Queueing dynamic considering road network congestion sub-region double-layer boundary control method and system

Abstract

The application provides a road network congestion sub-area double-layer boundary control method and system considering queuing dynamics, and relates to the technical field of intelligent transportation, which comprises the following steps: acquiring basic data of a target road network; establishing a macroscopic traffic flow model according to the basic data, deriving a vehicle conservation equation and performing linearization processing to obtain a linearized macroscopic traffic flow model; constructing a network layer boundary control framework according to the linearized macroscopic traffic flow model, generating a global boundary control rate; establishing a road intersection layer control mechanism according to the global boundary control rate, generating a differentiated road intersection control rate; obtaining a double-layer boundary control model according to the global boundary control rate and the differentiated road intersection control rate; inputting real-time acquired vehicle density and traffic flow data into the double-layer boundary control model to perform signal timing optimization and output an optimized traffic flow state. The application effectively improves the stability of the road network traffic flow, improves the queuing distribution balance and enhances the overall operation efficiency.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and more specifically, to a two-layer boundary control method and system for road network congestion sub-regions that takes into account queuing dynamics. Background Technology

[0002] With the acceleration of urbanization and the rapid increase in motor vehicle ownership, urban traffic congestion has become increasingly severe. Traditional local control methods are insufficient to cope with the complex traffic dynamics in large-scale road networks. To address this, the academic community has introduced macroscopic fundamental graph theory, which provides a new approach to regional traffic management by describing the overall traffic flow characteristics of the road network. This theory establishes a macroscopic relationship between vehicle density and flow rate, enabling a holistic understanding of the road network's operational status and deriving boundary control strategies based on sub-zone division to alleviate congestion by adjusting flow rates between zones. However, existing methods often rely on fixed timing of traffic lights or a single-level control framework, making it difficult to adapt to the real-time evolution of traffic conditions. Furthermore, insufficient coordination between the network layer and intersection layer leads to uneven distribution of queues at boundaries and local oversaturation, limiting further improvements in control effectiveness.

[0003] Based on the shortcomings of the existing technology, there is an urgent need for a two-layer boundary control method and system for road network congestion sub-areas that takes into account queuing dynamics. Summary of the Invention

[0004] The purpose of this invention is to provide a two-layer boundary control method and system for road network congestion sub-regions that considers queuing dynamics, in order to improve the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows:

[0005] Firstly, this application provides a two-layer boundary control method for road network congestion sub-regions that considers queuing dynamics, including:

[0006] Acquire basic data of the target road network, including road network topology information and historical traffic data;

[0007] A macroscopic traffic flow model is established based on the aforementioned basic data. By deriving the vehicle conservation equation and performing linearization, a linearized macroscopic traffic flow model is obtained.

[0008] A network layer boundary control framework is constructed based on the linearized macroscopic traffic flow model. The global release intensity is calculated using the proportional-integral control principle, and the global boundary control rate is generated.

[0009] An intersection layer control mechanism is established based on the global boundary control rate. By integrating the queue change trend, storage occupancy level and demand intensity three-dimensional coupling operator, a differentiated intersection control rate is generated.

[0010] A two-layer boundary control model is established based on the global boundary control rate and the differentiated intersection control rate. The two-layer boundary control model is obtained by coupling the control logic of the network layer and the intersection layer.

[0011] The real-time acquired vehicle density and traffic flow data are input into the dual-layer boundary control model to perform signal timing optimization, and the optimized traffic flow state is output by dynamically adjusting the green light duration.

[0012] Secondly, this application also provides a two-layer boundary control system for road network congestion sub-regions that considers queuing dynamics, including:

[0013] The acquisition module is used to acquire basic data of the target road network, including road network topology information and historical traffic data.

[0014] The processing module is used to establish a macroscopic traffic flow model based on the basic data, and to obtain a linearized macroscopic traffic flow model by deriving the vehicle conservation equation and performing linearization processing.

[0015] The construction module is used to construct a network layer boundary control framework based on the linearized macro traffic flow model, calculate the global release intensity using the proportional-integral control principle, and generate the global boundary control rate.

[0016] The coupling module is used to establish an intersection layer control mechanism based on the global boundary control rate. It generates a differentiated intersection control rate by integrating the three-dimensional coupling operators of queue change trend, storage occupancy degree and demand intensity.

[0017] An integration module is used to establish a two-layer boundary control model based on the global boundary control rate and the differentiated intersection control rate. The two-layer boundary control model is obtained by coupling the control logic of the network layer and the intersection layer.

[0018] The output module is used to input real-time acquired vehicle density and traffic flow data into the dual-layer boundary control model to perform signal timing optimization, and output the optimized traffic flow state by dynamically adjusting the green light duration.

[0019] The beneficial effects of this invention are as follows:

[0020] This invention acquires basic road network data and establishes a linearized macroscopic traffic flow model. It generates a global boundary control rate using the proportional-integral control principle, combines a three-dimensional coupling operator to achieve differentiated control at intersections, and finally dynamically optimizes signal timing through a two-layer boundary control model. This effectively improves the stability of road network traffic flow, enhances the balance of queue distribution, and improves overall operational efficiency. Attached Figure Description

[0021] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0022] Figure 1 This is a flowchart illustrating a two-layer boundary control method for road network congestion sub-regions that considers queuing dynamics, as described in an embodiment of the present invention.

[0023] Figure 2 This is a schematic diagram of a two-layer boundary control system for road network congestion sub-areas that considers queuing dynamics, as described in an embodiment of the present invention.

[0024] Figure 3 A schematic diagram of macroscopic traffic flow modeling;

[0025] Figure 4 This is a schematic diagram of PI control;

[0026] Figure 5 Flowchart for intersection-level signal timing optimization;

[0027] Figure 6 This is a schematic diagram of the controlled sub-region;

[0028] Figure 7 A schematic diagram of the fitting results for the macroscopic basic map of the controlled sub-region;

[0029] Figure 8 A schematic diagram of the fitting results for the macroscopic basic map of the outer sub-region;

[0030] Figure 9 Input traffic demand diagrams for the controlled area;

[0031] Figure 10 This is a schematic diagram showing the location of the boundary intersection.

[0032] The diagram is labeled as follows: 901, Acquisition Module; 902, Processing Module; 903, Construction Module; 904, Coupling Module; 905, Integration Module; 906, Output Module. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0034] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0035] Example 1:

[0036] This embodiment provides a two-layer boundary control method for road network congestion sub-regions that takes into account queuing dynamics.

[0037] See Figure 1 The figure shows that the method includes steps S100 to S600.

[0038] Step S100: Obtain basic data of the target road network, including road network topology information and historical traffic data;

[0039] Understandably, step S100 acquires basic data of the target road network through actual traffic detection infrastructure. This data includes road network topology information such as road grade, number of lanes, road segment length, and intersection connections (vector data). Historical traffic data originates from traffic flow and occupancy data recorded by fixed loop detectors at preset sampling intervals for each road segment over various time periods. The occupancy data is used to calculate the vehicle density of the road segment. This historical traffic data (including traffic flow, occupancy, and calculated vehicle density) collectively forms the original input basis for subsequent model building.

