A multi-center trajectory-level active traffic control system and method thereof

Abstract

The present invention relates to a multi-center trajectory-level active traffic control system and method thereof. The system includes a data acquisition module, a decision generation module, and a command execution module connected in sequence. The data acquisition module is used to collect trajectory information of networked autonomous driving vehicles in a road network as input to the decision generation module. The decision generation module, based on the input vehicle trajectory information and with the goal of multi-segment coordination, determines trajectory control instructions that can dynamically adjust the instantaneous traffic flow rate of each road section in the road network. The command execution module is used to issue the trajectory control instructions to the networked autonomous driving vehicles on each road section, so that the networked autonomous driving vehicles execute the corresponding trajectory control commands. Compared with the existing technology, the present invention can perform accurate and real-time traffic flow prediction, can promptly and accurately execute multi-center control instructions, and can accurately perform trajectory-level traffic control based on traffic capacity in different scenarios.

Description

Technical Field

[0001] The present invention relates to the field of intelligent traffic control technology, and in particular to a multi-centralized trajectory-level active traffic control system and method thereof. Background Art

[0002] With the rapid growth of cities, the number of cars has increased dramatically, leading to an increasingly prominent conflict between road construction and vehicle ownership, and worsening traffic congestion. Active traffic management, a temporary traffic management method based on real-time traffic condition prediction, plays a significant role in alleviating traffic congestion. By adjusting traffic demand to match traffic supply, active traffic management reduces the mismatch between supply and demand in the transportation system, thereby achieving the goal of alleviating traffic congestion.

[0003] Existing active traffic management methods primarily involve continuous traffic measurement and data analysis; evaluating the operational status of road networks, making short-term traffic forecasts, and making corresponding control strategy decisions. However, in practical applications, the following problems remain: First, it is difficult to accurately predict traffic demand. Traditional demand forecasting methods rely heavily on steady-state traffic flow on road sections. This means that traffic flow forecasts can only be made when traffic flow on long sections reaches steady state. Furthermore, operating conditions are relatively stable, and the scope of application is relatively limited, making it difficult to accurately and timely predict traffic flow. Second, traffic control measures are difficult to implement promptly and accurately on traffic entities, making it impossible to achieve behavioral control at the micro level and ensuring the efficient implementation of control measures. In summary, these shortcomings significantly reduce the effectiveness of active traffic management and can easily have a serious negative impact on existing transportation systems, hindering the improvement of vehicle cruising comfort and the stability of traffic flow. Summary of the Invention

[0004] The purpose of the present invention is to overcome the defects of the above-mentioned existing technologies and provide a multi-centralized trajectory-level active traffic control system and method thereof. Based on networked autonomous driving technology, it can accurately control traffic demand at the trajectory level, and effectively improve the control capability of traffic demand in terms of control range, efficiency and accuracy.

[0005] The objectives of the present invention can be achieved by the following technical solutions: a multi-centralized trajectory-level active traffic control system, comprising a data acquisition module, a decision generation module, and a command execution module connected in sequence, wherein the data acquisition module is used to collect trajectory information of networked autonomous driving vehicles in a road network as input to the decision generation module;

[0006] The decision generation module determines, based on the input vehicle trajectory information and with the goal of multi-segment coordination, trajectory control instructions that can dynamically adjust the instantaneous traffic flow rate of each road segment in the road network;

[0007] The command execution module is used to send trajectory control instructions to the connected autonomous driving vehicles on each road section, so that the connected autonomous driving vehicles execute the corresponding trajectory control commands.

[0008] Furthermore, the decision generation module includes a terminal segment decision maker, multiple intermediate segment decision makers, and multiple multi-segment coordinated instruction generators. The terminal segment decision maker generates a terminal segment traffic flow target adjustment flow rate decision plan based on the downstream bottleneck segment traffic capacity, which serves as a basis for generating coordinated instructions for the upstream intermediate segment multi-segment coordinated instruction generator.

