Highway network-level collaborative active management and control method and system

Abstract

The invention discloses an expressway road network level collaborative active management and control method and system, and the method comprises the steps: obtaining traffic data collected by a detector, and calculating an expressway region operation state key index; a fuzzy C-means clustering method is collected to determine the operation state of the highway road network and the cooperative control boundary of the regional road network; the method comprises the following steps: acquiring position information of highway network information issuing equipment, constructing a directed graph, and determining a free flow reachable matrix of the directed graph and a plurality of neighborhood matrixes of different levels; calculating a convolution operator of the constructed directed graph at each moment in a set time period; and inputting a convolution operator of the constructed directed graph at each moment in a set time period into the neural network model to obtain a traffic state of a section where each information issuing device is located at each moment in a prediction time period, and determining a final regional road network cooperative control decision means under a regional road network cooperative control boundary condition. According to the invention, road network-level collaborative active management and control for improving the passing efficiency of the highway road network can be realized.

Description

Technical Field

[0002] The present invention relates to the field of intelligent transportation technology, and in particular to a highway network-level collaborative active control method and system. Background Art

[0003] With the development of the economy and the increase in the number of cars, the traffic flow on highways has continued to rise, and the actual traffic flow is far greater than the designed saturation flow, causing highway congestion. After investigating and analyzing a large number of congestion incidents, scholars found that highway congestion follows the "point-line-surface" propagation logic, that is, the congestion in the upstream section will be transmitted to the downstream section and cause congestion in the entire road network. If the traditional means of physical expansion of highways only rely on the increase in the number of physical lanes, the frequency of vehicle lane changes will increase, the traffic interweaving behavior will be serious, and the single-lane capacity will show a marginal decreasing effect. After the induced traffic volume enters the road network in large quantities, the regional road network will return to the dilemma of congestion. Therefore, it is necessary to improve the efficiency of highway traffic through intelligent traffic flow control means.

[0004] Existing smart expansion control methods include speed coordination control, variable speed limit control, temporary opening of emergency lanes, ramp control strategies, etc. Most of these methods are aimed at traffic control in congested sections. The highway network is a whole, and coordinated control is required between sections. We cannot only focus on alleviating the traffic status of congested sections while ignoring the other sections. When congestion occurs in one section, adjacent sections can take measures such as flow limitation and diversion to avoid the spread of congestion. Road conditions information should be shared between sections to achieve unified dispatch and coordinated command, and to achieve a transition from passive control of a single section to active control of the entire road network.

[0005] Therefore, it is necessary to provide a highway network-level active management and control collaborative control technology to improve the traffic efficiency and safety of highways. Summary of the invention

[0006] In order to overcome the problem that traditional highway control often focuses on a single road section or specific area and lacks global consideration of the entire road network, and to achieve network-level collaborative active control to improve the traffic efficiency of the highway network, the present application provides a highway network-level collaborative active control method and system.

[0007] In a first aspect, the present application provides a highway network-level collaborative active control method, comprising: Obtaining the locations of different detectors in the expressway network and the traffic data of the sections where the detectors are collected, and calculating key indicators of the expressway area operation status; the key indicators of the expressway area operation status include the average vehicle speed of the road network, the road network congestion and the road network travel time; Based on the key indicators of the regional operation status of expressways, the fuzzy C-means clustering method is collected to determine the coordinated control boundary between the expressway network operation status and the regional network; Obtain the location information of the highway network information release device, and construct a directed graph based on the device location information, and determine the free flow reachability matrix of the directed graph and multiple neighborhood matrices of different levels based on the free flow speeds between the road sections corresponding to the information release device and the constructed directed graph; According to the neighborhood matrix of all orders, the free flow reachability matrix and the traffic status of the section where each information release device is located at each time in the set time period, the convolution operator of the constructed directed graph at each time in the set time period is calculated; The convolution operator of the constructed directed graph at each moment in the set time period is input into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period; According to the traffic status of each information release device on the road section at each moment in the set time period and the traffic status of each information release device on the road section at each moment in the predicted time period, under the boundary conditions of the regional road network collaborative control, the final regional road network collaborative control decision-making means is determined.