[0040] Step S200: Establish a macroscopic traffic flow model based on the basic data. By deriving the vehicle conservation equation and performing linearization, a linearized macroscopic traffic flow model is obtained.

[0041] It should be noted that step S200 establishes a macroscopic traffic flow model based on the acquired basic data. By deriving the conservation relationship between vehicle inflow and outflow and selecting the optimal operating point for linearization, the nonlinear traffic flow is dynamically transformed into an analytically processable linear system, providing a theoretical basis for control design.

[0042] Step S300: Construct a network layer boundary control framework based on the linearized macro traffic flow model, calculate the global release intensity using the proportional-integral control principle, and generate the global boundary control rate;

[0043] Understandably, step S300 uses a linearized model to construct a network layer boundary control framework and adopts the proportional-integral control principle to dynamically adjust the boundary release intensity, so that the accumulated number of vehicles in the controlled sub-area is stabilized near the optimal operating point, thereby achieving congestion control at the macro level.

[0044] Step S400: Establish an intersection layer control mechanism based on the global boundary control rate. Generate a differentiated intersection control rate by integrating the three-dimensional coupling operator of queue change trend, storage occupancy degree and demand intensity.

[0045] It should be noted that step S400 establishes an intersection-level control mechanism based on the global control rate. By integrating coupled operators of three dimensions—queue change trend, storage occupancy level, and demand intensity—it generates a differentiated control rate that reflects the real-time load differences of each intersection, thereby improving the precision of boundary control.

[0046] Step S500: Establish a two-layer boundary control model based on the global boundary control rate and the differentiated intersection control rate. By coupling the control logic of the network layer and the intersection layer, the two-layer boundary control model is obtained.

[0047] It is understandable that step S500 couples the control logic of the network layer and the intersection layer, and establishes a two-layer boundary control model by coordinating the global release intensity and the intersection-level allocation relationship to form a unified control architecture.

[0048] Step S600: Input the real-time acquired vehicle density and traffic flow data into the dual-layer boundary control model to perform signal timing optimization, and output the optimized traffic flow state by dynamically adjusting the green light duration.

[0049] It should be noted that step S600 inputs the real-time detected vehicle density and traffic flow data into the two-layer control model, dynamically optimizes the signal timing parameters and implements green light duration adjustment, and finally outputs the optimized traffic flow status.

[0050] Further, step S200 includes steps S210 to S230.

[0051] Step S210: Based on the basic data, establish the vehicle conservation equation and construct the vehicle inflow and outflow balance relationship in the sub-region based on the macroscopic basic graph theory to obtain the initial vehicle conservation equation.

[0052] Step S220: Linearize the initial vehicle conservation equation by selecting the optimal vehicle accumulation in the sub-region as the set point and performing Taylor series expansion to obtain the linearized vehicle conservation equation.

[0053] Step S230: Extract parameters based on the linearized vehicle conservation equation, and obtain the linearized macroscopic traffic flow model by deriving the coefficient matrix in the linearized equation.

[0054] Specifically, such as Figure 3 As shown, in macro-level traffic flow modeling, based on the macro-level fundamental graph theory, it is assumed that after sub-region division, the urban road network can be divided into two sub-regions with relatively small macro-level fundamental graph dispersion: the controlled sub-region and the peripheral sub-region. (See figure.) and These represent two sub-regions within the road network. It is a controlled area (city center). It is the outer region. The transfer flow between sub-regions is controlled by the boundary control rate. and Adjustment, control from Flow direction Traffic and Flow direction The traffic, of which and . and These represent sub-regions respectively. Flow to sub-region The transferred traffic, and from the sub-region Flow to sub-region The transferred flow; , Representing sub-regions Hezi District The total number of all vehicles in the area is the cumulative number of vehicles. This indicates internally generated demand (e.g., vehicles entering the road network from roadside parking areas). This represents the uncontrolled outflow from the controlled sub-region to the outer sub-region.

[0055] For the controlled sub-region The cumulative number of vehicles in its sub-zone clearly satisfies the following:

[0056] (1-1);

[0057] in, express Time zone The number of vehicles inside, express From time to time Flow direction The number of vehicles. The sum of internal demand and uncontrolled outflows. To indicate:

[0058] (1-2);

[0059] in, express The sum of internal demand and uncontrolled outflow at any given moment. express Time zone Internally generated traffic demand express The time interval represents the uncontrolled outflow from the controlled sub-region to the outer sub-region.

[0060] Assuming controlled sub-region There exists a clear macro-level chart that shows the cumulative number of vehicles. and total production Connect them. Total completed traffic in the sub-region. (That is, the number of vehicles leaving the area per unit of time, including vehicles that have completed their journey or are heading to adjacent sub-areas) can be used It means that among them For sub-region The average travel length is assumed to be independent of time and destination. Then, based on the data from the coil detector, the total completed flow rate can be estimated using formula (1-3):

[0061] (1-3);

[0062] in, Sub-region Total completed traffic, express Time zone Internal traffic completion express From time to time Flow direction The transferred traffic.

[0063] at the same time, The vehicle conservation equation for a vehicle whose destination is in the controlled sub-region at any given time can be expressed as:

[0064] (1-4);

[0065] in, express Time zone The cumulative number of vehicles inside. The differential symbol, For a moment, express Time zone To the sub-district Traffic flow control rate express From time to time Flow direction The transfer flow, express Time zone Internal traffic completion express Time zone Internally generated traffic demand.

[0066] This formula consists of three parts, namely Transfer flow under constant boundary control ; Time zone Internal completion flow ;as well as Traffic is generated within the time sub-region based on internal needs. .

[0067] Similarly, for the case where the vehicle's destination is in an outer sub-region, the conservation equation can be expressed as:

[0068] (1-5);

[0069] in, express From time to time Flow direction The number of vehicles, express Time zone Internally generated traffic demand express From time to time Flow direction The transferred traffic.

[0070] Based on formulas (1-1) to (1-5), the sub-regions can be further derived. The overall conservation equation is:

[0071] (1-6);

[0072] in, express Time zone The cumulative number of vehicles, express Time zone The number of vehicles inside, express From time to time Flow direction The number of vehicles, express Time zone To the sub-district Traffic flow control rate express From time to time Flow direction The transfer flow, This represents the total completed flow of the sub-region. express The sum of internal demand and uncontrolled outflow at any given moment.

[0073] In practical applications, traffic light control typically adjusts signal timing using discrete time steps, and the traffic flow data from loop detectors is discrete data collected at specific time intervals. By converting a given continuous-time equation into a discrete form, then... time The cumulative number of vehicles in a sub-region can be expressed by the following formula (1-7):

[0074] (1-7);

[0075] in, Indicates in time The cumulative number of vehicles in the sub-area. Indicates in time The cumulative number of vehicles in the sub-area. express From time to time Flow direction The transfer flow, Indicates in Time zone Total completed traffic, Indicates in time The sum of internal demand and uncontrolled outflow in the sub-region Indicates the discrete time step. Represents discrete time points.