[0009] The multi-section coordinated instruction generator transmits decision information from one section unit to the upstream adjacent section unit according to the target of traffic flow rate control of the end section, forming a multi-section coordinated instruction, which is the decision basis of each intermediate section decision maker;

[0010] The intermediate section decision maker is used to control the unit traffic flow of the intermediate section. Based on the traffic flow rate of the last section, multi-section coordination instructions and vehicle information on related sections, it forms a decision that can optimize the traffic flow rate of the intermediate section, thereby reducing the shock fluctuations generated during traffic operation.

[0011] Furthermore, the optimization objectives of the end-segment decision maker include: discomfort during vehicle cruising, oscillations and fluctuations in traffic flow status, and tracking errors in traffic flow rate. The optimization objectives of the end-segment decision maker are specifically:

[0012]

[0013] Among them, k and ω correspond to the time step index and the target vehicle number index respectively, ω ′ is the non-target vehicle number index, K and Ω correspond to the time step index set and vehicle index set respectively, a ω (k) and v ω (k) are the instantaneous acceleration and instantaneous speed of vehicle ω at time step k, q is the predicted traffic flow rate at the end of the optimization time domain, q t is the target expected flow rate, α, β and γ correspond to the target weight of discomfort during vehicle cruising, the target weight of oscillation fluctuation during traffic flow operation and the target weight of traffic flow rate tracking error, respectively.

[0014] Furthermore, the optimization objectives of the intermediate section decision maker include: discomfort during vehicle cruising and oscillation fluctuations in traffic flow. The optimization objectives of the intermediate section decision maker are specifically:

[0015]

[0016] Among them, a ω (k) and v ω (k) are the instantaneous acceleration and instantaneous speed of the target vehicle ω at time step k, ω′ is the non-target vehicle number index, K and Ω correspond to the time step index set and vehicle index set, respectively, α and β are the target weights of discomfort during vehicle cruising and the target weights of oscillation fluctuations caused by the state during traffic flow operation, respectively.

[0017] Furthermore, the constraints of the intermediate section decision maker include operational constraints of the connected autonomous driving vehicle and lane management constraints;

[0018] The constraints of the end-section decision maker include operational constraints of connected autonomous driving vehicles, lane management constraints, mileage constraints, and upstream section capacity constraints.

[0019] Furthermore, the kinematic constraints of the connected autonomous driving vehicle are specifically:

[0020]

[0021]

[0022]

[0023]

[0024] x ω (k)-x ω′ (k)-L ω ≥τ0v ω′ (k)+ζ0if x ω (k)≥x ω′ (k),

[0025]

[0026]

[0027]

[0028]

[0029]

[0030] (f ω,ω′ (k)-1)·M≤x ω (k)-x ω′ (k)≤f ω,ω′ (k)·M

[0031]

[0032] Among them, a ω (k) and v ω (k) represent the instantaneous acceleration and instantaneous velocity of vehicle ω at time step k, and are the minimum and maximum operating speeds of the vehicle, respectively. and are the maximum deceleration and maximum acceleration, respectively, x ω (k) is the position of vehicle ω at time step k, x ω (0), v ω (0) and a ω (0) are the position, velocity and acceleration of the vehicle ω at the initial moment, l ω is the length of vehicle ω, τ0 is the safe vehicle headway, ζ0 is the safe vehicle spacing, f ω,ω′ (k) is a binary variable indicator. If the target vehicle ω is in front of the vehicle ω′, then f ω,ω′ (k)=1, otherwise, f ω,ω′ (k)=0.

[0033] Furthermore, the lane management constraints of the connected autonomous driving vehicle include lane change feasibility constraints, lane reduction constraints, and boundary constraints for vehicle exit sections, specifically:

[0034]

[0035]

[0036]

[0037]

[0038]

[0039] Where M is a constant, j and j ′ is the index of different lane numbers, J is the lane number index set, is a lane selection indicator, if vehicle ω is in lane k at time step k, then Otherwise it is 0, L b is the location of the lane reduction area, J b is the set of lane indexes of lanes closed in the lane reduction area, L d It is the starting boundary position of the downstream segment unit.