[0008] By adopting the above scheme, the key indicators of the regional operating status of the expressway are calculated using the collected multi-source traffic data, and the fuzzy C-means clustering method is used to obtain the results of the expressway road network operating status, and the regional road network collaborative control boundary is determined to provide a basis for the final regional road network system management and control; considering the different important factors affecting high-speed traffic, the directed graph is constructed, the convolution operator of the directed graph at each moment in the set time period is obtained, and the convolutional long short-term memory neural network model is introduced to accurately predict the traffic status of each section at each moment in the predicted time period, and combined with the regional road network collaborative control boundary, the final regional road network collaborative control decision-making means is obtained to improve traffic efficiency.

[0009] Preferably, the obtaining of the positions of different detectors in the expressway network and the traffic data of the sections where the detectors collect data, and the calculation of the key indicators of the expressway area operation status include: Use different detectors to collect traffic data on the section, including: section flow, section speed between two adjacent gantries, traffic volume, single vehicle speed and average speed; According to the collected traffic data, the key indicators of the highway area operation status are calculated; among them, the average vehicle speed calculation formula of the road network is: Where v is the average vehicle speed of the road network; v i is the average vehicle speed of the i-th section of the regional road network; l i is the mileage of the i-th section of the regional road network; q i is the traffic volume of the i-th section of the regional road network; According to the average speed of the road network, the road network speed threshold range under the preset highway section congestion conditions is compared to determine the highway section congestion, including: unblocked, basically unblocked, light congestion, heavy congestion and severe congestion; the road network congestion is calculated based on the highway section congestion, and the calculation formula is: Among them, F N is the road network congestion; L is the total mileage of the road network; d i is the total mileage of road section i with moderate congestion; w j is the mileage of j in the severely congested section; m is the number of sections in moderate congestion; n is the number of sections in severe congestion; the road network travel time reliability index is calculated as follows: Among them, TTK is the road network travel time reliability index; m is the number of road sections; VKT i is the vehicle-kilometers, which is equal to the number of cars passing through section i during the statistical period multiplied by the length of section i; TBI i is the travel time reliability index of section i, calculated by: Among them, TBI i is the travel time reliability index; t 0.90,i is the 90% percentile travel time of road section i under congestion conditions, that is, the travel time taken by travelers to pass through road section i under 90% probability conditions; is the average travel time required to pass through road section i under congestion conditions; is the average travel time required to pass through road section i under ideal conditions; wherein, the congestion condition can be determined by comparing the calculated road network congestion with the preset road network congestion classification standard to determine whether the congestion classification level is medium or above.

[0010] By adopting the above scheme, different detectors are used to obtain various types of traffic data, so as to calculate and obtain the main indicators of road network operation status such as average vehicle speed, road network congestion and road network travel time based on multi-source data, providing a basis for accurate control of road network operation status.

[0011] Preferably, the method of collecting the fuzzy C-means clustering method based on the key indicators of the highway regional operation status to determine the highway network operation status and the regional network collaborative control boundary includes: The fuzzy C-means clustering method is used to perform cluster analysis on the road network operation status to obtain the highway network operation status results; the highway network operation dynamics are divided into five states: smooth, basically smooth, slightly congested, moderately congested, and severely congested, which correspond to the matching preset indexes A, B, C, D, and E respectively; The regional road network collaborative control boundary is determined based on the preset index matched with the results of the highway network operation status; if the preset index is [D, E], it indicates that the overall operation status of the road network is not good, and the regional road network collaborative control strategy is determined as follows: temporarily opening the emergency lane, the main line variable speed limit strategy, and the entrance ramp control strategy are adopted, but the induced diversion strategy is not adopted; if the preset index is C, it indicates that the overall operation status of the road network is general, and the regional road network collaborative control strategy is determined as follows: main / ramp control strategy and induced diversion strategy are adopted at nodes and roads with remaining traffic capacity; if the preset index is [A, B], it indicates that the overall operation status of the road network is good, and the regional road network collaborative control strategy is determined as follows: in the event of sudden congestion on some sections, an induced diversion strategy is adopted.

