A feedback control-based collaborative control method for multiple bottleneck ramps and connecting road networks on expressways

Abstract

This invention provides a collaborative control method for multiple bottleneck ramps and connecting road networks on expressways based on feedback control. This method treats the entire expressway mainline and each sub-region of the urban road network as continuous entities that can be represented by a macroscopic basic graph model, constructing a feedback control system based on real-time traffic detection data. The entire expressway mainline adopts dynamic pre-set point proportional-integral (PI) feedback control based on the macroscopic basic graph, while each urban road network unit adopts classical PI feedback control. The method includes the following steps: S1, construction of the research object and control unit; S2, traffic state parameters and model assumptions; S3, construction of the upper-level control framework; S4, feedback control decision system based on the macroscopic basic graph; S5, allocation and execution of the lower-level control total flow in individual control objects. The method proposed in this invention utilizes a simple and highly robust macroscopic basic graph to efficiently achieve collaborative control of large-scale expressways and urban road networks, thereby improving the overall traffic efficiency of the transportation system and providing a feasible new approach for urban traffic congestion management.

Description

Technical Field

[0001] This invention relates to the field of intelligent traffic control technology, specifically to a method for coordinated control of multiple bottleneck ramps and connecting road networks on expressways based on feedback control. Background Technology

[0002] With the continuous acceleration of urbanization and the rapid growth of motor vehicle ownership, the contradiction between urban traffic supply and demand is becoming increasingly prominent. Expressways, as the backbone of the urban road system, are characterized by independent right-of-way, high traffic volume, and rapid transit. The dense distribution of ramps leads to multiple adjacent traffic bottlenecks on long-distance expressway sections during peak hours. Furthermore, the close connection between urban surface road networks and expressways exacerbates the congestion range of these bottlenecks, increasingly creating synchronous congestion with connecting surface road networks, severely impacting the efficiency and operational stability of the urban traffic system. Under critical congestion conditions, significant traffic flow interactions exist between bottlenecks on the expressway mainline, and the traffic flow between expressways and the surface road network exhibits a clear competitive and cooperative relationship. This necessitates coordinated control of the expressway mainline as a whole, as well as its integration with the urban surface road network system. Coordinated control of expressways and connecting road networks is not only an effective means of alleviating traffic congestion and improving road network efficiency, but also a prerequisite for tapping the potential of existing road resources and achieving refined urban traffic management. It has become a pressing issue in my country's current urban traffic congestion management.

[0003] However, current collaborative control technologies for multiple bottlenecks and connecting road networks on expressways face significant technical bottlenecks in practical applications. On the one hand, existing expressway ramp coordination control methods are mostly based on segment basic maps, making it difficult to achieve overall collaborative control of the mainline, especially when the expressway mainline is large. In recent years, feedback control technology based on the macroscopic basic map (MFD) of expressways has provided a new approach to expressway collaborative control due to its advantages of being able to reflect the overall operating status of the road network simply and with high robustness, and adapting to large-scale collaborative regulation. However, existing MFD-based feedback control methods generally use fixed preset points, making it difficult to adapt to the control requirements of dynamic changes in the overall and local bottlenecks of the expressway mainline. On the other hand, in traffic control of expressways and connecting road networks, feedback control based on macroscopic basic maps is only used for the ground road network and has not yet been simultaneously applied to the expressway section. This affects the effective balance between computational efficiency and control effect in large-scale road network collaborative control; furthermore, at the lower level of control, the vehicle queuing problem caused by control still needs to be handled with finer detail.

[0004] Patent publication number "CN115035734A" discloses a method for controlling the entrance ramp signals of urban expressways based on congestion status classification. The proposed control method uses congestion index classification as its core, employing mainline traffic flow as input and ramp signal green ratio as output for graded feedback adjustment to achieve local merging management at the entrance of a single ramp (or a small number of ramps), thereby improving ramp merging efficiency. However, it suffers from limitations, including restricting the control object to a few local ramp areas of the expressway, failing to consider traffic control coordination between the expressway and the urban road network, and lacking a feasible hierarchical traffic allocation mechanism.

[0005] Patent publication number "CN116913093A" discloses a smart highway collaborative management and control method based on feedback control. The proposed control method emphasizes state restoration and prediction, and establishes a collaborative model for variable speed limits and ramp management to obtain the optimal strategy. It divides the highway into control sections according to the location of control equipment, and conducts collaborative management and control of the highway mainline and entrance / exit ramps to improve highway traffic efficiency. However, it suffers from several drawbacks: the control object does not cover the connection to the urban road network; it lacks a unified and constrained upper-level control action (i.e., total ramp traffic flow, or total traffic flow at the directional boundaries between various road network units) for ramp coordination control and urban road network signal control; and it lacks an executable mechanism for allocating total traffic flow to individual ramps or control sections.