[0040] Furthermore, the mileage constraint is specifically:

[0041]

[0042] D ω ≥x ω (k max )-L u ,ω∈Ω I

[0043] D ω ≥0,ω∈Ω I

[0044] D ω ≤(1-λ ω )·M+x ω (k max )-L u ,ω∈Ω I

[0045] D ω ≤λ ω ·M,ω∈Ω I

[0046] D ω ≤x ω (k max )-L u ,ω∈Ω s

[0047] D ω ≤L d -L u ,ω∈Ω s

[0048] D ω ≥(1-λ ω )·M+x ω (k max )-L u ,ω∈Ω s

[0049] D ω ≥λ ω M+L d -L u ,ω∈Ω s

[0050] Among them, D ω is the mileage of vehicle ω, k max is the final time step in the optimization time domain, L u is the end boundary position of the upstream section unit, Ω I Represents the vehicle number index set that will pass through the upstream, Ω s Represents the vehicle number index set that is about to leave this road section unit, λ ω is a binary variable indicator. If the vehicle ω travels mileage D ωEqual to the position x of vehicle ω at the end of the optimization time domain ω (k max ) minus the end boundary position L of the upstream section unit u , then λ ω =1, otherwise, λ ω =0, M is a constant.

[0051] Furthermore, the upstream section traffic capacity constraint is specifically:

[0052]

[0053] Among them, C is the traffic capacity of the upstream section unit, and the left term represents the traffic volume of vehicles crossing the boundary of the downstream section unit.

[0054] Furthermore, the multi-segment coordinated instruction generator is used to constrain the state of the connected autonomous driving vehicle that crosses the downstream boundary of the target segment unit. The specific form of the constraint is:

[0055]

[0056]

[0057]

[0058]

[0059]

[0060] in, is the time step when vehicle ω crosses the downstream boundary of the road segment, The position of vehicle ω when it crosses the downstream boundary of the road segment, is the speed of vehicle ω crossing the downstream boundary of the road segment, is the lane selection when vehicle ω crosses the downstream boundary of the road segment, Ω II Represents the vehicle number index set that is about to cross the downstream boundary of the road segment unit.

[0061] A multi-center trajectory-level active traffic control method includes the following steps:

[0062] S1, the data acquisition module collects the trajectory information of the connected autonomous driving vehicles in the road network in real time;

[0063] S2, the end segment decision maker determines the end segment vehicle trajectory plan based on the trajectory information and the preset bottleneck traffic information;

[0064] S3, a multi-section coordinated instruction generator transmits decision information from one section unit to the upstream adjacent section units according to the target of traffic flow rate control of the last section, forming a multi-section coordinated instruction;

[0065] S4: The intermediate segment decision maker determines the vehicle trajectory plan for the current intermediate segment under the constraints of the vehicle information and coordination instructions on the segment. When the plan is transmitted to the upstream segment and the corresponding decision is made, a trajectory control instruction with multi-segment coordination is formed.

[0066] S5. The command execution module sends the trajectory control instructions to the connected autonomous driving vehicles on each road section, so that the connected autonomous driving vehicles execute the corresponding trajectory control commands.

[0067] Compared with the prior art, the present invention has the following advantages:

[0068] First, based on connected autonomous driving technology, this invention divides long road sections, which originally require steady-state traffic flow for flow prediction, into several smaller sections where steady-state traffic flow can be predicted without strictly requiring it. By sensing and collecting microscopic information from connected autonomous vehicles, combined with a designed data acquisition module, decision generation module, and command execution module, the data acquisition module collects trajectory information such as position, speed, and acceleration from connected autonomous vehicles in the road network. The decision generation module uses vehicle trajectory information as input and aims to coordinate multiple road sections to generate trajectory control instructions that can dynamically adjust the instantaneous traffic flow rate of each road section in the road network. The command execution module then issues the trajectory control instructions to connected autonomous vehicles on each road section, enabling them to execute the corresponding trajectory control instructions. This enables real-time and precise trajectory-level traffic control based on road capacity in different scenarios.