[0012] By adopting the above scheme, the regional road network collaborative control strategy is quickly matched and determined according to the preset index corresponding to the highway network operation status results, providing a basis for the final regional road network system management and control.

[0013] Preferably, the method of determining the final regional road network collaborative control decision means according to the traffic status of each intelligence release device at each time in the set time period and the traffic status of each intelligence release device at each time in the predicted time period, under the regional road network collaborative control boundary conditions, includes: Determine the regional road network coordinated control time, including: Assuming that all vehicles can immediately obtain the guidance information at the guidance starting point, we introduce the characterization of the traveler's choice of road section a ij The lowest expected speed tolerance v tol According to the traffic status of each information release device at each time in the set time period and the traffic status of each information release device at each time in the predicted time period, determine the traffic status of each information release device at each time in the predicted time period. 0 During the period, section a ij The queue length does not overflow; t 0 During the +Δt period, section a ij The queue overflows to the upstream section a i-1,i To ensure that the vehicle enters section a i-1,i After receiving the induction information, the vehicle has enough time to leave the congested section before the road is blocked. The length of the induction cycle is defined as follows: Among them, l i-1,i is the length of the road section; is the tolerable speed v tol The order function of ; For section a i-1,i Tolerable travel time at tolerable speed; As the induced cycle interval increases, the total travel time of travelers shows a trend of first decreasing and then increasing; the induced cycle interval is t low When H ≥ t low When, it is generally better than H <t low , then the induction time needs to meet the following conditions: Determine the spatial scope of regional road network coordinated control, including: Determine the induced space distance L space , there exists u such that: but: Pre-control space length L space After the determination, the node where the queuing event occurs is set as the induction key point, and the key point is used as the starting point for classification. The nodes are classified in sequence along the upstream direction. The point set composed of all pre-controlled nodes is N r , N r As the action point of traffic control and the selection point of the starting point of the induced path; Determine the final regional road network collaborative control decision-making means, including: If the regional network collaborative control strategy is determined to be the temporary opening of emergency lanes, main line variable speed limit strategy, and entrance ramp control strategy based on the index matching the highway network operation status results, but the induced diversion strategy is not adopted, then the final regional network collaborative control decision means is to select the ramp or main line active control strategy for separate control; If the regional road network collaborative control strategy is determined to be the main / ramp control strategy and the induced diversion strategy at the nodes and roads with the remaining traffic capacity according to the index matched by the highway network operation status results, then the final regional road network collaborative control decision means is determined to be the selection of ramp or main line active control strategy for separate control or the implementation of the induced diversion strategy according to the determined road network collaborative control time and road network collaborative control space range; If the regional road network collaborative control strategy is determined according to the index matching the highway network operation status results to adopt an induced diversion strategy in the event of sudden congestion on some sections, then the final regional road network collaborative control decision-making means is to implement the induced diversion strategy based on the determined road network collaborative control time and road network collaborative control spatial range.

[0014] By adopting the above scheme, the congested sections are determined according to the traffic status of each section at each time in the set time period and at each time in the predicted time period, and the time range and spatial range of regional road network coordinated control are determined in combination with the travelers' tolerance speed for the selected sections, so as to obtain the action point of traffic control and the selection point of the starting point of the induced path, realize induced diversion, and improve the efficiency of road network traffic operation.

[0015] Preferably, the tolerable speed acquisition includes: The weather conditions and holiday information of the regional road sections are obtained according to the highway network intelligence release equipment, and the preset tolerance speed is obtained based on the matching of the weather conditions and holiday information of the regional road sections; wherein, different tolerance speeds are preset for different weather conditions and holidays, and the preset tolerance speed corresponding to each combination of weather conditions and holidays is determined according to the average tolerance speed of travelers under the same historical combination of weather conditions and holidays.

