested speeds to CAV vehicles to guide them through the intersection. At the same time, it sends the signal control plans to the LED variable message signs to inform HV vehicles of the lane function. The adjustment area is connected to the central control area at one end and to the sensing area at the other end; all lane lines in the adjustment area are dashed lines, and except for vehicles in the CAV dedicated lane, other vehicles can change lanes in this area according to their driving direction. The sensing area is the road surface area from the end of the adjustment area to the entrance lane, used to collect traffic volume, number of CAVs, traffic flow in different directions, and vehicle speed in real time within the area.

[0020] The advantages of this invention are: 1. This invention sets up a dedicated CAV lane and places it in the first lane to the right of the dedicated left-turn lane. It stipulates that when the dedicated CAV lane is open, only autonomous vehicles are allowed to travel in this lane; when the dedicated CAV lane is closed, all vehicles are allowed to travel in this lane. This reasonable division of lane functions improves the traffic capacity of CAV vehicles and further enhances the traffic capacity of trunk roads in mixed traffic environments. The lane function control method proposed in this invention takes into account the differences in power performance between different vehicles and the differences in traffic capacity between CAV dedicated lanes and mixed lanes. Different lane function settings are adopted under different CAV penetration rates, which fully leverages the advantages of CAV dedicated lanes, avoids the waste of road resources, and effectively improves traffic efficiency while ensuring safety. This invention provides an adaptive control method for arterial intersections with dedicated lanes for connected autonomous vehicles in mixed traffic flow environments. By using lane function switching judgment conditions, the method realizes dynamic changes in lane functions and then coordinates and optimizes these dynamic changes with the signal control scheme, thereby greatly improving the traffic efficiency of arterial intersections. 4. The multi-agent deep reinforcement learning algorithm based on PS-PPO adopted in this invention allows agents at each intersection to share strategies and execute actions according to their current state. This enables agents to learn a general optimal strategy "under similar traffic conditions", which greatly reduces the number of training parameters. This invention discloses a method for setting up CAV (Caravan Driver) lanes at arterial intersections in mixed-traffic environments and an adaptive control system. The CAV lane is positioned in the first lane to the right of the left-turn lane. The lane functions are rationally divided based on changes in CAV penetration rate, and the differences in power performance between different vehicles and the differences in traffic capacity between the CAV lane and the mixed-traffic lane are considered. The functions of the mixed-traffic lane and the CAV lane are dynamically adjusted. Through the coordinated optimization of lane function and signal control schemes, the adaptability of arterial traffic management to future mixed-traffic flows is enhanced, maximizing the traffic efficiency of arterial intersections and achieving efficient traffic flow management and control. This invention has significant innovation and broad application prospects. Attached Figure Description

[0021] Figure 1 This is a flowchart of the method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to the present invention. Figure 2 This is a schematic diagram illustrating the setup of a dedicated lane for autonomous driving at a trunk road intersection according to an embodiment of the present invention. Figure 3 This is a schematic diagram of LED variable lane sign state 1 (CAV dedicated straight lane) according to an embodiment of the present invention; Figure 4This is a schematic diagram of LED variable lane sign state 2 (CAV dedicated all-turn lane) according to an embodiment of the present invention; Figure 5 This is a schematic diagram of LED variable lane sign state 3 (regular straight lane) according to an embodiment of the present invention; Figure 6a This is a schematic diagram of a variable lane driving direction sign (CAV dedicated straight lane) according to an embodiment of the present invention; Figure 6b This is a schematic diagram of a variable lane driving direction sign (CAV dedicated all-turn lane) according to an embodiment of the present invention; Figure 6c This is a schematic diagram of a variable lane driving direction sign (conventional straight lane) according to an embodiment of the present invention; Figure 7 This is a flowchart illustrating the dynamic switching of lane functions according to an embodiment of the present invention; Figure 8 This is a schematic diagram of the architecture of a multi-agent deep reinforcement learning system based on PS-PPO according to an embodiment of the present invention; Figure 9 This is a schematic diagram of the signal control phase sequence according to an embodiment of the present invention. Detailed Implementation

[0022] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0023] In the description of this invention, it should be understood that the terms "length", "width", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", etc., indicate the orientation or positional relationship based on the orientation or positional relationship shown in the accompanying drawings. They are only for the convenience of describing this invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific orientation, or be constructed and operated in a specific orientation. Therefore, they should not be construed as limitations on the invention.