[0069] 2. The present invention designs a decision generation module to include a terminal section decision maker, a multi-section coordinated instruction generator and an intermediate section decision maker. For the downstream bottleneck section, the terminal section decision maker is used to generate a terminal section traffic flow target adjustment flow rate decision plan; the multi-section coordinated instruction generator transmits the decision information from one section unit to the upstream adjacent section unit according to the target of the terminal section traffic flow rate control, and forms a multi-section coordinated instruction, which is the decision basis for each intermediate section decision maker; the intermediate section decision maker is used to control the traffic flow of the intermediate section unit, and based on the terminal section traffic flow rate, multi-section coordinated instructions and vehicle information on the relevant sections, a decision is formed that can optimize the adjustment of the intermediate section traffic flow rate, thereby reducing the shock fluctuations generated during the traffic operation process. The present invention greatly improves the control capability of traffic demand in terms of control range, efficiency and accuracy.

[0070] 3. The present invention designs the optimization objectives and constraints of the end-segment decision maker and the middle-segment decision maker, and designs a multi-segment collaborative instruction generator to constrain the state of the connected autonomous driving vehicle crossing the downstream boundary of the target segment unit. It adopts a multi-center optimization method and performs multi-center solution for the end segment and the middle segment. The connected autonomous driving vehicles on each sub-segment can realize behavior control at the micro level, and can execute the multi-center control instructions in a timely and accurate manner, achieving the effect of trajectory-level active traffic control. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 Schematic diagram of the system structure of the present invention;

[0072] Figure 2 Schematic diagram of the method flow of the present invention;

[0073] Figure 3 Schematic diagram of the application framework of the embodiment;

[0074] Description of the marks in the figure: A, data acquisition module, B, decision generation module, C, command execution module. DETAILED DESCRIPTION

[0075] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0076] Example

[0077] like Figure 1 As shown in FIG, a multi-centralized trajectory-level active traffic control system includes a data acquisition module A, a decision generation module B, and a command execution module C connected in sequence. The data acquisition module A is used to collect trajectory information of networked autonomous driving vehicles in the road network as input to the decision generation module B.

[0078] Decision generation module B, based on the input vehicle trajectory information and with the goal of multi-segment coordination, determines trajectory control instructions that can dynamically adjust the instantaneous traffic flow rate of each section in the road network;

[0079] The command execution module C is used to send trajectory control instructions to the connected autonomous driving vehicles on each road section, so that the connected autonomous driving vehicles execute the corresponding trajectory control commands.

[0080] The decision generation module B includes a terminal segment decision maker, multiple intermediate segment decision makers, and multiple multi-segment coordinated instruction generators. The terminal segment decision maker generates a target traffic flow rate adjustment decision plan for the terminal segment based on the capacity of the downstream bottleneck segment, which serves as the basis for generating coordinated instructions for the upstream intermediate segment multi-segment coordinated instruction generator.

[0081] The multi-section coordinated instruction generator transmits decision information from one section unit to the upstream adjacent section units according to the traffic flow rate control target of the terminal section, forming a multi-section coordinated instruction. This coordinated instruction is the basis for the decision-making of each intermediate section decision maker.

[0082] The intermediate section decision maker is used to control the traffic flow of the intermediate section unit. Based on the traffic flow rate of the last section, multi-section coordination instructions and vehicle information on related sections, it forms a decision that can optimize the traffic flow rate of the intermediate section, thereby reducing the oscillations and fluctuations generated during traffic operation.

[0083] The above system is applied in practice to realize a multi-center trajectory-level active traffic control method, such as Figure 2 As shown, the following steps are included:

[0084] S1, the data acquisition module collects the trajectory information of the connected autonomous driving vehicles in the road network in real time;

[0085] S2, the end segment decision maker determines the end segment vehicle trajectory plan based on the trajectory information and the preset bottleneck traffic information;

[0086] S3, a multi-section coordinated instruction generator transmits decision information from one section unit to the upstream adjacent section units according to the target of traffic flow rate control of the last section, forming a multi-section coordinated instruction;

[0087] S4: The intermediate segment decision maker determines the vehicle trajectory plan for the current intermediate segment under the constraints of the vehicle information and coordination instructions on the segment. When the plan is transmitted to the upstream segment and the corresponding decision is made, a trajectory control instruction with multi-segment coordination is formed.