[0016] By adopting the above scheme, taking into account the impact of different weather conditions and the tolerance speed of travelers during holidays, the time range and spatial range of regional road network coordinated control are determined to adapt to the current weather conditions and the tolerance speed during holidays.

[0017] Preferably, obtaining the road network vehicle speed threshold range under the preset highway section congestion condition includes: The weather conditions and holiday information of the regional road network are obtained according to the highway network intelligence release equipment, and the road network speed threshold range under the preset highway section congestion conditions is obtained based on the matching of the weather conditions and holiday information of the regional road network; among them, different weather conditions and holidays correspond to different highway section congestion conditions, and each combination of weather conditions and holidays corresponds to the road network speed threshold range under the preset highway section congestion conditions, which is determined according to the average speed of the road network of travelers under the historical combination of the same weather conditions and holidays and different speed limit conditions of the highway.

[0018] By adopting the above scheme, taking into account the impact of different weather conditions and the average driving speed of travelers during holidays, combined with the speed limit standards of various highways, the road network speed threshold range is determined to adapt to different weather conditions and holidays and different congestion conditions of highway sections, thereby completing a more accurate congestion level classification standard and achieving accurate calculation of key indicators of highway regional operating status.

[0019] Preferably, the convolution operator of the constructed directed graph at each moment in the set time period is input into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period using the DGC-GRU neural network model. The input of the DGC-GRU neural network model is the convolution operator of the constructed directed graph at each moment in the set time period, and the output of the model is the traffic status of each road section in the measured time period.

[0020] By adopting the above solution and using the DGC-GRU neural network model that takes into account the spatiotemporal characteristics of traffic flow, the complex spatial dependencies and dynamic temporal dependencies in traffic data can be extracted to achieve more accurate predictions.

[0021] In a second aspect, the present application provides a highway network-level collaborative active control system, including: The highway area operation status indicator acquisition module is used to obtain the positions of different detectors in the highway network and the traffic data of the sections where the detectors collect, and calculate the key indicators of the highway area operation status; the key indicators of the highway area operation status include the average vehicle speed of the road network, the road network congestion and the road network travel time; The highway network operation status determination module is used to determine the highway network operation status and the regional network collaborative control boundary based on the key indicators of the highway regional operation status and the fuzzy C-means clustering method; The highway network directed graph and parameter acquisition module is used to obtain the location information of the highway network intelligence release equipment, and construct a directed graph based on the equipment location information, and determine the free flow reachability matrix of the directed graph and multiple neighborhood matrices of different levels based on the free flow speed between the sections corresponding to the intelligence release equipment and the constructed directed graph; The highway network directed graph convolution operator acquisition module is used to calculate the convolution operator of the constructed directed graph at each moment in the set time period according to the neighborhood matrix of all orders, the free flow reachability matrix and the traffic status of the section where each information release device is located at each moment in the set time period; The highway network traffic status acquisition module is used to input the convolution operator of the constructed directed graph at each moment in the set time period into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period; The highway network collaborative control module is used to determine the final regional road network collaborative control decision-making means under the regional road network collaborative control boundary conditions based on the traffic status of each information release device on the road section at each moment within the set time period and the traffic status of each information release device on the road section at each moment within the predicted time period.

[0022] By adopting the above scheme, various factors affecting highway congestion are taken into account to calculate the road network operation status, and by introducing the dilated causal convolution module, long-term dependencies are captured to achieve more accurate prediction of highway traffic congestion, and then determine the regional road network collaborative control decision-making means, effectively improving the highway network traffic efficiency.

[0023] In a third aspect, the present application provides a computer-readable storage medium, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method as described above.

[0024] In a fourth aspect, the present application provides a computer device, comprising a memory, a processor, and a program stored and executable on the memory, wherein the program implements the steps of the above method when executed by the processor.