[0024] like Figure 1 As shown, a method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment includes the following steps.

[0025] S1) Sensing traffic flow status, specifically including: real-time collection of traffic flow information, traffic signal status information, and CAV speed and quantity information, and sharing the collected information to the following steps S2) and S3).

[0026] The traffic flow information includes the number of vehicles arriving at and departing from the intersection, the number of vehicles queuing in each direction at the intersection, the average vehicle delay time, and the CAV penetration rate.

[0027] The traffic signal status information includes the current phase, phase duration, phase sequence, and the execution time of the current phase.

[0028] Specifically, traffic flow information is collected in real time through intelligent sensing devices installed on the roadside, such as millimeter-wave radar and camera sensors; traffic signal control information is obtained in real time through networked signal controllers; and CAV speed and quantity information is obtained through information uploaded in real time by CAVs.

[0029] S2) Based on the current CAV penetration rate, calculate the trunk road capacity when the CAV dedicated lane is closed, the trunk road capacity when the CAV dedicated lane is open and serves as a CAV straight-ahead dedicated lane, and the trunk road capacity when the CAV dedicated lane is open and serves as a CAV all-turn dedicated lane. Calculate the maximum net capacity benefit from the trunk road capacity under these three different states. By comparing the maximum net capacity benefit, set lane function switching judgment conditions and generate a signal control scheme. The lane function switching judgment conditions are used to determine whether to open or close the CAV dedicated lane and to determine whether the CAV dedicated lane is set as a CAV all-turn dedicated lane or a CAV straight-ahead dedicated lane. The specific process is attached. Figure 7 As shown.

[0030] The trunk line capacity refers to the combined capacity of several lanes on the trunk line.

[0031] The dedicated CAV lane includes an entrance lane and a driving lane, and is located in the first lane to the right of the dedicated left-turn lane; when the dedicated CAV lane is open, only autonomous vehicles are allowed to travel in this lane; when the dedicated CAV lane is closed, all vehicles are allowed to travel in this lane.

[0032] Under the first condition described above, CAV penetration rate p The value is low, all CAV vehicles are mixed with HV vehicles, and the dedicated CAV lane is closed.

[0033] In the second state described above, CAV penetration rate p As the value gradually increases, right-turning CAV vehicles are located in the right-turn lane, while straight-going CAV vehicles are located in the dedicated straight-going CAV lane.

[0034] In the third state described above, CAV penetration rate p The value continued to increase, and all CAV vehicles were placed in the dedicated CAV all-steering lane.

[0035] The maximum net benefit of the traffic capacity is calculated using the following formula. In the formula, This represents the maximum net gain in accessibility among the three different states. C off (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is closed. C onT (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated lane for CAV full steering.

[0036] like This indicates that, theoretically, the CAV dedicated lane can improve the traffic capacity of this entrance; if If so, it should not be turned on.

[0037] This application solves the equation: If the equation is in There is a solution Then The critical permeability is used as the threshold, and the hysteresis width is used to give the on / off permeability to prevent frequent switching.

[0038] The open and closed permeability rates are expressed by the following formulas: In the formula, Indicates the activation of penetration rate. Indicates the closure of penetration rate. This represents the hysteresis width, which is typically between (0.03, 0.10).

[0039] The hysteresis throughput corresponding to the open and closed penetration rates are respectively In the formula, This indicates the hysteresis throughput capacity corresponding to the activated penetration rate. Indicates the activation of penetration rate. This indicates that the CAV dedicated lane is closed, and the trunk-direction traffic capacity is as shown by the opening penetration rate. This indicates the hysteresis throughput capacity corresponding to the penetration rate that is closed. Indicates the closure of penetration rate. This indicates that the CAV dedicated lane is closed, and the trunk-direction traffic capacity is at the closure penetration rate.

[0040] The lane function switching judgment condition is: If the current lane is a dedicated CAV lane, the traffic capacity in the main direction is... If so, the CAV lane will be closed; If the current lane is a regular straight-ahead lane, the main road's traffic capacity... If the current state continues for more than 3 cycles, the CAV dedicated lane will be activated; Then based on changes in traffic capacity Determine whether the CAV dedicated lane is set as a dedicated lane for CAV all-turning or a dedicated lane for CAV straight-ahead travel. If Then the CAV dedicated lane is set as a dedicated lane for CAV straight-through traffic; if Then the CAV dedicated lane is set as a dedicated lane for CAV all-turning vehicles.