[0088] S5. The command execution module sends the trajectory control instructions to the connected autonomous driving vehicles on each road section, so that the connected autonomous driving vehicles execute the corresponding trajectory control commands.

[0089] This embodiment applies the above technical solution, such as Figure 3 The main contents are:

[0090] Data acquisition module A, which collects trajectory information such as position, speed, and acceleration of connected autonomous vehicles on the road network for input into decision-making module B;

[0091] Decision generation module B is divided into three submodules: the end-segment decider, the multi-segment coordination instruction generator, and the intermediate segment decider. Its function is to take the vehicle trajectory information provided by data acquisition module A as input. The end-segment decider first determines the vehicle trajectory plan for the end segment based on the collected vehicle information and bottleneck capacity. The coordination scope of the multi-segment coordination instruction generator gradually transfers the coordination scope from the end segment decision to the upstream intermediate segments to achieve the goal of multi-segment coordination. The intermediate segment decider uses the vehicle information of the current segment and the coordination instructions generated by the multi-segment coordination instruction generator as constraints to determine the vehicle trajectory plan for the current intermediate segment. When this plan is transferred to the upstream segment and the corresponding decision is made, a trajectory control instruction with multi-segment coordination is formed.

[0092] The command execution module C takes the trajectory control instructions generated by the decision generation module B as input and sends the trajectory control instructions to the networked autonomous driving vehicles on each road section, so that the networked autonomous driving vehicles can execute the corresponding trajectory control instructions.

[0093] In the decision generation module B, a multi-center optimization method is used to carry out multi-center solutions for the end section and the middle section. The solution form is as follows:

[0094]

[0095]

[0096] Among them, J1 is the optimization control target of the decision maker of the last section, J2 is the optimization control target of the decision maker of the middle section, K and ω are the time step index and target vehicle number index respectively, ω′ is the non-target vehicle number index, K and Ω are the time step index set and vehicle index set respectively;

[0097] a ω (k) and v ω (k) represents the instantaneous acceleration and instantaneous speed of vehicle ω at time step k, q is the predicted traffic flow rate at the end of the optimization time domain, and q t is the target expected flow rate, α, β and γ are the target weights of discomfort during vehicle cruising, the target weights of oscillation fluctuations in the state during traffic flow operation and the target weights of traffic flow rate tracking error, respectively.

[0098] To ensure the authenticity of the optimization control model and the safe driving of the connected autonomous driving vehicle on all road sections, the following constraints are set:

[0099]

[0100]

[0101]

[0102] x ω (k)-x ω′ (k)-l ω ≥τ0v ω′ (k)+ζ0if x ω (k)≥x ω′ (k),

[0103]

[0104]

[0105]

[0106]

[0107]

[0108] (f ω,ω′ (k)-1)·M≤x ω (k)-x ω′ (k)≤f ω,ω′ (k)·M

[0109]

[0110] where a ω (k) and v ω (k) represent the instantaneous acceleration and instantaneous velocity of vehicle ω at time step k, and are the minimum and maximum operating speeds of the vehicle, respectively. and are the maximum deceleration and maximum acceleration respectively. ω (k) is the position of vehicle ω at time step k, x ω (0), v ω (0) and a ω (0) are the position, velocity and acceleration of the vehicle ω at the initial moment. ω is the length of vehicle ω, τ0 is the safe vehicle headway, ζ0 is the safe vehicle spacing, f ω,ω′ (k) is a binary variable indicator. If the target vehicle ω is in front of the vehicle ω′, then f ω,ω′ (k)=1, otherwise, f ω,ω′ (k)=0.

[0111] Connected autonomous vehicles on each road section should comply with lane management constraints in lane use, including lane change feasibility constraints, lane reduction constraints, and boundary constraints for vehicles leaving the road section.