[0025] In summary, this application has the following beneficial effects: 1. Taking into account various types of traffic data collected, the key indicators of the regional operation status of highways are calculated based on multi-source data, and the fuzzy C-means clustering method is used to perform cluster analysis on the road network operation status to obtain more accurate results of the highway network operation status, and the regional road network collaborative control boundary is determined to provide a restriction basis; considering factors such as the traffic conditions, traffic flow distribution and road capacity of each section, a directed graph is constructed based on the location information of the intelligence release equipment, and the free flow reachability matrix and neighborhood matrix are determined. The convolutional long short-term memory neural network model is introduced to accurately predict the traffic status of each section at each time within the prediction time period, realize the active control of the regional road network, obtain the final regional road network collaborative control decision-making means, and improve the traffic efficiency of the highway network; 2. Combined with the upstream and downstream relationships between the sections where each intelligence release device is located and the congestion source, accurately identify the propagation direction and impact range of traffic congestion, design different active control strategies according to different situations, determine the control measures that should be taken for each section in the road network, and realize refined road network-level large-volume traffic control and coordinated induction. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 is a flow chart of the method described in a specific embodiment; Figure 2 A schematic diagram of a convolution operator input into a neural network model in the method described in a specific embodiment; Figure 3 A schematic diagram of the traffic status of a vehicle at each time on a road network section in the method described in a specific embodiment; Figure 4 It is a schematic diagram of executing the induction strategy at each time and spatial range of the road network section in the method described in the specific embodiment; Figure 5 A schematic diagram of a road network collaborative control decision-making means in the method described in a specific embodiment; Figure 6 It is a schematic diagram of the structure of the system described in the specific embodiment. DETAILED DESCRIPTION

[0027] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0028] like Figure 1 As shown, the embodiment of the present application discloses a method for highway network-level collaborative active control, and the specific steps include: S1. Obtain the locations of different detectors in the highway network and the traffic data of the sections where the detectors collect data, and calculate the key indicators of the highway area operation status.

[0029] Specifically, traffic data is collected through a variety of detectors installed at different locations on the highway, including ETC gantries, radars, and videos. The traffic data provided by the ETC gantries include cross-sectional flow and the speed between two adjacent gantries; the traffic data provided by the radars include traffic flow and single-vehicle speed; and the traffic data provided by the videos include cross-sectional flow, average speed and other traffic parameters.

[0030] Based on the traffic data collected by the detector, the key indicators of the highway area operation status are calculated. The key indicators of the highway area operation status specifically include: average vehicle speed v of the road network, road network congestion (road section congestion, road network congestion F N ) and network travel time (link travel time reliability index TBI i , road network travel time reliability index TTK) The calculation formula for the average vehicle speed of the road network is: Where v is the average vehicle speed of the road network; v i is the average vehicle speed of the ith section of the regional road network (km / h); l i is the mileage of the i-th section of the regional road network (km); q i is the traffic flow of the i-th section of the regional road network (veh / h).

[0031] According to the average speed of the road network, the road network speed threshold range under the preset highway section congestion condition is compared to determine the highway section congestion. Specifically, considering the different weather conditions (such as sunny days, rainy days and rainstorms, etc.) and holidays (holidays and non-holidays), the number of traffic is different. In order to ensure traffic safety, the road network speed threshold range under the preset highway section congestion condition is preset. Specifically, the combination of each weather condition and holiday condition corresponds to the road network speed threshold range under the preset highway section congestion condition, which is determined according to the combination of the same weather conditions and holidays in history and the average speed of the traveler road network under different speed limit conditions of the highway. Among them, the weather conditions of the regional road network, holiday information and speed limit information of each highway section can be collected and obtained according to the highway network intelligence release equipment. Combined with the speed limit information, the road network speed threshold range under the preset highway section congestion condition is obtained based on the weather conditions of the regional road network and the holiday information matching.

[0032] In this embodiment, the congestion level of highway sections is divided into five levels, including: unimpeded, basically unimpeded, light congestion, heavy congestion and severe congestion. The lower the congestion level, the higher the average travel speed of the section. The specific standards for the classification of highway section congestion levels are shown in Table 1 below.