[0006] Patent publication number "CN114023068A" discloses a short-term traffic flow prediction and active control system and method for short-distance weaving areas. The proposed control method is designed for local scenarios in short-distance weaving areas. It combines real-time multi-target radar acquisition and edge computing, employing methods such as spatiotemporal graph convolutional networks to construct a traffic flow prediction model. This model performs short-term prediction of traffic flow in the weaving area and outputs entrance and exit ramp signal control schemes to achieve active control of local bottlenecks. However, its application is limited to microscopic local areas, failing to consider ramp coordination control on large-scale expressways; it relies on a learning-based prediction model, making it difficult to adapt to complex and changing traffic environments, potentially leading to significant control performance losses due to fluctuations in traffic models and traffic demand; and it does not consider traffic control coordination across bottlenecks and between expressways and urban road networks, making it difficult to achieve traffic efficiency improvements on a large scale or even globally.

[0007] Therefore, facing the challenges of coupled control of multiple bottlenecks on expressways and connecting road networks, and the practical needs of urban traffic congestion management in my country, there is an urgent need to propose a collaborative control method for multiple bottleneck ramps and connecting road networks on expressways based on feedback control. This method involves constructing a large-scale road network traffic improvement model, feedback decision-making for the entire expressway and individual urban road network units, and a collaborative control mechanism for refined control of total traffic flow distribution at various control points (ramps or control sections) and closed-loop regulation. This approach effectively balances the control effect and computational cost of large-scale road networks, providing a reliable, efficient, and precise collaborative control solution for urban traffic congestion management in my country. Summary of the Invention

[0008] The purpose of this invention is to propose a collaborative control method for multiple bottleneck ramps and connecting road networks on expressways based on feedback control. The method abstracts the entire expressway mainline and each urban road network unit as continuous entities that can be represented by a macroscopic basic diagram. Upper-level control uses dynamic preset-point PI feedback control for the total traffic flow at expressway entrance ramps, and classical PI feedback control for the total inflow at the boundary sections of each urban road network unit. Lower-level control considers constraints such as traffic light green light duration and vehicle queuing caused by direct control, allocating the total traffic flow to specific ramps and boundary control sections. This is further mapped to signal timing instructions and executed in the next cycle, forming a rolling closed-loop control, thereby improving the traffic efficiency of the transportation system.

[0009] To achieve the above objectives, the technical solution adopted by the present invention is as follows:

[0010] This invention provides a method for coordinated control of multiple bottleneck ramps and connecting road networks on expressways based on feedback control, comprising the following steps:

[0011] S1. Construction of research objects and control units;

[0012] S2, Traffic State Parameters and Model Assumptions;

[0013] S3. Construction of the upper-level control framework;

[0014] S4. Feedback control decision system based on macroscopic basic diagram;

[0015] S5. The allocation and execution of total lower-level control flow within a single control object.

[0016] As a further improvement, the construction of the research object and control unit in step S1 is carried out in the following specific process:

[0017] Step 1: Select long-distance expressway sections with multiple bottlenecks, deploy a series of detectors along the route, and collect data including at least the unit density, flow rate, speed of each section of the main line, and the vehicle queue length of each entrance ramp;

[0018] Step 2: Divide the urban surface road network that connects with the expressway into a certain number of sub-road network units as appropriate, so that each urban sub-road network unit can meet the requirements of having a macroscopic basic map with low dispersion, and define the set of boundary control road segments that connect with adjacent urban sub-road network units in each urban sub-road network unit.

[0019] As a further improvement, the object and control unit construction process in step S1 also includes the following:

[0020] Multiple sets of coil detectors are deployed along the main line, and the main line is divided into sections using the midpoint of adjacent detectors as endpoints. Each road segment unit, at each discrete time step Based on the traffic data obtained from the detectors, road segment units are obtained. density With traffic Based on this, the overall weighted average density of the main line is calculated. Weighted average flow :

[0021] The formula for calculating the overall weighted average density of the main line is as follows:

[0022] ;

[0023] The formula for calculating the overall weighted average flow rate of the main line is as follows:

[0024] ;

[0025] in, It is a road segment unit The length.