[0076] The conservation equations (1-6) are nonlinear equations. To facilitate subsequent controller design and better apply them to the control theory framework, these equations are linearized. Furthermore, introducing a desired steady state simplifies the controller design process and effectively improves system stability and performance, enabling the system to operate near the steady state. In the context of the macroscopic fundamental traffic flow diagram (MFD), this nonlinear model is linearized around the following setpoint. ,in Choice and Related, It is by The optimal vehicle accumulation amount can be obtained by fitting a cubic polynomial function and then using this function. For the given , and Solution of the steady-state equation To stabilize the system near the equilibrium point, the steady-state equation is:

[0077] (1-8);

[0078] in, This represents the sum of internal demand and uncontrolled outflows under steady-state conditions. Sub-region To sub-region The value of the boundary control law under steady-state conditions. Indicates in sub-region The cumulative number of vehicles is From Flow direction The transfer flow, Sub-region The cumulative number of vehicles is Tokiko District Total completed traffic.

[0079] Linear expansion of formula (1-6) around the selected set point:

[0080] (1-9);

[0081] (1-10);

[0082] (1-11);

[0083] get:

[0084] (1-12);

[0085] in, This indicates the deviation of the cumulative vehicle volume from the steady-state value. This represents the actual cumulative number of vehicles. This represents the optimal vehicle accumulation under steady-state conditions. This represents the deviation of the boundary control rate from the steady-state value. Indicates the actual boundary control rate. This represents the boundary control law under steady state. This represents the deviation of the sum of internal demand and uncontrolled outflows from the steady-state value. This represents the sum of actual internal demand and uncontrolled outflows. This represents the sum of the corresponding steady-state internal demand and uncontrolled outflows.

[0086] Transform a given continuous-time equation into its discrete form:

[0087] (1-13);

[0088] in, Indicates at discrete time sub-region The deviation of the cumulative vehicle volume from the steady-state value. Indicates at discrete time sub-region The deviation of the cumulative vehicle volume from the steady-state value. Indicates at discrete time sub-region Xiangzi District The deviation of the boundary control rate from the steady-state value, Indicates at discrete time sub-region The cumulative number of vehicles is From Flow direction The transfer flow, Sub-region The macroscopic fundamental graph function at the optimal cumulative The first derivative at that point, Indicates at discrete time sub-region The deviation of the sum of internal demand and uncontrolled outflows from the steady-state value. Sub-region To sub-region The value of the boundary control law under steady-state conditions. Indicates sub-region The macroscopic fundamental graph function at the optimal cumulative The first derivative at that point, Indicates at discrete time sub-region The deviation of the cumulative vehicle volume from the steady-state value.

[0089] The initial vehicle conservation equation (1-6) established in step S210, used as a description of a nonlinear system, is transformed into a linearized vehicle conservation equation (1-12) in step S220 by selecting the optimal vehicle accumulation amount in a sub-region as the setpoint and performing a Taylor series expansion. This transformation is achieved through the step-by-step derivation of formulas (1-8) to (1-11), converting the nonlinear dynamic system into a system that can be linearly approximated near the setpoint. In step S230, the continuous-time equation is further converted into a discrete form using formula (1-13), and a model is constructed based on the linearized vehicle conservation equation. By deriving its discrete state-space expression, a linearized macroscopic traffic flow model is obtained. This series of processes makes the complex traffic flow dynamics applicable to modern control theory analysis methods, laying a theoretical foundation for the subsequent design of boundary controllers.

[0090] Further, step S300 includes steps S310 to S330.

[0091] Step S310: Determine the control setpoint based on the linearized macroscopic traffic flow model. By selecting the optimal vehicle accumulation in the macroscopic basic map as the target operating point, the desired accumulation setpoint is obtained.

[0092] Step S320: Perform proportional-integral control processing based on the desired cumulative amount set value. Calculate the error between the current vehicle cumulative amount and the set value and apply the proportional-integral control law for dynamic adjustment to obtain the preliminary boundary control law.

[0093] Step S330: Perform control rate constraint processing based on the preliminary boundary control rate, and ensure the feasibility of the control rate by applying minimum and maximum release strength limits to obtain the global boundary control rate.

[0094] Specifically, in the construction of network layer boundary control frameworks, PI controllers (Proportional-Integral controllers) are widely used due to their simplicity, robustness, and ease of implementation. By simultaneously adjusting the proportional and integral components of the system, PI controllers can effectively handle steady-state errors in dynamic systems while maintaining the stability of the control system over a wide range of system parameters. In sub-area boundary control of traffic systems, the boundary control rate is dynamically adjusted... This makes the controlled sub-region To maintain within the expected range, the PI controller can be selected for: (1) Real-time dynamic adjustment: Boundary flow needs to be adjusted according to real-time traffic conditions. The PI controller can calculate the flow error in real time through the feedback loop and quickly adjust the output to achieve dynamic response. (2) Steady-state performance requirements: In sub-area boundary control, steady-state error may cause the accumulation of vehicles in the sub-area to deviate from the expected value, affecting the efficiency of the road network. The PI controller can eliminate the error through integral action.

[0095] The core of a PI controller is feedback control, and its control law can be expressed as:

[0096] (1-14);

[0097] In the formula, Indicates control input; This represents the systematic error, which is the difference between the target value and the current value. This is the proportional coefficient, used to adjust the controller's immediate response to errors; The integral coefficient is used to eliminate the steady-state error of the system and improve the long-term performance of the system.

[0098] This embodiment establishes as follows: Figure 4 The boundary control system shown in the figure, in which the boundary control system is based on the controlled sub-region vehicle cumulative Adjusting the input of the control system Define error Based on the control law of PI control, the boundary control law can be expressed as:

[0099] (1-15);

[0100] in, express time to Boundary flow control rate, express time Flow direction Boundary flow control rate, This is the proportionality coefficient. The integral coefficient is... express Controlled sub-regions at all times Cumulative number of vehicles express Controlled sub-regions at all times Cumulative number of vehicles Sub-region The optimal vehicle accumulation under steady-state conditions express Time zone The error between the optimal vehicle accumulation and the actual vehicle accumulation.

[0101] Furthermore, considering the signal timing at boundary intersections in practical applications, the control rate setting needs to balance several factors. First, the control rate should not be too low, ensuring that the shortest green light time at the intersection meets the needs of pedestrians crossing the street safely; simultaneously, the control rate should not be too high, as excessively long green light times may prevent the signal timing from meeting the time requirements of pedestrians crossing on transverse roads. Therefore, the control rate needs to be carefully considered. Apply constraints:

[0102] (1-16);

[0103] in, and Control rate Minimum and maximum constraints to be satisfied. express time Flow direction The control rate. If obtained If the constraints are not met, adjustments should be made:

[0104] (1-17);

[0105] Meanwhile, vehicle accumulation It should also satisfy the maximum constraint. ,Right now Because when The time zone will be locked, causing traffic flow to be unable to move effectively, thus affecting the operational efficiency of the entire road network.