[0112] 1) The specific form of the lane-changing feasibility constraint is as follows:

[0113]

[0114]

[0115] Among them, j and j′ are different lane number indexes, J is the lane number index set, is a lane selection indicator, if vehicle ω is in lane j at time step k, then Otherwise it is 0.

[0116] 2) The specific form of lane reduction constraint is as follows

[0117]

[0118] Among them, M is a sufficiently large constant, L b is the location of the lane reduction area, J b It is a collection of lane number indices of lanes closed in the lane reduction area.

[0119] 3) The specific form of the boundary constraint for the vehicle leaving the road section is as follows:

[0120]

[0121]

[0122] Among them, M is a sufficiently large constant, L d is the starting location of the downstream segment unit boundary.

[0123] At the same time, since long road sections are divided into several shorter road section units, the connected autonomous driving vehicle must meet the mileage constraints in each road section unit. The specific form of the constraints is:

[0124]

[0125] Among them, D ω is the mileage of the vehicle, k max is the final time step in the optimization time domain, L d is the starting boundary position of the downstream section unit, L u is the end boundary position of the upstream section unit, Ω I Represents the vehicle number index set that will pass through the upstream, Ω s Represents the vehicle number index set that is about to leave this road section unit.

[0126] Since the model is solved by mixed integer linear programming, the above constraints need to be linearized. The specific form of the linearized constraints after conversion is:

[0127] D ω ≥x ω (k max )-L u ,ω∈Ω I

[0128] D ω ≥0,ω∈Ω I

[0129] D ω ≤(1-λ ω )·M+x ω (k max )-L u ,ω∈Ω I

[0130] D ω ≤λ ω ·M,ω∈Ω I

[0131] D ω ≤x ω (k max )-L u ,ω∈Ω s

[0132] D ω ≤L d -L u ,ω∈Ω s

[0133] D ω ≥(1-λ ω )·M+x ω (k max )-L u ,ω∈Ω s

[0134] D ω ≥λ ω M+L d -L u ,ω∈Ω s

[0135] Among them, M is a sufficiently large constant, λ ω is a binary variable indicator. If the vehicle ω travels mileage D ω Equal to the position x of vehicle ω at the end of the optimization time domain ω (k max ) minus the end boundary position L of the upstream section unit u , then λω =1, otherwise, λ ω =0.

[0136] The traffic throughput of the last road segment is limited by the traffic capacity of the upstream road segment unit, so the traffic throughput constraints are as follows:

[0137]

[0138] Among them, C is the traffic capacity of the upstream section unit, and the term on the left represents the traffic volume of vehicles crossing the boundary of the downstream section unit.

[0139] In addition, the multi-segment coordinated instruction generator mainly generates the state instructions of the connected autonomous driving vehicle that crosses the downstream boundary of the target segment unit. The instructions complete the control in the form of constraints:

[0140]

[0141]

[0142]

[0143]

[0144]

[0145] in, is the time step when vehicle ω crosses the downstream boundary of the road segment, The position of vehicle ω when it crosses the downstream boundary of the road segment, is the speed of vehicle ω crossing the downstream boundary of the road segment, is the lane selection when vehicle ω crosses the downstream boundary of the road segment, Ω II Represents the vehicle number index set that is about to cross the downstream boundary of the road segment unit.

[0146] In summary, this technical solution takes into account the ability of connected autonomous driving technology to accurately and in real time collect vehicle operational information on the road network. Based on this information, an active traffic management method can accurately predict traffic flow on a road segment. Unlike traditional methods that require traffic flow to reach steady state on a long road segment before predicting flow, this method divides long road segments into several smaller sections where flow prediction can be performed without strictly achieving steady state. This method leverages the microscopic information perception of connected autonomous vehicles to achieve accurate and real-time flow prediction. Furthermore, connected autonomous vehicles on each sub-segment can manage their behavior at the micro level through a multi-center control platform, enabling timely and precise execution of multi-center control commands, achieving trajectory-level active traffic control. This technical solution significantly improves traffic demand control capabilities in terms of control range, efficiency, and accuracy, enhancing vehicle cruising comfort and traffic flow stability. It also achieves significant improvements in computational efficiency, addressing the curse of dimensionality that traditional methods often fall prey to. Furthermore, it offers excellent compatibility across various scenarios and is applicable to new mixed traffic flow scenarios consisting of connected autonomous vehicles and human-driven vehicles.