[0026] As a further improvement, the traffic state parameters and model assumptions in step S2 are as follows:

[0027] Step 1: Assuming the overall traffic flow dynamics of the expressway mainline can be discretized using a macroscopic basic graphical model:

[0028] In a discrete time step Within this system, a weighted average traffic flow can be constructed for the main line of the expressway. With weighted average density The functional relationship between density distribution standard deviation and entrance ramp traffic flow is considered as a controllable constraint on the input flow of the expressway mainline.

[0029] Step 2: For each urban sub-road network unit, assume that it can be discretized and modeled according to the macroscopic basic graph theory:

[0030] In a discrete time step Within this, the weighted average traffic flow of urban sub-road network units can be provided. With weighted average density The functional relationship is defined, and the traffic flow of the road segment overlapping with the urban road network is regarded as a controllable constraint on the input flow of the sub-road network unit.

[0031] As a further improvement, the construction of the upper-level control framework in step S3 is carried out as follows:

[0032] Step 1: Construct the upper-level control action vector:

[0033] Each control cycle The total traffic flow of the ramps corresponding to the main line of the expressway, and the total traffic flow of the control sections at the directional boundaries between urban sub-road network units, wherein the total traffic flow of the expressway entrance ramps is denoted as... The total inflow of the boundary road segment of urban sub-road network unit m is denoted as ;

[0034] Step 2: Control Objectives and Evaluation Indicators

[0035] With the overall control objective of reducing the total travel time (TTS) of the system, the control objective can be further specified in each control cycle as "increasing the outbound flow of the system in each control cycle".

[0036] The evaluation weight of each road network unit can be set according to the proportion of its own traffic capacity to the system's traffic capacity;

[0037] Step 3: Constructing Urban Road Network Units The feedback control framework needs to obtain its weighted average density through data acquisition and processing. Specifically, this can be achieved by installing detectors on key road sections within urban road network units.

[0038] As a further improvement, step S4 is based on a feedback control decision system of macroscopic basic graph, and the specific process is as follows:

[0039] Step 1: Expressway Mainline Feedback Control:

[0040] Weighted average density based on main line Multiply by the total length of the main line to get the total number of vehicles. As input, the total flow of traffic entering the main line from the ramp. For the output quantity, a dynamic preset point PI control algorithm is adopted:

[0041] Discrete equilibrium equation for cumulative vehicle count on the main line:

[0042] ;

[0043] in, It is the collection of all entrance ramps. It is the entrance ramp. During the control cycle Traffic, It is the total traffic demand for uncontrolled entry into the main line. This refers to the total outbound traffic volume on the main line. To control the duration of the cycle;

[0044] The total outbound traffic flow on the main line includes the outbound traffic flow from the exit ramps and the main line terminus. It is assumed that the weighted average flow rate can be used to calculate the flow rate. Linear estimation, the Estimated using a macroscopic fundamental graph model;

[0045] Dynamic preset point PI feedback control algorithm:

[0046] ;

[0047] in, This refers to the total traffic flow from the entrance ramp (entering the main line). and These are the non-negative proportional and integral parameters, respectively. It is a dynamic weighted average density preset value;

[0048] The update rules are as follows:

[0049] ;

[0050] in, This refers to the numbering of the main line segment unit. This is the total number of main line segment units. and They are road segment units Actual density and critical density, It is the preset value of the (fixed) weighted average density of the main line. In the cycle The critical weighted average density when congestion occurs on the main line;

[0051] The The value is slightly smaller than the key weighted average density determined by the principal MFD, the The calculation formula is as follows:

[0052] ;

[0053] ;

[0054] in, It is a road segment unit Length;

[0055] Step 2: Urban road network unit feedback control:

[0056] For each urban road network unit m, its weighted average density The input quantity is the total inflow of its boundary segment. For the output quantity, the classic PI feedback control method is adopted, and the specific formula is as follows:

[0057] ;

[0058] in, and These are the non-negative proportional and integral parameters, respectively. It is a (fixed) weighted average density preset value, which is slightly smaller than the key weighted average density determined by the urban road network unit MFD.

[0059] As a further improvement, the PI feedback control parameters , , as well as Obtained through offline calibration, and is to be determined. and Constraints are imposed to ensure that control actions meet the traffic flow constraints of ramps and boundary control sections.