[0106] The aforementioned PI controller design theory provides a complete control framework for the implementation of steps S310 to S330. In step S310, by selecting the optimal vehicle accumulation amount in the macroscopic basic graph as the target operating point, the desired accumulation amount setpoint is determined, laying the foundation for control error calculation. Based on this setpoint, step S320 implements proportional-integral control processing through formulas (1-14) and (1-15), calculates the error between the current vehicle accumulation amount and the setpoint, and applies dynamic adjustment of the control law to obtain the preliminary boundary control law. Step S330 further constrains the preliminary control law through formulas (1-16) and (1-17), applying minimum and maximum release intensity limits to ensure the engineering feasibility of the control law, ultimately generating the global boundary control law. This complete control process combines theoretical control laws with actual engineering constraints, realizing the entire process from setting the control target to generating an executable control law.

[0107] Further, step S400 includes steps S410 to S430.

[0108] Step S410: Configure the basic parameters of the intersection based on the global boundary control rate. By determining the minimum effective green time and saturation flow parameters of each boundary intersection, the basic control parameters of the intersection are obtained.

[0109] Step S420: Based on the basic control parameters of the intersection, a three-dimensional coupling operator is constructed. The comprehensive weight is calculated by integrating the ratio of adjacent cycle queue lengths, the road segment storage capacity occupancy rate, and the saturated flow demand intensity to obtain the intersection coupling weight coefficient.

[0110] Step S430: Generate differentiated control rates based on intersection coupling weight coefficients. The differentiated intersection control rates are obtained by adjusting normalized weight allocation and boundary truncation constraints.

[0111] It should be noted that in traditional boundary control, a uniform control rate is typically applied to all boundary intersections. However, in actual urban road networks, different boundary segments vary in terms of connector length, number of lanes, and channelization patterns, resulting in inconsistent queue growth rates and available storage capacity. When the overall release intensity decreases, although single-layer boundary control can effectively suppress the overall inflow of external vehicles and significantly alleviate boundary overflow, the load between different intersections may still be uneven. Queues at some intersections may decrease only slightly, while others may still approach saturation.

[0112] Therefore, it is necessary to further refine the green light allocation at the intersection layer based on the network layer control rate adjustment. This embodiment proposes a three-dimensional coupled operator of "trend-occupancy-demand," which constructs a comprehensive weight by integrating queue change trends, storage occupancy levels, and demand intensity to achieve more targeted allocation of traffic capacity. This operator is not intended to replace the mitigation effect of single-layer control on overall overflow, but rather to improve the queue distribution among different intersections at the boundary, making the load at each intersection more balanced, thereby improving the refinement and stability of boundary control.

[0113] The three-dimensional coupling operator is designed as follows: Firstly, it measures the queue growth trend by the ratio of the queue lengths of two adjacent control cycles. The indicators directly reflect whether congestion is accumulating at an accelerated pace, among which... For the border intersection exist The queue length at any given time, if If the intersection is widening, it is widening; otherwise, it is contracting. The second factor is saturation, which is the ratio of queue size to the road segment's storage capacity. describe, The first factor is the length of the road segment; a high saturation level means the intersection is approaching its capacity and needs more green light time. The second factor is the relative demand intensity. Measurement, among which For the border intersection exist Flow rate at any moment Intersection The saturation flow of the import channel. These three factors correspond to "trend, critical point, and demand," and together determine which intersection is more likely to become the trigger point for overflow in the current cycle, or where the effective green energy released will generate higher output.

[0114] Based on the above three-dimensional information, this embodiment constructs a mouth-level coupling operator, which takes the following form:

[0115] (1-18);

[0116] in These are weighting coefficients, used to adjust the relative sensitivity of the three types of factors. For three-dimensional coupling operators, Indicates the boundary intersection exist Queue length at any given time Indicates the boundary intersection exist Queue length at any given time Time was lost at the intersection. For the border intersection exist Flow rate at any moment Intersection The saturation flow of the approach lanes. To allocate more resources to more congested intersections, green light time is redistributed. The normalized results are used to construct weights, and the global effective green budget given by the network layer is allocated to the intersections that need it more according to the actual situation.

[0117] The construction process of the aforementioned three-dimensional coupling operator corresponds to the complete implementation path of steps S410 to S430: In step S410, the minimum effective green time and saturation flow parameters of each boundary intersection are determined through the global boundary control rate, establishing basic control parameters for the intersection and providing fundamental support for weight calculation; Step S420, based on the basic control parameters of the intersection, integrates three-dimensional information such as the ratio of adjacent cycle queue lengths, road segment storage capacity occupancy rate, and saturation flow demand intensity to calculate the comprehensive weight and obtain the intersection coupling weight coefficient; Step S430, through normalized weight allocation and boundary truncation constraint adjustment, transforms the intersection coupling weight coefficient into a differentiated intersection control rate, ultimately achieving refined allocation of intersection-level flow. This series of processes ensures the coordination and unity between the network layer control objective and the real-time state of the intersection layer.

[0118] Further, step S500 includes steps S510 to S530.

[0119] Step S510: Perform total control budget processing based on the global boundary control rate and the differentiated intersection control rate. By calculating the sum of the effective green time corresponding to the global release intensity of the network layer and each control rate of the intersection layer, a preliminary control budget framework is obtained.

[0120] Step S520: Perform hierarchical coupling coordination processing based on the preliminary control budget framework, and dynamically adjust the control rate allocation between the network layer and the intersection layer by applying the water level correction mechanism to ensure consistency of the total control amount, thereby obtaining the coupled control strategy;

[0121] Step S530: Based on the coupling control strategy, the model is structured and processed. By integrating the proportional-integral control logic of the network layer with the three-dimensional coupling operator of the intersection layer to form a unified control architecture, a two-layer boundary control model is obtained.

[0122] Understandably, to avoid the queuing overflow under high demand conditions from spreading upstream at the boundary intersection, this embodiment organizes the boundary control as a two-layer closed loop of "network layer - intersection layer": the network layer provides the current global release intensity (control rate) through the PI controller based on the sub-area MFD, and the intersection layer refines this total amount into executable timing schemes for each boundary intersection within a fixed control cycle, and performs timing adjustments through a three-dimensional coupling operator to establish two-layer PI control (TL-PI, two-layer PI control).

[0123] In each control cycle Internally, the two-layer control is executed according to the following steps.