[0041] The change in traffic capacity Calculated using the following formula In the formula, Indicates changes in traffic capacity. p This represents CAV penetration rate, ranging from [0,1]. This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated lane for CAV full steering.

[0042] In this embodiment, the appendix is ​​attached. Figure 2 The following explanation will be given as an example. Figure 2 In the middle, the main road has four lanes. The first lane from the left is the left-turn lane, the second lane from the left is the CAV dedicated lane (CAV dedicated straight lane, CAV dedicated full-turn lane or regular straight lane), the third lane from the left is the regular straight lane, and the fourth lane from the left is the right-turn lane.

[0043] When CAV penetration rate p When the value is low, all CAV vehicles are mixed with HV vehicles, and the dedicated CAV lane is closed.

[0044] The trunk-direction traffic capacity when the CAV dedicated lane is closed is calculated using the following formula. In the formula, C off (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is closed. p This represents CAV penetration rate, ranging from [0,1]. s 1 This represents the saturation flow rate of two straight-ahead mixed traffic lanes, in units of... pcu / h / ln， g / C Indicates the green credit ratio. This represents the average headway of mixed traffic flow, in seconds. This represents the average headway of a single CAV under the current control strategy, in seconds. This represents the average headway of manually driven vehicles, expressed in seconds.

[0045] When CAV penetration rate p As the value gradually increases, right-turning CAVs share the fourth lane from the left, and when straight-going CAVs are in the dedicated CAV straight-going lane, one lane on the main road is a dedicated CAV lane, and the other is a regular straight-going lane. Figure 2 As shown. The saturation flow rate in the main direction is the sum of the saturation flow rates of the CAV dedicated straight-ahead lane and the regular straight-ahead lane.

[0046] Therefore, the trunk-direction traffic capacity when the CAV dedicated lane is open and is a CAV straight-through dedicated lane is calculated by the following formula. C onT (p) This indicates the trunk road traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. p This represents CAV penetration rate, ranging from [0,1]. s 2 This represents the sum of the saturation flow rates of the dedicated CAV straight-ahead lane and the regular straight-ahead lane. This represents the average headway of a single CAV under the current control strategy, in seconds. This represents the average headway of manually driven vehicles, expressed in seconds.

[0047] g / C This indicates the green credit ratio.

[0048] When CAV penetration rate p As the value continues to increase, all CAV vehicles are located in the dedicated CAV all-turn lane. One of the straight lanes in the trunk direction is the dedicated CAV all-turn lane, and the other is the regular straight lane. The saturation flow rate in the trunk direction is the sum of the saturation flow rates of the dedicated CAV all-turn lane and the regular straight lane.

[0049] Therefore, the trunk-direction traffic capacity when the CAV dedicated lane is open and is a CAV all-turn dedicated lane is calculated by the following formula. In the formula, C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV all-turn lane. p This represents CAV penetration rate, ranging from [0,1]. s3 represents the sum of the saturation flow rates of the CAV all-turn lane and the regular straight lane. This indicates the traffic volume of the CAV (Caravan All-Turning) lane in the straight-ahead direction. It is the total traffic volume of the dedicated lane for all turns. g / C This indicates the green credit ratio.

[0050] By calculating the trunk-direction saturated flow rate of three lane function settings under different CAV penetration rates, the optimal lane function setting scheme for different penetration rates is determined.

[0051] Step S2) is used to control the switching between the CAV dedicated lane and the regular straight lane, so as to maximize the use of green light time and improve traffic efficiency.

[0052] Step S3) Employs a multi-agent deep reinforcement learning algorithm based on PS-PPO (parameter-shared proximal policy optimization) to iteratively optimize the signal control scheme generated in step S2), generate the optimal signal control scheme in real time, and output signal control commands.

[0053] The experience of all agents is used to jointly train the shared policy and evaluate the network. By defining the algorithm's state space, action space, and reward function, the reinforcement learning algorithm iteratively optimizes the algorithm, generating the optimal signal control scheme in real time and outputting signal control commands. (See attached) Figure 8 As shown.

[0054] The state space definition includes traffic flow characteristics, traffic signal states, and lane function vectors.

[0055] The traffic flow characteristics include: the number of vehicles queuing in each direction, average vehicle delay, and CAV penetration rate. The traffic signal status includes: current phase, phase duration, phase sequence, and the execution time of the current phase. The lane function vector... (0 represents a regular straight-ahead lane, 1 represents a dedicated straight-ahead lane for CAVs, and 2 represents a dedicated lane for CAVs with full steering).