[0060] As a further improvement, the allocation and execution of the lower-level control total flow in step S5 within a single control object is as follows:

[0061] Step 1: In the current control cycle The total traffic flow of the expressway ramps obtained in step S4 and the total flow at the boundaries of each city's road network unit The traffic flow is then allocated to the corresponding specific ramps and boundary control sections to obtain the expected traffic flow for each ramp and boundary control section.

[0062] During this allocation process, the following constraints must be met simultaneously:

[0063] The upper and lower limits of green light time, road segment receiving capacity, and the minimum and maximum total flow constraints in step S3, as well as the vehicle queuing overflow directly caused by suppression control;

[0064] Step 2: Map the expected traffic flow of each ramp and boundary control section to the executable signal phase green light duration, and execute it in the next control cycle;

[0065] After the control is executed, the detection data is collected again, the state vector is updated, and steps S3–S5 are repeated to form a rolling closed-loop control, realizing real-time coordinated control of multiple bottlenecks on the expressway and connecting urban road networks.

[0066] As a further improvement, the control is executed in a fixed control cycle, completing "data acquisition - traffic status update - feedback control decision - total flow allocation and signal execution" in each control cycle, and repeating iteratively in the next control cycle to form a closed-loop control.

[0067] As a further improvement, a feedback control-based collaborative control system for multiple bottleneck ramps and connecting road networks on expressways includes:

[0068] The data acquisition module is used to collect data including at least the unit density, flow rate, and speed of each section of the main line, as well as the queue length of vehicles at each entrance ramp and boundary control section.

[0069] The modular modeling module is used to construct expressway segment units and urban sub-road network units, and to define the set of boundary control segments for urban sub-road network units;

[0070] The traffic condition modeling module is used in step S2 to establish the overall traffic flow evolution model of the expressway and the macro-basic graph model of the urban sub-road network units;

[0071] The upper-level control collaborative decision-making module is used to calculate the overall weighted average density of the expressway mainline in steps S3 and S4, and output the total flow of the entrance ramps; at the same time, it calculates the fixed preset points in the feedback control for each urban road network unit, and outputs the total inbound flow of the boundary control section.

[0072] The lower-level control allocation and execution module is used in step S5 to allocate various total traffic flows to individual ramps and boundary control sections, and map them to the corresponding signal phase green light durations;

[0073] The closed-loop feedback module is used to update the macro traffic status of the expressway mainline and various urban road network units in the next control cycle, and to execute collaborative feedback control in a rolling iterative manner.

[0074] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention adopts a feedback control strategy based on macroscopic fundamental graphs, directly carrying out macroscopic traffic control decisions at the overall level of the expressway mainline and at the level of each urban road network unit. It has the characteristics of low computational cost, robustness to model errors and random disturbances, and easy acceptance by managers and drivers, making it suitable for real-time control of large-scale mixed road networks. At the same time, this invention uses macroscopic fundamental graph theory to construct a hierarchical control architecture in both long-distance expressway mainlines and multi-city road network sections. The upper-level control determines the total traffic flow of the expressway entrance ramps and the total inbound traffic flow of each urban road network unit boundary segment, while the lower-level control realizes the executable mapping of each total traffic flow to specific ramps and intersection control segments. This architecture has the advantages of combining overall and local traffic control, balancing the computational cost and implementation effect of large-scale traffic control. Attached Figure Description

[0075] Figure 1 Satellite map of the research section and connecting road network of the Shanghai Yan'an Elevated Road (expressway);

[0076] Figure 2 A collaborative feedback control framework for multiple bottlenecks and connecting road networks of expressways based on macroscopic basic maps;

[0077] Figure 3 A macro-level basic map of the urban road network for each control scheme;

[0078] Figure 4 A macroscopic basic diagram of the main line of the expressway for each control scheme;

[0079] Figure 5 Accumulate the number of vehicles for different road networks under each control scheme;

[0080] Figure 6 Spatiotemporal heatmap of expressway density for each control scheme;

[0081] Figure 7 A comparison of the total travel time (TTS) of each control scheme system. Detailed Implementation

[0082] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the following specific embodiments are only used to illustrate the present invention and are not intended to limit the scope of the present invention:

[0083] like Figure 1 As shown, this embodiment selects an approximately 11.4km section of expressway from the Shanghai Yan'an Elevated Outer Ring Expressway to the North-South Elevated Expressway and its connecting road network as the research object. The main line of this expressway has a speed limit of 80km / h, 2-5 lanes, and includes 7 entrance ramps and 4 exit ramps. Detectors are deployed on the main line to provide data such as cross-sectional density, traffic flow, and speed. The expressway section is directly connected to two urban road networks and is further divided into 4 urban sub-road network units, with areas of 11.9km², 9.1km², 7.3km², and 7.2km² from left to right.