Claims

1. A multi-center trajectory-level active traffic control system, characterized in that: It includes a data acquisition module, a decision generation module and a command execution module connected in sequence, wherein the data acquisition module is used to collect trajectory information of the networked autonomous driving vehicle in the road network as input to the decision generation module; The decision generation module determines, based on the input vehicle trajectory information and with the goal of multi-segment coordination, trajectory control instructions that can dynamically adjust the instantaneous traffic flow rate of each road segment in the road network; The command execution module is used to send trajectory control instructions to the networked autonomous driving vehicles on each road section, so that the networked autonomous driving vehicles execute the corresponding trajectory control commands; The decision generation module includes a terminal segment decision maker, multiple intermediate segment decision makers, and multiple multi-segment coordination instruction generators. The terminal segment decision maker generates a terminal segment traffic flow target adjustment flow rate decision plan based on the downstream bottleneck segment traffic capacity, which serves as a basis for the upstream intermediate segment multi-segment coordination instruction generator to generate coordination instructions. The multi-section coordinated instruction generator transmits decision information from one section unit to the upstream adjacent section unit according to the target of traffic flow rate control of the end section, forming a multi-section coordinated instruction, which is the decision basis of each intermediate section decision maker; The intermediate section decision maker is used to control the traffic flow of the intermediate section unit. Based on the traffic flow rate of the last section, multi-section coordination instructions and vehicle information on the relevant sections, it forms a decision that can optimize the traffic flow rate of the intermediate section, thereby reducing the fluctuations generated during traffic operation. The optimization objectives of the end-of-road segment decision maker include: discomfort during vehicle cruising, oscillations and fluctuations in traffic flow status, and tracking errors in traffic flow rate. The optimization objectives of the end-of-road segment decision maker are specifically: Among them, k and ω correspond to the time step index and target vehicle index respectively, ω ′ is the non-target vehicle index, K and Ω correspond to the time step index set and vehicle index set respectively, a ω (k) and v ω (k) are the instantaneous acceleration and instantaneous speed of the target vehicle ω at time step k, q is the predicted traffic flow rate at the end of the optimization time domain, and q t is the target expected flow rate, α, β, and γ correspond to the target weights of discomfort during vehicle cruising, the target weights of oscillation fluctuations in the state during traffic flow operation, and the target weights of traffic flow rate tracking error, respectively; The optimization objectives of the intermediate section decision maker include: discomfort during vehicle cruising and oscillations in traffic flow. The optimization objectives of the intermediate section decision maker are specifically: Among them, a ω (k) and v ω (k) are the instantaneous acceleration and instantaneous velocity of the target vehicle ω at time step k, ω ′ is the non-target vehicle index, K and Ω correspond to the time step index set and vehicle index set respectively, α and β are the target weights of discomfort during vehicle cruising and the target weights of oscillation fluctuations during traffic flow operation respectively; The constraints of the intermediate section decision maker include operational constraints of the connected autonomous driving vehicle and lane management constraints; The constraints of the end-segment decision maker include operational constraints of connected autonomous vehicles, lane management constraints, mileage constraints, and upstream segment capacity constraints; The multi-segment coordinated instruction generator is used to constrain the state of the connected autonomous driving vehicle that crosses the downstream boundary of the target segment unit. The specific form of the constraint is: in, is the time step when the target vehicle ω crosses the downstream boundary of the road segment, is the position of the target vehicle ω when it crosses the downstream boundary of the road section, L u is the end boundary position of the upstream section unit, is the speed of the target vehicle ω crossing the downstream boundary of the road segment. If the target vehicle ω is in lane j at time step k, then Otherwise it is 0. is the lane selection when the target vehicle ω crosses the downstream boundary of the road segment, Ω II Represents a collection of vehicle indexes that are about to cross the downstream boundary of the road segment unit.