[0056] The action space definition includes the green light extension time for selecting the next phase or adjusting the current phase. A schematic diagram of the signal control phase sequence in this embodiment is attached. Figure 9 As shown.

[0057] The reward function is defined by including the queue length at intersection approach lanes and the sum of vehicle delays across all intersections. It employs a combination of local rewards and arterial coordination rewards, expressed by the following formula: This represents the numerical value of the reward function. This indicates a local reward system that minimizes queue length and average vehicle delay at this intersection. This indicates the overall reward, specifically the total traffic volume on the main lines. This represents a penalty to prevent jitter caused by frequent phase switching. w 1 , w 2 , w 3 These represent the weights.

[0058] The present invention also designs an adaptive control system for CAV dedicated lanes at arterial intersections in mixed traffic environments, including a central control area, an adjustment area, and a perception area.

[0059] The central control area is connected to the intersection at one end and to the adjustment area at the other end. The central control area includes a central control unit, traffic lights, and LED variable message signs. The central control unit sends signal control plans and suggested speeds to CAV vehicles to guide them through the intersection. At the same time, it sends the signal control plans to the LED variable message signs to inform HV vehicles of the lane function.

[0060] Specifically, the LED variable information board includes a primary LED variable information board and a secondary LED variable information board. The primary LED variable information board is an LED variable lane sign, which is set between the adjustment area and the sensing area. The secondary LED variable information board is a variable lane driving direction sign, which is set before the intersection entrance lane.

[0061] The adjustment zone is connected to the central control zone at one end and to the sensing zone at the other end. All lane lines in the adjustment zone are dashed lines. Except for vehicles in the CAV dedicated lane, other vehicles can change lanes in this area according to their driving direction.

[0062] The sensing area is the road surface area from the end of the adjustment area to the entrance lane, used to collect traffic flow information such as traffic volume, number of CAVs, traffic flow in different directions, and vehicle speed in real time.

[0063] As attached Figure 2 As shown, in this embodiment, the main road intersections of the dedicated autonomous driving lanes are urban arterial roads with no fewer than eight lanes in both directions, and there are no fewer than three intersections, including a central control area, an adjustment area, and a perception area. Arriving connected autonomous vehicles form a CAV convoy or a mixed CAV and HV convoy in the adjustment area and wait to enter the central control area.

[0064] Figure 2 The number of central control area, adjustment area and sensing area is 1 each.

[0065] The dedicated lane for autonomous driving includes an entrance lane and a driving lane. It is located in the first lane to the right of the dedicated left-turn lane and only autonomous vehicles are allowed to travel in this lane.

[0066] The central control area is connected to the intersection at one end and to the adjustment area at the other end. The central control area is equipped with LED variable message signs, traffic lights, and a central control unit. Within this area, the CAV vehicle passes through the intersection according to the signal control plan issued by the central control unit. The suggested speed is issued to the CAV vehicle by the central control unit and then communicated to the HV (manual driver) vehicle via the two-level LED variable message signs.

[0067] Specifically, the primary LED variable information panel is an LED variable lane sign, positioned between the adjustment zone and the sensing zone. It is a gantry structure spanning four lanes, with the second lane's sign panel being an LED display screen that dynamically shows lane functions. For example... Figure 3 , Figure 4 , Figure 5 As shown, Figure 3 LED variable lane sign status 1. Figure 4 The LED variable lane sign is in state 2. Figure 5 The LED variable lane sign is in state 3.

[0068] from Figure 3 It can be seen that the LED variable lane sign is in effect when the second lane is used as a dedicated straight-ahead lane for CAVs. From Figure 4 It can be seen that the LED variable lane sign status is as follows when the second lane is used as a dedicated all-steering lane for CAVs. From Figure 5 It can be seen that the LED variable lane sign is in effect when the second lane is used as a regular straight lane.

[0069] The secondary LED variable message sign is a variable lane direction indicator, installed before the intersection approach lane. It is a cantilevered sign, with the second lane's sign panel featuring a dynamic LED display screen. For example... Figures 6a-6c The image shows three states of the variable lane driving direction sign: the variable lane driving direction sign state when the second lane is used as a dedicated straight lane for CAVs, the variable lane driving direction sign state when the second lane is used as a dedicated all-turn lane for CAVs, and the variable lane driving direction sign state when the second lane is used as a regular straight lane.