[0084] In this embodiment, control is achieved through dynamic signal updates at boundary intersections: since the length of the entrance ramps is between 200 and 400 meters and no traffic lights are installed, ramp control is adjusted by the signal of the adjacent upstream urban intersection; the control cycle is uniformly set to 60 seconds, and the expected traffic flow of each controlled road segment is converted into phase green light time for implementation; at the same time, queuing space constraints are considered, for example, the total queuing space of the controlled road segments leading to the expressway via entrance ramps 1-3 and 4-6 in urban road network 1 and 2 is 410 vehicles and 305 vehicles, respectively.

[0085] like Figure 2 As shown, this embodiment provides a method for coordinated control of multiple bottleneck ramps and connecting road networks on expressways based on feedback control, including the following steps:

[0086] S1. Construction of research objects and control units;

[0087] S2, Traffic State Parameters and Model Assumptions;

[0088] S3. Construction of the upper-level control framework;

[0089] S4. Feedback control decision system based on macroscopic basic diagram;

[0090] S5. The allocation and execution of total lower-level control flow within a single control object.

[0091] Step S1 is as follows:

[0092] Step 1: Select a long-distance expressway section with multiple bottlenecks, deploy a series of detectors along the route, and divide the main line into several road segment units with the midpoint of adjacent detectors as the endpoint. Collect data including at least the density, flow rate, speed of each road segment unit of the main line, and the vehicle queue length of each entrance ramp.

[0093] Step 2: Divide the urban surface road network connecting with the expressway into a certain number of sub-road network units as appropriate, so that each urban sub-road network unit can meet the requirements of having a macroscopic basic map with low dispersion, and define the set of boundary control road segments connecting with adjacent urban sub-road network units in each urban sub-road network unit for implementing dynamic traffic flow control.

[0094] Step S2 is as follows:

[0095] Step 1: Assuming the overall traffic flow dynamics of the expressway mainline can be discretized using a macroscopic basic graphical model: in a discrete time step Within this system, a weighted average traffic flow can be constructed for the main line of the expressway. With weighted average density The functional relationship between the standard deviation of the density distribution and the traffic flow. Furthermore, the traffic flow at the entrance ramps is considered a controllable constraint on the input traffic flow to the main expressway.

[0096] Step 2: For each city sub-road network unit, assume it can be discretized and modeled according to the macroscopic basic graph theory: in a discrete time step Within this, the weighted average traffic flow of urban sub-road network units can be provided. With weighted average density The functional relationship is as follows. Furthermore, the traffic flow of the road segments overlapping with the urban road network boundary is considered a controllable constraint on the input flow of the sub-road network units.

[0097] Step S3 is as follows:

[0098] Step 1: Construct the upper-level control action vector:

[0099] Each control cycle This refers to the total traffic flow of the ramps corresponding to the main line of the expressway, and the total traffic flow of the control sections at the directional boundaries between urban sub-road network units. The total traffic flow of the expressway entrance ramps is denoted as... The total inflow of the boundary road segment of urban sub-road network unit m is denoted as .

[0100] Step 2: Control Objectives and Evaluation Indicators

[0101] The overall control objective is to reduce the total travel time (TTS) of the system. Within each control cycle, the control objective can be further specified as "increasing the outbound flow rate of the system within each control cycle."

[0102] The evaluation weight of each road network unit can be set according to the proportion of its own traffic capacity to the system's traffic capacity.

[0103] Step S4 is as follows:

[0104] Step 1: Expressway Mainline Feedback Control: Based on the weighted average density of the entire mainline (Multiplying this by the total length of the main line gives the cumulative number of vehicles) () is the input quantity, representing the total flow rate entering the main line from the entrance ramp. For the output quantity, a dynamic preset point PI control algorithm is adopted:

[0105] Discrete equilibrium equation for cumulative vehicle count on the main line:

[0106] ;

[0107] in, It is the collection of all entrance ramps; It is the entrance ramp. During the control cycle Traffic; It is the total traffic demand that enters the main line without control. It is the total outbound traffic flow on the main line (including exit ramps and outbound traffic flow at the end of the main line), assuming it can be calculated from the weighted average traffic flow. Linear estimation, while the latter can be estimated using a macroscopic fundamental graph model; To control the duration of the cycle.