[0124] First, at the network layer, based on the calculations of the PI controller... Control rate of movement from the outer sub-region to the controlled sub-region at any time Under controlled conditions Flow direction The globally effective green at the boundary intersection is:

[0125] (1-19);

[0126] (1-20);

[0127] in, Indicates under controlled conditions Flow direction Globally effective green at boundary intersections express time Flow direction The input of the control system For the index of the intersection, The total number of intersections. For the intersection cycle, For effective green duration, Time was lost at the intersection.

[0128] Secondly, at the intersection level, we first present the general traffic flow distribution principle: road sections with higher saturation traffic can accommodate more traffic. Assume an intersection... The saturation flow rate of the inlet is Therefore, the base green time for this intersection is:

[0129] (1-21);

[0130] in, Intersection The saturation flow rate of the inlet channel, Indicates an intersection The baseline green time, Indicates under controlled conditions Flow direction The boundary intersection is a globally effective green space.

[0131] Considering queue variations, storage requirements, and demand intensity, to ensure basic traffic capacity and safety constraints, a minimum effective green space is reserved at the intersection level for each boundary intersection. :

[0132] (1-22);

[0133] in, This indicates that the intersection level reserves the minimum effective green space for each boundary intersection. For effective green duration, The minimum effective green constant for intersections is given by the project. The sum of the minimum green constants for all intersections is obtained as follows:

[0134] (1-23);

[0135] in, Find the minimum green summation value for all intersections. This indicates that the intersection layer reserves the minimum effective green space for each boundary intersection.

[0136] like This indicates that the current overall traffic flow intensity is insufficient to meet the minimum green light requirements of all intersections. In this case, the minimum green light requirement for each intersection will be scaled down proportionally to:

[0137] (1-24);

[0138] in, This indicates that the intersection level reserves the minimum effective green space for each boundary intersection. Indicates under controlled conditions Flow direction Globally effective green at boundary intersections The minimum effective green constant for the intersection is given by the project.

[0139] Set the remaining freely distributable effective green to zero; otherwise, calculate the remaining distributable effective green:

[0140] (1-25);

[0141] in, For the remaining available effective green, Indicates under controlled conditions Flow direction Globally effective green at boundary intersections Find the minimum green summation value for all intersections.

[0142] When allocating the remaining green light time, priority should be given to intersections with high expansion and saturation levels. To achieve reasonable traffic distribution, a three-dimensional coupling operator should be used. Normalization yields the weights:

[0143] (1-26);

[0144] in, To normalize the weights, For three-dimensional coupling operators, Intersection Three-dimensional coupling operator, For the index of the intersection, This represents the total number of intersections.

[0145] These weights reflect the overall urgency of each intersection in the current cycle across three dimensions: rapid queue growth, high storage usage, and strong demand. Subsequently, the remaining allocable effective green space is distributed among the intersections according to these weights, resulting in the initial intersection-level effective green space:

[0146] (1-27);

[0147] in, For the initial effective green at the intersection level, This indicates that the intersection level reserves the minimum effective green space for each boundary intersection. To normalize the weights, The remaining available green space.

[0148] Then, the initial effective green is converted into the corresponding initial intersection-level control rate:

[0149] (1-28);

[0150] in, The initial intersection-level control rate, For the initial effective green at the intersection level, For effective green time.

[0151] In addition, upper and lower bound constraints on the proportion are applied to each intersection to avoid excessively high or low traffic intensity at a single intersection:

[0152] (1-29);

[0153] in, and These are the lower and upper limits of the effective green ratio set for the project, respectively. for Always at the intersection The applied control rate, This represents the initial intersection-level control rate. At this point, the control rates for each intersection are... The differential tilt brought about by the three-dimensional coupling operator has been demonstrated, but due to the truncation effect of the upper and lower bounds, their total effective green may not be consistent with the target given by the network layer.

[0154] To ensure consistency in total control between the two layers, a simple water level adjustment is needed based on the truncated proportion. The actual allocated total effective green space is calculated based on the current truncated proportion:

[0155] (1-30);

[0156] in, This represents the total effective green space actually allocated. for Always at the intersection The applied control rate, For effective green time.

[0157] And obtain the difference between the target budget and the actual budget:

[0158] (1-31);

[0159] in, This represents the difference between the total effective green area globally and the actual allocated total effective green area. Indicates under controlled conditions Flow direction Globally effective green at boundary intersections This indicates the remaining green volume.

[0160] when When there is a significant deviation from zero, the set only occurs at intersections that do not touch the upper or lower bounds. This difference will be shared internally. Specifically, this involves... Based on the effective greening capacity of each intersection Weighted allocation shifts the proportion of active intersections up or down by the same level simultaneously, thus restoring overall consistency as much as possible in a single update. Mathematically, the adjustment of the proportion of active intersections can be expressed as:

[0161] (1-32);

[0162] in, for Always at the intersection The applied control rate, This is a uniform control law correction introduced during a control update to eliminate the global effective green deficit. This represents the difference between the total effective green area globally and the actual allocated total effective green area. and For the index of the intersection, Meet at the intersection. For effective green time.

[0163] And project back onto the interval after the update. If all intersections have reached the boundary, the current solution is retained, and a warning is issued at the implementation level. This "single-step water level backfilling / recycling" process is equivalent to evenly distributing the total effective green difference among adjustable intersections without violating the upper and lower bound constraints, so that the final allocation scheme conforms as closely as possible to the total amount given by the network layer. .

[0164] Finally, the final effective green duration is generated based on the revised intersection-level control rate:

[0165] (1-33);

[0166] in, For the final green light time, for Always at the intersection The applied control rate, For effective green time.

[0167] This is then implemented in specific signal timing. In practice, all phase adjustments must meet engineering constraints such as minimum duration and single-cycle amplitude modulation to ensure the safety and feasibility of the control strategy in actual operation.

[0168] Specifically, such as Figure 5 As shown, the overall process of traffic flow allocation at the boundary control point is as follows:

[0169] First, calculate the optimal cumulative amount of the controlled sub-region. The system determines whether the current accumulated amount exceeds the optimal value; if it does not, it maintains the fixed signal timing and proceeds to the next time step update; if it exceeds the optimal value, it initiates boundary control and calculates the overall release strength. Then, a minimum effective green time is assigned to each boundary intersection. Calculate the remaining allocable green space. If the remaining green volume Then, the minimum green time is compressed proportionally to meet the budget requirements. Next, based on the three-dimensional coupling operator... Calculate the weights of each intersection and perform initial effective green light calculations. The allocation is performed, and the results are converted into intersection-level control rates, which are then truncated to upper and lower boundaries. Finally, the allocation difference for intersections that have not reached the boundary is adjusted through water level-based consistency correction to generate the final green light time. And timing adjustments are implemented. This process achieves precise control of traffic congestion through dynamic judgment and a multi-level allocation mechanism.