2. A multi-center trajectory-level active traffic control system according to claim 1, characterized in that: The kinematic constraints of the connected autonomous driving vehicle are specifically: Among them, a ω (k) and v ω (k) represent the instantaneous acceleration and instantaneous velocity of the target vehicle ω at time step k, and are the minimum and maximum operating speeds of the vehicle, respectively. and are the maximum deceleration and maximum acceleration, respectively, x ω (k) is the position of the target vehicle ω at time step k, x ω (0), v ω (0) and a ω (0) are the position, velocity and acceleration of the target vehicle ω at the initial moment, l ω is the length of the target vehicle ω, τ0 is the safe vehicle headway, ζ0 is the safe vehicle spacing, M is a constant, f ω,ω′ (k) is a binary variable indicator if the target vehicle ω is within the position of the non-target vehicle ω ′ In front, then f ω,ω′ (k)=1, otherwise, f ω,ω′ (k)=0.

3. A multi-center trajectory-level active traffic control system according to claim 2, characterized in that: The lane management constraints for connected autonomous vehicles include lane change feasibility constraints, lane reduction constraints, and boundary constraints for vehicle exit sections. Specifically, they are: Where M is a constant, j and j ′ is the index of different lane numbers, J is the lane number index set, is a lane selection indicator, if the target vehicle ω is in lane j at time step k, then Otherwise it is 0, L b is the location of the lane reduction area, J b is the set of lane indexes of lanes closed in the lane reduction area, L d It is the starting boundary position of the downstream segment unit.

4. A multi-center trajectory-level active traffic control system according to claim 3, characterized in that: The mileage constraints are specifically: D ω ≥x ω (k max )-L u ,ω∈Ω I # D ω ≥0,ω∈Ω I # D ω ≤(1-λ ω )·M+x ω (k max )-L u ,ω∈Ω I # D ω ≤λ ω ·M,ω∈Ω I # D ω ≤x ω (k max )-L u ,ω∈Ω s # D ω ≤L d -L u ,ω∈Ω s # D ω ≥(1-λ ω )·M+x ω (k max )-L u ,ω∈Ω s # D ω ≥λ ω ·M+L d -L u ,ω∈Ω s # Among them, D ω is the mileage of the target vehicle ω, k max is the final time step in the optimization time domain, L u is the end boundary position of the upstream section unit, Ω I Represents the vehicle number index set that will pass through the upstream, Ω s Represents the vehicle number index set that is about to leave this road section unit, λ ω is a binary variable indicator. If the target vehicle ω travels mileage D ω Equal to the position x of vehicle ω at the end of the optimization time domain ω (k max ) minus the end boundary position L of the upstream section unit u , then λ ω =1, otherwise, λ ω =0, M is a constant; The upstream section capacity constraint is specifically: Among them, C is the traffic capacity of the upstream section unit, and the term on the left represents the traffic volume of vehicles crossing the boundary of the downstream section unit.

5. A multi-center trajectory-level active traffic control method, applied to a multi-center trajectory-level active traffic control system according to claim 1, characterized in that: The following steps are involved: S1, the data acquisition module collects the trajectory information of the connected autonomous driving vehicles in the road network in real time; S2, the end segment decision maker determines the end segment vehicle trajectory plan based on the trajectory information and the preset bottleneck traffic information; S3, a multi-section coordinated instruction generator transmits decision information from one section unit to the upstream adjacent section units according to the target of traffic flow rate control of the last section, forming a multi-section coordinated instruction; S4: The intermediate segment decision maker determines the vehicle trajectory plan for the current intermediate segment under the constraints of the vehicle information and coordination instructions on the segment. When the plan is transmitted to the upstream segment and the corresponding decision is made, a trajectory control instruction with multi-segment coordination is formed. S5. The command execution module sends the trajectory control instructions to the connected autonomous driving vehicles on each road section, so that the connected autonomous driving vehicles execute the corresponding trajectory control commands.

Images