[0070] The adjustment zone is connected to the central control area at one end and the perception area at the other end. The lane lines within the adjustment zone are dashed lines. Except for vehicles in the dedicated autonomous driving lane, other vehicles can change lanes within this area according to their travel direction. The length of the adjustment zone is set according to the road segment length and traffic volume. In this embodiment, the length of the adjustment zone is 30 to 50 meters. If the road segment is long and the traffic volume is large, the length of the adjustment zone can be appropriately extended.

[0071] The perception zone is the road surface area from the end of the adjustment zone to the entrance lane. Through intelligent sensing devices set up on the roadside, traffic flow information such as traffic volume, number of CAVs, traffic flow in different directions, and vehicle speed are collected in real time within the area. This area includes dedicated lanes for autonomous driving and mixed traffic lanes.

[0072] The lane lines within the designated entrance area are solid lines, and all vehicles are prohibited from changing lanes within this area.

[0073] This invention discloses a method for setting up CAV (Caravan Driver) lanes at arterial intersections in mixed-traffic environments and an adaptive control system. The CAV lane is positioned in the first lane to the right of the left-turn lane. The lane functions are rationally divided based on changes in CAV penetration rate, and the differences in power performance between different vehicles and the differences in traffic capacity between the CAV lane and the mixed-traffic lane are considered. The functions of the mixed-traffic lane and the CAV lane are dynamically adjusted. Through the coordinated optimization of lane function and signal control schemes, the adaptability of arterial traffic management to future mixed-traffic flows is enhanced, maximizing the traffic efficiency of arterial intersections and achieving efficient traffic flow management and control. This invention has significant innovation and broad application prospects.

[0074] The above embodiments are preferred embodiments of the present invention, but the embodiments of the present invention are not limited to the above embodiments. Any changes, modifications, substitutions, combinations, or simplifications made without departing from the spirit and principle of the present invention shall be considered equivalent substitutions and shall be included within the protection scope of the present invention.

Claims

1. A method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment, characterized in that, Includes the following steps: S1) Collect traffic flow information, traffic signal status information, and CAV speed and quantity information in real time, and share the collected information to the following steps S2) and S3). S2) Based on the current CAV penetration rate, calculate the trunk line traffic capacity when the CAV dedicated lane is closed, the trunk line traffic capacity when the CAV dedicated lane is open and is a CAV straight-through dedicated lane, and the trunk line traffic capacity when the CAV dedicated lane is open and is a CAV all-turn dedicated lane. Calculate the maximum net traffic capacity benefit from the trunk line traffic capacity under the above three different states, and then set lane function switching judgment conditions by comparing the maximum net traffic capacity benefit to generate a signal control scheme. The lane function switching judgment condition is used to determine whether to open or close the CAV dedicated lane, and to determine whether the CAV dedicated lane is set as a CAV all-turn dedicated lane or a CAV straight-through dedicated lane. The maximum net benefit of the traffic capacity is calculated using the following formula. In the formula, This represents the maximum net gain in accessibility among the three different states. C off (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is closed. C onT (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV all-turn lane. Step S3) Employs a multi-agent deep reinforcement learning algorithm based on PS-PPO to iteratively optimize the signal control scheme generated in step S2), generate the optimal signal control scheme in real time, and output signal control commands.

2. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 1, characterized in that: In S1), the traffic flow information includes the number of vehicles arriving at and leaving the intersection, the number of vehicles queuing in each direction of the intersection, the average vehicle delay time, and the CAV penetration rate; the traffic signal status information includes the current phase, phase duration, phase sequence, and the execution time of the current phase.

3. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 1, characterized in that: In S2), the CAV dedicated lane includes an entrance lane and a driving lane, and is located in the first lane to the right of the left-turn dedicated lane; when the CAV dedicated lane is open, only autonomous vehicles are allowed to drive in this lane; when the CAV dedicated lane is closed, all vehicles are allowed to drive in this lane.

4. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 3, characterized in that: In S2), the lane function switching judgment condition is: If the current lane is a dedicated CAV lane, the traffic capacity in the main direction is... If so, the CAV lane will be closed; If the current lane is a regular straight-ahead lane, the main road's traffic capacity... If the current state continues for more than 3 cycles, the CAV dedicated lane will be activated; in, In the formula, This indicates the hysteresis throughput capacity corresponding to the activated penetration rate. Indicates the activation of penetration rate. This indicates that the CAV dedicated lane is closed, and the trunk-direction traffic capacity is as shown by the opening penetration rate. Indicates the hysteresis width. Equation The solution, This indicates the hysteresis throughput capacity corresponding to the penetration rate that is closed. Indicates the closure of penetration rate. This indicates that the CAV dedicated lane is closed, and the trunk-direction traffic capacity is at the closure penetration rate.

5. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 4, characterized in that: In S2), when the CAV dedicated lane is opened, the traffic capacity will be adjusted accordingly. The CAV dedicated lane shall be designated as either a dedicated lane for CAV full-turning or a dedicated lane for CAV straight-through traffic. like Then the CAV dedicated lane is set as a CAV straight-through dedicated lane; like Then the CAV dedicated lane is set as a dedicated lane for CAV full steering; The change in traffic capacity Calculated using the following formula In the formula, Indicates changes in traffic capacity. p This represents CAV penetration rate, ranging from [0,1]. This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated lane for CAV full steering.

6. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 1, characterized in that: In S2), when a certain number of lanes in a certain direction are four lanes, with the middle two lanes being a dedicated CAV lane and a regular through lane respectively, the trunk-direction traffic capacity when the dedicated CAV lane is closed is calculated by the following formula. In the formula, C off (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is closed. p This represents CAV penetration rate, ranging from [0,1]. s 1 This represents the saturation flow rate of two straight-ahead mixed traffic lanes, in units of... pcu / h / ln, g / C Indicates the green credit ratio. This represents the average headway of mixed traffic flow, in seconds. This represents the average headway of a single CAV under the current control strategy, in seconds. This represents the average headway of manually driven vehicles, expressed in seconds.

7. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 6, characterized in that: In S2), when several lanes in a certain direction are four lanes, with the middle two lanes being a dedicated CAV lane and a regular through lane respectively, the trunk direction traffic capacity when the dedicated CAV lane is open and is a dedicated CAV through lane is calculated by the following formula. C onT (p) This indicates the trunk road traffic capacity when the CAV dedicated lane is open and is a dedicated CAV straight-ahead lane. p This represents CAV penetration rate, ranging from [0,1]. s 2 This represents the sum of the saturation flow rates of the dedicated CAV straight-ahead lane and the regular straight-ahead lane. This represents the average headway of a single CAV under the current control strategy, in seconds. This represents the average headway of manually driven vehicles, in seconds. g / C This indicates the green credit ratio.

8. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 7, characterized in that: In S2), when several lanes in a certain direction are four lanes, with the middle two lanes being a dedicated CAV lane and a regular straight-through lane respectively, the trunk-direction traffic capacity when the dedicated CAV lane is open and is a dedicated CAV all-turn lane is calculated by the following formula. In the formula, C onA (p) This indicates the trunk-direction traffic capacity when the CAV dedicated lane is open and is a dedicated CAV all-turn lane. p This represents CAV penetration rate, ranging from [0,1]. s3 represents the sum of the saturation flow rates of the CAV all-turn lane and the regular straight lane. This indicates the traffic volume of the CAV (Caravan All-Turning) lane in the straight-ahead direction. It is the total traffic volume of the dedicated lane for all turns. g / C This indicates the green credit ratio.

9. The method for setting up a dedicated CAV lane at a trunk intersection in a mixed-traffic environment according to claim 1, characterized in that: In S3), the PS-PPO-based multi-agent deep reinforcement learning algorithm defines the algorithm's state space, action space, and reward function, and iteratively optimizes the reinforcement learning algorithm to generate the optimal signal control scheme in real time.

10. An adaptive control system for dedicated CAV lanes at arterial intersections in mixed-traffic environments, characterized in that, Includes a central control area, an adjustment area, and a sensing area; The central control area is connected to the intersection at one end and to the adjustment area at the other end. The central control area includes a central control unit, traffic lights, and LED variable message signs. The central control unit sends signal control plans and suggested speeds to CAV vehicles to guide them through the intersection. At the same time, it sends the signal control plans to the LED variable message signs to inform HV vehicles of the lane function. The adjustment area is connected to the central control area at one end and to the sensing area at the other end; all lane lines in the adjustment area are dashed lines, and except for vehicles in the CAV dedicated lane, other vehicles can change lanes in this area according to their driving direction. The sensing area is the road surface area from the end of the adjustment area to the entrance lane, used to collect traffic volume, number of CAVs, traffic flow in different directions, and vehicle speed in real time within the area.

Images