[0170] Based on the design of the specific control algorithm described above, this embodiment constructs a complete two-layer boundary control model. This model organically integrates the control logic of the network layer and the intersection layer through a systematic framework. Step S510 first performs total control budget processing. Based on the global release intensity given by the network layer and the sum of the effective green time corresponding to each control rate of the intersection layer, a preliminary control budget framework is established. This process, through the implementation of formulas (1-19) to (1-25), determines the allocation benchmark of the overall control resources. Step S520, based on this, performs hierarchical coupling coordination processing. The water level correction mechanism is applied to dynamically adjust the control rate allocation of the network layer and the intersection layer. The single-step backfill / recovery method of formulas (1-30) to (1-32) ensures the consistency of the total control, solves the allocation deviation problem caused by boundary constraints, and forms a coordinated and consistent coupled control strategy. Step S530 completes the model structuring process by integrating the proportional-integral control logic of the network layer with the three-dimensional coupling operator of the intersection layer to form a unified two-layer control architecture, as shown in formula (1-33). This transforms the theoretical control algorithm into a complete, executable control model. This construction process enables the two-layer boundary control model to maintain both the macro-control capability of the network layer and the fine-tuning advantage of the intersection layer, providing a complete solution for urban road network congestion management.

[0171] Further, step S600 includes steps S610 to S630.

[0172] Step S610: Based on the real-time acquired vehicle density and traffic flow data, perform data verification and preprocessing. By checking the data integrity and calculating the real-time traffic status indicators of each boundary intersection, the verified real-time traffic data is obtained.

[0173] Step S620: Based on the verified real-time traffic data input value, the dual-layer boundary control model performs dynamic optimization processing of signal timing. By adjusting the green light duration allocation of the network layer and the intersection layer to ensure consistency of the total control, an optimized green light timing scheme is obtained.

[0174] Step S630: Process traffic flow status output according to the optimized green light timing scheme. By implementing dynamic signal timing and monitoring changes in queue length and vehicle accumulation, the optimized traffic flow status is obtained.

[0175] To verify the effectiveness of the constructed boundary control model, this embodiment selects the actual urban road network of Zurich, Switzerland as an experimental case. Based on the completed road network sub-zone division, the macro-basic map (MFD) of the sub-zone is fitted based on real traffic data. The fitting of the macro-basic map not only reflects the overall characteristics of traffic flow within the sub-zone, but also provides key parameter support for the implementation of the boundary control strategy. According to the analysis of the division results, the traffic congestion problem in the red sub-zone is particularly prominent, especially during the morning and evening peak hours, when commuter traffic increases significantly, leading to a substantial decrease in road network operating efficiency. As a predominantly residential area, the red sub-zone bears a large amount of travel demand during peak hours, including commuter vehicles and public transportation traffic, resulting in a continuously high traffic load. To alleviate traffic congestion in this area and improve the overall operating efficiency of the road network, this embodiment selects the red area as the controlled sub-zone, such as... Figure 6 The controlled sub-regions will be subject to key boundary control measures. Simultaneously, other sub-regions will be designated as peripheral sub-regions to optimize traffic flow allocation between the controlled and peripheral sub-regions through boundary control, ensuring the coordinated operation of the entire road network.

[0176] Based on the traffic data collected by the detectors in each sub-zone, the relationship curve between vehicle accumulation and weighted flow rate in each sub-zone can be obtained, i.e., the macroscopic basic diagram (MFD). In this embodiment, the least squares method is used to fit the MFD of each sub-zone to ensure that the curve can accurately represent the flow-accumulation characteristics of each sub-zone. The fitting results are as follows: Figures 7-8 As shown in the figure, the MFD curves of each sub-region have a good fitting effect, and the fitting curves can clearly show the changing pattern between the cumulative vehicle volume and the weighted flow.

[0177] Furthermore, to quantify the fitting results, this embodiment further extracted and organized relevant parameters, the specific parameter values ​​of which are shown in Table 1. These parameters not only reflect the differences in traffic characteristics among the sub-regions, but also provide important references for subsequent traffic flow assignment and boundary control studies, such as the optimal cumulative volume for the region. .

[0178] Table 1 MFD Fitting Parameters

[0179]

[0180] Based on the extraction of macroscopic basic map parameters, this embodiment further conducted simulation experiments to verify the actual effect of the boundary control strategy. The initial simulation parameters were set based on real traffic data from the Zurich urban road network, selecting four hours of data from the afternoon to evening (3:00 PM to 7:00 PM) on weekdays as the research object. This time period covers the entire process of urban traffic gradually entering and exiting peak hours, and can comprehensively reflect the dynamic changes in traffic flow and the operational status of the road network. Specifically, the traffic demand input of the controlled area is as follows: Figure 9 As shown, the initial time is set to 15:00. To simulate real demand scenarios, the input traffic demand gradually increases until the evening peak begins. After maintaining high demand for a period of time, the cumulative number of vehicles in the road network gradually increases, traffic flow rises significantly, and the traffic pressure on some road sections gradually intensifies, leading to congestion. After the peak period ends, the demand gradually decreases.

[0181] For boundary intersections, based on actual road network data, the signal cycle range for sub-zone boundary intersections is [60s, 180s], with time loss within a single cycle. The saturation flow rate of the approach lanes at the boundary intersection is 1400 veh / h. Due to the differences in road classification among the approach lanes, their signal cycles and free-flow speeds also vary. The road segment length is determined based on actual measurement data, as detailed in Table 2. The relative location of the boundary intersection is... Figure 10 The text is marked with yellow dots.

[0182] Table 2 Specific information on intersection sections

[0183]

[0184] In simulation experiments, boundary control law The initial setting is 1, which sets the sub-region boundary control rate. and the control rate of each intersection The range is [0.1, 1], and the control period is 180s. Furthermore, to prevent drastic fluctuations in signal timing within a short period, reasonable constraints are imposed on the change in the boundary control rate. as well as To ensure the smoothness of the control process and the stability of the system, this embodiment conducts a comparative experiment based on a real urban road network to fully verify the effectiveness of the proposed two-layer boundary control method. Boundaryless control, single-layer Bang-Bang control, single-layer PI control, and two-layer Bang-Bang control are used as control strategies to evaluate the performance differences of each method under the same traffic demand. Experimental results show that the proposed two-layer PI control method performs best in terms of control rate smoothness, stability of road network vehicle accumulation, average speed improvement, and queue length suppression. Its control rate changes continuously and smoothly, effectively maintaining the sub-area near the optimal operating point. Simultaneously, the boundary queue distribution is more balanced, and local oversaturation is significantly alleviated, verifying the comprehensive advantages of "network layer and intersection layer coupled control" in improving traffic flow conditions.

[0185] Example 2:

[0186] like Figure 2 As shown, this embodiment provides a two-layer boundary control system for road network congestion sub-areas that considers queuing dynamics. The system includes:

[0187] The acquisition module 901 is used to acquire basic data of the target road network, including road network topology information and historical traffic data.

[0188] The processing module 902 is used to establish a macroscopic traffic flow model based on basic data. By deriving the vehicle conservation equation and performing linearization processing, a linearized macroscopic traffic flow model is obtained.