[0033] Table 1 Classification criteria for highway section congestion levels (data in the table are speed data) The road network congestion is calculated based on the determined highway section congestion. The calculation formula is: Among them, F N is the road network congestion; L is the total mileage of the road network; d i is the total mileage of road section i with moderate congestion; w j is the mileage of j in the severely congested section; m is the number of sections in moderate congestion; n is the number of sections in severe congestion.

[0034] According to the road network congestion, the road network congestion classification level of the highway is determined by comparing it with the preset road network congestion classification standards as shown in Table 2 below.

[0035] Table 2 Classification of road network congestion The congestion conditions in the subsequent calculation parameters can be judged from Table 2.

[0036] The section travel reliability index is calculated based on the determined highway section congestion, and the calculation formula is: Among them, TBI i is the travel time reliability index; t 0.90,i is the 90% percentile travel time of road section i under congestion conditions (for medium, high, and relatively high congestion levels according to the classification of road network congestion), that is, the travel time spent by travelers through road section i under 90% probability conditions; is the average travel time required to pass through road section i under congestion conditions; is the average travel time required to pass through road section i under ideal conditions.

[0037] Therefore, the road network travel time reliability index is calculated as follows: Among them, TTK is the road network travel time reliability index; m is the number of road sections; VKT i is the vehicle-kilometers, which is equal to the number of cars passing through section i during the preset statistical time multiplied by the length of section i.

[0038] S2. Collect fuzzy C-means clustering method to determine the operation status of the highway network and the coordinated control boundary of the regional network.

[0039] Specifically, based on the key indicators of the regional operation status of expressways, the fuzzy C-means clustering method is used to perform cluster analysis on the road network operation status, and finally the results of the expressway road network operation status are obtained. The road network operation dynamics are divided into five states: unblocked, basically unblocked, lightly congested, moderately congested, and severely congested, which correspond to the matching preset indexes A=1, B=2, C=3, D=4, and E=5 respectively. The specific results are shown in Table 3.

[0040] Table 3 Road network operation status judgment table According to the preset index matching the highway network operation status results, the regional network collaborative control boundary is determined, as shown in Table 4.

[0041] Table 4 Regional road network collaborative control boundaries Specifically, if the matching preset index is [4,5], it indicates that the overall operation state of the road network is not good, it is relatively congested, and most roads are also relatively congested, which means that there is no extra space in the road network for diversion. The corresponding regional road network collaborative control strategy is: choose to adopt technologies such as temporary opening of emergency lanes, main line variable speed limit strategy, entrance ramp control strategy, but do not adopt induced diversion strategy; if the matching preset index is 3, it indicates that the overall operation state of the road network is general, and the regional road network collaborative control strategy is determined as: adopt main / ramp control strategy and induced diversion strategy at nodes and roads with remaining traffic capacity; if the matching preset index is [1,2], it indicates that the overall operation state of the road network is good. Under normal circumstances, there is no need for induced diversion of the road network. Generally, in the case of emergencies such as traffic accidents, there is a need for induced diversion. The regional road network collaborative control strategy is determined as: in the case of sudden congestion on some sections, induced diversion strategy is adopted.

[0042] S3. Obtain the location information of the highway network intelligence release device to construct a directed graph, and determine the free flow reachability matrix and neighborhood matrix of the directed graph.

[0043] Specifically, a number of highway network information release devices are installed on the highway to release real-time traffic information, road conditions, weather, warnings and other information. Based on the location information of the highway network information release devices, a highway network directed graph G = (V, E) is constructed, where E is the set of roads; V = {V 1 ,V 2 ,…,V N} is the section where the information release equipment is located in the highway network, that is, there are n nodes in the network.

[0044] Determine the plurality of neighborhood matrices A of different orders according to the directed graph, A∈R N×N ; The neighborhood matrix A represents the n×n dimensional matrix A composed of the relationships between the nodes. The element a in the matrix i,j Represents two sections a i and a j The shortest path distance (a i，i =0).