[0108] Weighted average density of the main line of the expressway As the input, a dynamic preset point PI feedback control is used to calculate the target value of the total flow rate at the entrance ramp. Its discrete control law is:

[0109] ;

[0110] in, These are the preset points for the dynamic density of the entire expressway mainline.

[0111] Step 2: Urban Road Network Unit Feedback Control: Based on urban road network units Weighted average density As the input, the target value of the total inflow at the boundary of the urban road network unit is calculated using classical PI feedback control. Its control law is:

[0112] ;

[0113] in, For the first Fixed density preset points for each urban sub-road network unit, subscript This corresponds to the four urban sub-road network units divided in this embodiment.

[0114] Step S5 is as follows:

[0115] Step 1: Distribute total flow by direction:

[0116] The total traffic flow of expressway ramps output by the macro control layer Total traffic flow at the boundaries of each city's road network unit The traffic flow is allocated to specific ramps and boundary intersection control sections to obtain their expected flow rates in the current control cycle. The total traffic flow allocation for urban road network units is constructed with the core objectives of "reducing the heterogeneity of density distribution of urban road network units" and "avoiding queuing saturation and overflow in control sections". In this allocation process, the following constraints are simultaneously met: upper and lower bounds of green light time, road segment receiving capacity, and the minimum and maximum total traffic flow constraints in step S3, as well as suppressing vehicle queuing overflow directly caused by control.

[0117] Step 2: Generate executable signal timing instructions:

[0118] The desired flow is mapped to an executable signal phase green light duration and executed in the next control cycle;

[0119] For dynamic signal control intersections, only the phases of "through-traffic expressways or urban road network units" are actively regulated, while the other phases are executed according to the predetermined timing or minimum guaranteed timing.

[0120] Step 3: Closed-loop rolling:

[0121] After the control is executed, the detection data is collected again, the state vector is updated, and steps S3–S5 are repeated to form a rolling closed-loop control, realizing real-time coordinated control of multiple bottlenecks on the expressway and connecting urban road networks.

[0122] To verify the effectiveness of the method of the present invention, this embodiment sets up a simulation experiment and comparison scheme, as follows:

[0123] The main line and ramps of the expressway are simulated using the METANET model to simulate traffic evolution, while the urban sub-road network is simulated using the macroscopic basic graph model. Traffic demand is obtained from the morning peak demand (06:30–08:50, including 20 minutes of warm-up, effective simulation time of 2 hours) collected from taxi GPS, expressway detectors and field surveys, and discretized with a Gaussian distribution to reflect random fluctuations (mean is taken from the survey value, standard deviation is taken as 4% of the mean).

[0124] This embodiment sets up only two comparison schemes: the no-control scheme (NC) and the scheme of the present invention (MFD-FC); the evaluation index is uniformly adopted as the total travel time (TTS) of the hybrid road network system, which is used to intuitively compare the control effects of the two schemes.

[0125] like Figure 7 As shown, the MFD-FC scheme of this invention achieves significant benefits compared to the uncontrolled NC scheme: the total system TTS is reduced by 6.5%, with the TTS of the expressway reduced by 15.9%, the TTS of urban road network 2 reduced by 10.3%, and the TTS of urban road network 1 slightly increased (2.3%). This result demonstrates that under the constraint of limited control space (limited control queuing capacity), the control strategy of this invention can dynamically balance the efficiency of the expressway mainline with the queuing overflow risk at the boundary of the urban sub-road network, ultimately achieving a significant improvement in the overall system efficiency.

[0126] Furthermore, combined Figures 3-6 The following supplementary explanations are provided regarding the control mechanism and congestion propagation process of this invention:

[0127] (1) From the perspective of macro-basic charts, such as Figure 3 The results show that the MFD-FC scheme can improve the efficiency of urban sub-road networks by adjusting the boundary flow size in advance during peak hours, making the operation of the sub-road network closer to its maximum flow.

[0128] (2) From the perspective of expressway congestion, such as Figure 4 The results show that, under the combined effect of queuing space constraints and urban road network congestion control requirements, local bottlenecks on expressways may still cause congestion during specific periods, but the MFD-FC scheme can effectively reduce the scope of congestion and prevent its further deterioration.

[0129] (3) From the perspective of the evolution of the cumulative number of vehicles in the urban road network, such as Figure 5 The results show that the MFD-FC scheme, through feedback control, can shorten the duration of the high cumulative vehicle count state in the system, thus maintaining the overall cumulative vehicle count of the mixed road network at an ideal level.