[0189] Module 903 is used to construct a network layer boundary control framework based on a linearized macroscopic traffic flow model, calculate the global release intensity using the proportional-integral control principle, and generate the global boundary control rate.

[0190] The coupling module 904 is used to establish an intersection layer control mechanism based on the global boundary control rate. It generates a differentiated intersection control rate by integrating the three-dimensional coupling operators of queue change trend, storage occupancy degree and demand intensity.

[0191] The integration module 905 is used to establish a two-layer boundary control model based on the global boundary control rate and the differentiated intersection control rate. By coupling the control logic of the network layer and the intersection layer, the two-layer boundary control model is obtained.

[0192] The output module 906 is used to input real-time acquired vehicle density and traffic flow data into the two-layer boundary control model to perform signal timing optimization, and output the optimized traffic flow state by dynamically adjusting the green light duration.

[0193] In one specific embodiment of this application, the processing module 902 includes:

[0194] The first processing unit is used to establish vehicle conservation equations based on basic data, construct the balance relationship between vehicle inflow and outflow in sub-regions based on macroscopic basic graph theory, and obtain the initial vehicle conservation equations.

[0195] The second processing unit is used to linearize the initial vehicle conservation equation by selecting the optimal vehicle accumulation in the sub-region as the set point and performing Taylor series expansion to obtain the linearized vehicle conservation equation.

[0196] The third processing unit is used to extract parameters based on the linearized vehicle conservation equations and obtain a linearized macroscopic traffic flow model by calculating the coefficient matrix in the linearized equations.

[0197] In one specific embodiment of this application, the construction module 903 includes:

[0198] The first building unit is used to determine the control setpoint based on the linearized macroscopic traffic flow model. By selecting the optimal vehicle accumulation in the macroscopic basic map as the target working point, the desired accumulation setpoint is obtained.

[0199] The second building unit is used to perform proportional-integral control processing based on the desired cumulative amount set value. It calculates the error between the current vehicle cumulative amount and the set value and applies the proportional-integral control law to dynamically adjust and obtain the preliminary boundary control law.

[0200] The third building unit is used to perform control rate constraint processing based on the preliminary boundary control rate. By applying minimum and maximum release strength limits, the feasibility of the control rate is ensured, and the global boundary control rate is obtained.

[0201] In one specific embodiment of this application, the coupling module 904 includes:

[0202] The first coupling unit is used to configure the basic parameters of the intersection based on the global boundary control rate. By determining the minimum effective green time and saturation flow parameters of each boundary intersection, the basic control parameters of the intersection are obtained.

[0203] The second coupling unit is used to construct a three-dimensional coupling operator based on the basic control parameters of the intersection. It calculates the comprehensive weight by integrating the ratio of adjacent period queue lengths, the road segment storage capacity occupancy rate, and the saturated flow demand intensity, and obtains the intersection coupling weight coefficient.

[0204] The third coupling unit is used to generate differentiated control rates based on the intersection coupling weight coefficients. The differentiated intersection control rates are obtained through normalized weight allocation and boundary truncation constraint adjustment.

[0205] In one specific embodiment of this application, the integration module 905 includes:

[0206] The first integration unit is used to process the total control budget based on the global boundary control rate and the differentiated intersection control rate. It obtains the preliminary control budget framework by calculating the sum of the global release intensity of the network layer and the effective green time corresponding to each control rate of the intersection layer.

[0207] The second integration unit is used to perform hierarchical coupling coordination processing based on the preliminary control budget framework. By applying a water level correction mechanism, it dynamically adjusts the control rate allocation between the network layer and the intersection layer to ensure that the total control amount is consistent, thereby obtaining the coupled control strategy.

[0208] The third integration unit is used to perform model structuring processing according to the coupling control strategy. By integrating the proportional-integral control logic of the network layer with the three-dimensional coupling operator of the intersection layer to form a unified control architecture, a two-layer boundary control model is obtained.

[0209] In one specific embodiment of this application, the output module 906 includes:

[0210] The first output unit is used to perform data verification and preprocessing based on the real-time acquired vehicle density and traffic flow data. By checking the data integrity and calculating the real-time traffic status indicators of each boundary intersection, the verified real-time traffic data is obtained.

[0211] The second output unit is used to perform dynamic optimization of signal timing based on the verified real-time traffic data input value of the dual-layer boundary control model. By adjusting the green light duration allocation of the network layer and the intersection layer, the total control volume is kept consistent, and an optimized green light timing scheme is obtained.

[0212] The third output unit is used to process traffic flow status output based on the optimized green light timing scheme. By implementing dynamic signal timing and monitoring changes in queue length and vehicle accumulation, the optimized traffic flow status is obtained.

[0213] The above are merely specific embodiments of the present invention, but the scope of protection of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in the present invention should be included within the scope of protection of the present invention.

Claims

1. A two-layer boundary control method for road network congestion sub-regions considering queuing dynamics, characterized in that, include: Acquire basic data of the target road network, including road network topology information and historical traffic data; A macroscopic traffic flow model is established based on the aforementioned basic data. By deriving the vehicle conservation equation and performing linearization, a linearized macroscopic traffic flow model is obtained. A network layer boundary control framework is constructed based on the linearized macroscopic traffic flow model. The global release intensity is calculated using the proportional-integral control principle, and the global boundary control rate is generated. An intersection layer control mechanism is established based on the global boundary control rate. By integrating the queue change trend, storage occupancy level and demand intensity three-dimensional coupling operator, a differentiated intersection control rate is generated. A two-layer boundary control model is established based on the global boundary control rate and the differentiated intersection control rate. The two-layer boundary control model is obtained by coupling the control logic of the network layer and the intersection layer. The real-time acquired vehicle density and traffic flow data are input into the dual-layer boundary control model to perform signal timing optimization, and the optimized traffic flow state is output by dynamically adjusting the green light duration.

2. The two-layer boundary control method for road network congestion sub-regions considering queuing dynamics as described in claim 1, characterized in that, A macroscopic traffic flow model is established based on the aforementioned basic data. By deriving the vehicle conservation equation and performing linearization, a linearized macroscopic traffic flow model is obtained, including: Based on the aforementioned basic data, the vehicle conservation equation is established. The balance relationship between vehicle inflow and outflow in sub-regions is constructed based on the macroscopic basic graph theory, resulting in the initial vehicle conservation equation. The initial vehicle conservation equation is linearized by selecting the optimal vehicle accumulation in the sub-region as the setpoint and performing Taylor series expansion to obtain the linearized vehicle conservation equation. Based on the linearized vehicle conservation equation, parameter extraction is performed, and by deriving the coefficient matrix in the linearized equation, a linearized macroscopic traffic flow model is obtained.