[0045] The free flow reachable matrix HF is calculated according to the shortest distance matrix of the directed graph and the free flow speed between the sections where each detector is located. FR∈ R N×N , the elements in the matrix as follows: If the motor vehicle passes through section d i Start at free flow speed S and drive to section d j , the time does not exceed the specified time Then the elements in the matrix otherwise

[0046] In this embodiment, a K-order directed graph convolution operator is constructed and calculated.

[0047] S4. According to the neighborhood matrices of all orders, the free flow reachability matrix and the traffic status of the sections where each information release device is located at each moment in the set time period, the convolution operator of the constructed directed graph at each moment in the set time period is calculated.

[0048] Specifically, the traffic status of the section where each information release device is located at each moment in a set time period includes: congestion and queue information, etc.

[0049] The directed graph convolution operator is calculated according to the following formula Extract spatiotemporal information of transportation networks.

[0050] in, W represents the convolution operator of the directed graph when the order is k at time t; dgc_k ∈R N×1 represents the weight matrix of the neighborhood matrix of order k; represents the neighborhood matrix of order k, H FFR ∈R N×N represents the free flow reachability matrix; X t ∈R N×1 It represents the traffic status of the section where each detector is located at time t; ⊙ is the multiplication of the elements at the corresponding positions of the matrix; · is the matrix multiplication operator.

[0051] The K-order directed graph convolution operator is calculated according to the following formula: S5. Input the convolution operator of the constructed directed graph at each moment in the set time period into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period.

[0052] Specifically, a directed graph convolutional long short-term memory neural network model is constructed to match the real highway network to learn the complex spatial dependencies and dynamic temporal dependencies in traffic data.

[0053] like Figure 2 As shown, the directed graph convolution operator Combined with the GRU neural network model, a DGC-GRU model that takes into account the spatiotemporal characteristics of traffic flow is constructed; specifically, the input signal of the DGC-GRU neural network model at time t is from the sequence X t Transformed into a matrix containing the spatiotemporal information of the road network The output is the traffic status of each road section during the measurement period.

[0054] Specifically, the update gate, reset gate, candidate hidden state and hidden state structures of the DGC-GRU model are shown in the following formula: Among them, W z , W r , W∈R N×KN is the weight matrix of the input signal; U z , U r , U∈R N×N is the weight matrix of the previous state information; b z , b r , b∈R N×1 are three deviation vectors. σ and tanh are both activation functions, which are used to add some nonlinear factors to the neural network to adapt to the more complex nonlinear characteristics of traffic data; σ is a sigmoid function.

[0055] As shown in the following formula, calculate the updated output y at time t t : y t =σ(W 0 ·h t +b 0 ) Among them, W 0 is the weight matrix of the output layer; b 0 is the output layer coefficient vector; σ is the sigmoid activation function, and the output mapping range is (0, 1), thereby obtaining the traffic status of the section where each information release device is located at each moment in the predicted time period.

[0056] S6. According to the traffic status of each information release device on the road section at each moment in the set time period and the traffic status of each information release device on the road section at each moment in the predicted time period, under the boundary conditions of the regional road network collaborative control, determine the final regional road network collaborative control decision means.

[0057] Taking into account the different road network operating conditions, different regional road network coordinated control decisions are taken accordingly. The regional road network coordinated control decisions designed for different situations are combined with the traffic conditions of each section within the set time period and the preset time period, and refined road network-level large-volume traffic control coordinated induction. The specific steps include: First, determine the regional road network coordinated control time, that is, determine when to carry out coordinated induction, including: Assuming that all vehicles can immediately obtain the guidance information at the guidance starting point, we introduce the characterization of the traveler's choice of road section a ij The lowest expected speed tolerance v tol ; Taking into account that different weather conditions and holidays will affect travelers' tolerance for road section travel speeds, such as people will have a higher tolerance for communication speeds in sunny weather and on holidays; therefore, the weather conditions and holiday information of regional sections are obtained according to the highway network intelligence release equipment, and the preset tolerance speed is obtained based on the matching of the obtained weather conditions and holiday information of the regional sections; different tolerance speeds are preset for different weather conditions and holidays, and the preset tolerance speed for each combination of weather conditions and holidays is determined according to the average tolerance speed of travelers under the same historical combination of weather conditions and holidays.