[0130] (4) From the perspective of the spatiotemporal distribution of expressway density, such as Figure 6As shown, under the control of the MFD-FC scheme, the spatial range and duration of high-density areas are effectively suppressed, the congestion spread process is avoided, and the overall operation of the expressway is more stable.

[0131] In summary, the method of the present invention can significantly reduce the total travel time of the system compared with the uncontrolled method, and shows a consistent improvement effect in three aspects: macroscopic basic map morphology, cumulative number of vehicles in urban road network and spatiotemporal distribution of expressway density, thus verifying the effectiveness and practicality of the present invention.

[0132] The control system described in this invention can be implemented through a combination of hardware and software, as detailed below:

[0133] Typically, a feedback control-based collaborative control system for expressway multi-bottleneck ramps and connecting road networks can be deployed at the control center. This system includes the following functional modules:

[0134] The data acquisition module is used to collect data including at least the unit density, flow rate, and speed of each section of the main line, as well as the queue length of vehicles at each entrance ramp and boundary control section.

[0135] The modular modeling module is used to construct expressway segment units and urban sub-road network units, and to define the set of boundary control segments for urban sub-road network units;

[0136] The traffic condition modeling module is used in step S2 to establish the overall traffic flow evolution model of the expressway and the macro-basic graph model of the urban sub-road network units;

[0137] The upper-level control collaborative decision-making module is used to calculate the overall weighted average density of the main line of the expressway and output the total flow of the entrance ramps in steps S3 and S4. At the same time, it calculates the fixed preset points in the feedback control of each urban road network unit and outputs the total inflow of the boundary control section.

[0138] The lower-level control allocation and execution module is used in step S5 to allocate various total traffic flows to individual ramps and boundary control sections, and map them to the corresponding signal phase green light durations;

[0139] The closed-loop feedback module is used to update the macro traffic status of the expressway mainline and various urban road network units in the next control cycle, and to execute collaborative feedback control in a rolling iterative manner.

[0140] The above modules can be implemented using industrial computers, embedded controllers, or servers, and interact with field signal lights via communication networks to issue commands and ensure the real-time execution of control commands.

[0141] Furthermore, the present invention can also be implemented by a computer-readable storage medium storing a computer program that, when run on a processor, executes all the steps of "a feedback-based coordinated control of multiple bottleneck ramps and connecting road networks on expressways" described in the above embodiments.

[0142] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any other way. Any modifications or equivalent changes made based on the technical essence of the present invention shall still fall within the scope of protection claimed by the present invention.

Claims

1. A method for freeway multi-bottleneck ramp and coordinated control of the connected road network based on feedback control, characterized in that, It comprises the following steps: S1, research object and control unit construction; S2, traffic state parameters and model assumptions; S3, upper layer control framework construction; S4, feedback control decision system based on macroscopic fundamental diagram; S5, allocation and execution of total flow in single control object in lower layer control.

2. The method according to claim 1, wherein, The specific process of the research object and control unit construction in step S1 is as follows: Step 1: Select a long-distance expressway section containing multiple bottlenecks, arrange a series of detectors along the line, and collect data including at least the unit density, flow, and speed of each road section of the main line, the vehicle queue length of each entrance ramp, etc. Step 2: Divide each city ground road network connected with the expressway into a certain number of sub-road network units, so that each city sub-road network unit can meet the requirements of having a low-dispersion macroscopic fundamental diagram, and define the boundary control road section set connected with adjacent city sub-road network units in each city sub-road network unit.

3. The method of claim 1, wherein, The research object and control unit construction process in step S1 also includes the following content: A plurality of groups of coil detectors are arranged along the main line, and the main line is divided into a plurality of section units with the midpoints of adjacent detectors as endpoints . At each discrete time step , the density and flow of each section unit are obtained according to the traffic data obtained by the detectors, and the overall weighted average density and weighted average flow of the main line are calculated according to the density and flow of each section unit. The calculation formula of the overall weighted average density of the main line is: ； The calculation formula of the overall weighted average flow of the main line is: ； wherein is the length of the road segment unit .