3. The two-layer boundary control method for road network congestion sub-regions considering queuing dynamics as described in claim 1, characterized in that, Based on the linearized macroscopic traffic flow model, a network layer boundary control framework is constructed. The global release intensity is calculated using the proportional-integral control principle, generating a global boundary control rate, including: The control setpoint is determined based on the linearized macroscopic traffic flow model. The desired cumulative volume setpoint is obtained by selecting the optimal vehicle accumulation in the macroscopic basic map as the target operating point. Proportional-integral control is performed based on the desired cumulative amount set value. The error between the current vehicle cumulative amount and the set value is calculated and dynamically adjusted using a proportional-integral control law to obtain the preliminary boundary control law. Based on the preliminary boundary control rate, control rate constraint processing is performed. By applying minimum and maximum release strength limits, the feasibility of the control rate is ensured, and the global boundary control rate is obtained.

4. The two-layer boundary control method for road network congestion sub-regions considering queuing dynamics as described in claim 1, characterized in that, An intersection-level control mechanism is established based on the global boundary control rate. By fusing a three-dimensional coupling operator of queue change trends, storage occupancy levels, and demand intensity, a differentiated intersection control rate is generated, including: The basic parameters of the intersection are configured according to the global boundary control rate. The basic control parameters of the intersection are obtained by determining the minimum effective green time and saturation flow parameters of each boundary intersection. Based on the basic control parameters of the intersection, a three-dimensional coupling operator is constructed. The comprehensive weight is calculated by integrating the ratio of adjacent period queue lengths, the road segment storage capacity occupancy rate, and the saturated flow demand intensity to obtain the intersection coupling weight coefficient. The differentiated control rate is generated based on the intersection coupling weight coefficient. The differentiated intersection control rate is obtained by adjusting the normalized weight allocation and boundary truncation constraints.

5. The two-layer boundary control method for road network congestion sub-regions considering queuing dynamics as described in claim 1, characterized in that, A two-layer boundary control model is established based on the global boundary control rate and the differentiated intersection control rate. By coupling the control logic of the network layer and the intersection layer, the two-layer boundary control model is obtained, including: Based on the global boundary control rate and the differentiated intersection control rate, the total control budget is processed. By calculating the sum of the global release intensity of the network layer and the effective green time corresponding to each control rate of the intersection layer, a preliminary control budget framework is obtained. Based on the preliminary control budget framework, hierarchical coupling coordination is performed. By applying a water level correction mechanism, the control rate allocation between the network layer and the intersection layer is dynamically adjusted to ensure that the total control amount is consistent, thus obtaining a coupled control strategy. Based on the aforementioned coupling control strategy, the model is structured and a unified control architecture is formed by integrating the network layer proportional-integral control logic with the intersection layer three-dimensional coupling operator, resulting in a two-layer boundary control model.

6. A two-layer boundary control system for road network congestion sub-regions considering queuing dynamics, characterized in that, include: The acquisition module is used to acquire basic data of the target road network, including road network topology information and historical traffic data. The processing module is used to establish a macroscopic traffic flow model based on the basic data, and to obtain a linearized macroscopic traffic flow model by deriving the vehicle conservation equation and performing linearization processing. The construction module is used to construct a network layer boundary control framework based on the linearized macro traffic flow model, calculate the global release intensity using the proportional-integral control principle, and generate the global boundary control rate. The coupling module is used to establish an intersection layer control mechanism based on the global boundary control rate. It generates a differentiated intersection control rate by integrating the three-dimensional coupling operators of queue change trend, storage occupancy degree and demand intensity. An integration module is used to establish a two-layer boundary control model based on the global boundary control rate and the differentiated intersection control rate. The two-layer boundary control model is obtained by coupling the control logic of the network layer and the intersection layer. The output module is used to input real-time acquired vehicle density and traffic flow data into the dual-layer boundary control model to perform signal timing optimization, and output the optimized traffic flow state by dynamically adjusting the green light duration.

7. The dual-layer boundary control system for road network congestion sub-regions considering queuing dynamics as described in claim 6, characterized in that, The processing module includes: The first processing unit is used to establish vehicle conservation equations based on the basic data, construct the balance relationship between vehicle inflow and outflow in sub-regions based on macroscopic basic graph theory, and obtain the initial vehicle conservation equations. The second processing unit is used to linearize the initial vehicle conservation equation by selecting the optimal vehicle accumulation in the sub-region as the set point and performing Taylor series expansion to obtain the linearized vehicle conservation equation. The third processing unit is used to extract parameters based on the linearized vehicle conservation equation and obtain a linearized macroscopic traffic flow model by deriving the coefficient matrix in the linearized equation.

8. The dual-layer boundary control system for road network congestion sub-regions considering queuing dynamics according to claim 6, characterized in that, The building module includes: The first construction unit is used to determine the control setpoint based on the linearized macro traffic flow model, and obtain the desired cumulative setpoint by selecting the optimal vehicle accumulation in the macro basic map as the target working point. The second construction unit is used to perform proportional-integral control processing based on the desired cumulative amount set value. By calculating the error between the current vehicle cumulative amount and the set value and applying the proportional-integral control law for dynamic adjustment, a preliminary boundary control law is obtained. The third construction unit is used to perform control rate constraint processing based on the preliminary boundary control rate, and to ensure the feasibility of the control rate by applying minimum and maximum release strength limits, thereby obtaining the global boundary control rate.

9. The dual-layer boundary control system for road network congestion sub-regions considering queuing dynamics according to claim 6, characterized in that, The coupling module includes: The first coupling unit is used to configure the basic parameters of the intersection according to the global boundary control rate. By determining the minimum effective green time and saturation flow parameters of each boundary intersection, the basic control parameters of the intersection are obtained. The second coupling unit is used to construct a three-dimensional coupling operator based on the basic control parameters of the intersection, and calculate the comprehensive weight by fusing the ratio of adjacent period queue lengths, the road segment storage capacity occupancy rate and the saturated flow demand intensity to obtain the intersection coupling weight coefficient. The third coupling unit is used to perform differentiated control rate generation processing based on the intersection coupling weight coefficient, and obtain the differentiated intersection control rate through normalized weight allocation and boundary truncation constraint adjustment.

10. The dual-layer boundary control system for road network congestion sub-regions considering queuing dynamics according to claim 6, characterized in that, The integration module includes: The first integration unit is used to perform total control budget processing based on the global boundary control rate and the differentiated intersection control rate. By calculating the sum of the effective green time corresponding to the global release intensity of the network layer and each control rate of the intersection layer, a preliminary control budget framework is obtained. The second integration unit is used to perform hierarchical coupling coordination processing based on the preliminary control budget framework. By applying a water level correction mechanism, the control rate allocation between the network layer and the intersection layer is dynamically adjusted to ensure that the total control amount is consistent, thereby obtaining a coupled control strategy. The third integration unit is used to perform model structuring processing according to the coupling control strategy. By integrating the network layer proportional-integral control logic with the intersection layer three-dimensional coupling operator to form a unified control architecture, a two-layer boundary control model is obtained.

Images