4. The method of claim 1, wherein, The specific process of traffic state parameters and model assumptions in step S2 is as follows: Step 1: Assume that the dynamic evolution of the overall traffic flow of the main line of the expressway can be discretely described by the macroscopic fundamental diagram model: At a discrete time step The weighted average flow of the mainline of the expressway can be constructed The weighted average density The function relationship between the density distribution standard deviation, while the entrance ramp traffic flow is considered as a controllable constraint on the input flow of the mainline of the expressway Step 2: For each city sub-road network unit, assume that it can be discretely modeled according to the macroscopic fundamental diagram theory: At a discrete time step The weighted average flow of the urban sub-road network unit can be given And the function relationship of the weighted average density At the same time, the traffic flow of the overlapping boundary section of the urban road network is regarded as a controllable constraint on the input flow of the sub-road network unit.

5. The method of claim 1, wherein, The specific process of upper layer control framework construction in step S3 is as follows: Step 1: Construct the upper layer control action vector: Each control cycle , the total flow of the ramp corresponding to the whole main line of the expressway, and the total flow of the control road section between the sub-units of the urban road network in the split direction, wherein the total flow of the entrance ramp of the expressway is denoted as , the total incoming flow of the boundary road section of the sub-unit m of the urban road network is denoted as ; Step 2: Control objectives and evaluation indicators: Take reducing the total travel time TTS as the overall control objective, and in each control period, further specify the control objective as "improving the system exit flow in each control period"; Wherein, the evaluation weight of each road network unit can be set according to the proportion of its traffic capacity in the system traffic capacity; Step 3: Constructing the urban road network unit The feedback control framework of the proposed method requires the weighted average density of the traffic flow to be obtained through data collection and processing , which can be achieved by setting detectors on the (key) road segments within the urban road network unit.

6. The method of claim 1, wherein, The specific process of the feedback control decision system based on the macroscopic fundamental diagram in step S4 is as follows: Step 1: Expressway main line feedback control: The main line weighted average density The main line total length is multiplied by the cumulative vehicle number As an input, the total flow of the ramp into the main line As an output, the dynamic preset point PI control algorithm is used: Main line cumulative vehicle number discrete balance equation: ； wherein, is the set of all on-ramps, is an on-ramp the flow of traffic, at the control period, is the total traffic demand entering the mainline without control, is the total flow of traffic exiting the mainline, is the length of the control period; The mainline total on- and off-flow includes off-ramp and mainline end off-flow, and the It is assumed that the weighted average flow Linear estimation, the Estimation by macro fundamental model Dynamic preset point PI feedback control algorithm: ； where is the total flow from the on-ramp (entering the mainline), and are non-negative proportional and integral parameters, respectively, is a dynamic weighted average density preset; The update rule is shown below: ； wherein, is the number of the main line section unit, is the total number of the main line section units, and are the actual density and the critical density of the section unit, respectively, is the (fixed) weighted average density preset value of the main line as a whole, is the critical weighted average density when the main line is congested at the period . The slightly less than the critical weighted average density determined by the main line MFD, the The calculation formula is as follows: ； ； wherein is the length of a road segment unit ​ Step 2: City road network unit feedback control: For each urban road network unit m, with its weighted average density For the input quantity, with its total incoming flow of boundary road segments For the output quantity, the classical PI feedback control method is adopted, and the specific formula is: ； wherein, and are non-negative proportional and integral parameters, respectively, is a (fixed) weighted average density preset value, which is slightly less than the critical weighted average density determined by the urban road network cell MFD.

7. The method of claim 6, wherein, The PI feedback control parameters , , And are obtained through offline calibration and are constrained to the to-be-solved and .

8. The method of claim 1, wherein, The specific process of the allocation and execution of total flow in single control object in lower layer control in step S5 is as follows: Step 1: in the current control cycle , the total flow of the expressway ramp obtained in step S4 and the total flow of each city road network unit boundary are distributed to the corresponding specific ramp and boundary control road section to obtain the expected flow of each ramp and boundary control road section. In this allocation process, the following restrictions are met simultaneously: Green light time upper and lower bounds, road section receiving capacity, and total flow minimum and maximum flow constraints in step S3, as well as suppression control directly causing vehicle queue overflow; Step 2: Map the expected flow of each ramp and boundary control road section to the executable signal phase green light duration, and execute it in the next control period; After the control execution is completed, the detection data is re-collected, the state vector is updated, and steps S3-S5 are repeated to form a rolling closed-loop control.

9. The method of claim 1, wherein, The control is executed in a fixed control period, and in each control period, "data collection-traffic state update-feedback control decision-total flow allocation and signal execution" is completed, and the iteration is repeated in the next control period to form a closed-loop control.

Images