[0058] like Figure 3 As shown, according to the traffic status of each information release device at each time in the set time period and the traffic status of each information release device at each time in the predicted time period, the traffic flow at t 0 During the period, section a ij The queue length does not overflow; t 0 During the +Δt period, section a ij The queue overflows to the upstream section a i-1，i To ensure that the vehicle enters section a i-1，i After receiving the induction information, vehicles have enough time to leave the congested section before the road is blocked. The length of the induction cycle is defined so that the induction cycle interval is synchronized with the traffic volume prediction interval.

[0059] Among them, l i-1，i is the length of the road section; is the tolerable speed v tol The order function of ; For section a i-1，i Tolerable travel time at tolerable speed; As the induced cycle interval increases, the total travel time of travelers shows a trend of first decreasing and then increasing; the induced cycle interval is t low When H ≥ t low When the overall performance is better than H<t low , then the induction time needs to meet the following conditions: Secondly, on the basis of determining the regional road network coordinated control time, determine the regional road network coordinated control space scope, including: Determine the induced space distance L space , there exists u such that: but: Pre-control space length L space After confirmation, if Figure 4 As shown in the figure, the node where the queuing event occurs, that is, the section where the congestion occurs, is set as the induction key point. The key point is used as the starting point for classification, and the nodes are classified in sequence along the upstream direction. The point set composed of all pre-controlled nodes is N r , until the pre-control space length is reached, N r As the action point of traffic control and the selection point of the starting point of the induced path, different levels of pre-control can be set for different nodes.

[0060] Finally, under the boundary conditions of regional road network collaborative control, the final regional road network collaborative control decision means are determined, such as Figure 5 As shown, including: If the regional road network collaborative control strategy is determined to be the temporary opening of emergency lanes, main line variable speed limit strategy, entrance ramp control strategy, but the induced diversion strategy is not adopted according to the index matching the results of the highway network operation status, then the final regional road network collaborative control decision means is to select the ramp or main line active control strategy for separate control. After a preset period of time, it is re-judged whether the road section with the current congestion condition is still in a congested state. If it is still, it is determined that the control effect is not good, and the ramp + main line active control strategy is selected for collaborative control. If the regional road network collaborative control strategy is determined to be the main / ramp control strategy and the induced diversion strategy at the nodes and roads with the remaining traffic capacity according to the index matched by the highway network operation status results, then the final regional road network collaborative control decision means is determined to be the selection of ramp or main line active control strategy for separate control or the implementation of the induced diversion strategy according to the determined road network collaborative control time and road network collaborative control space range; If the regional road network collaborative control strategy is determined according to the index matching the highway network operation status results to adopt an induced diversion strategy in the event of sudden congestion on some sections, then the final regional road network collaborative control decision-making means is to implement the induced diversion strategy based on the determined road network collaborative control time and road network collaborative control spatial range.

[0061] like Figure 6 As shown, the embodiment of the present application discloses a highway network-level collaborative active control system, which specifically includes: a highway area operation status indicator acquisition module 101, which is used to obtain the positions of different detectors in the highway network and the traffic data of the section where the detectors collect, and calculate the key indicators of the highway area operation status; the key indicators of the highway area operation status include the average vehicle speed of the road network, the road network congestion and the road network travel time; The highway network operation status determination module 102 is used to collect fuzzy C-means clustering method based on the key indicators of the highway regional operation status to determine the highway network operation status and the regional network collaborative control boundary; The highway network directed graph and parameter acquisition module 103 is used to obtain the location information of the highway network information release device, and construct a directed graph according to the device location information, and determine the free flow reachability matrix of the directed graph and multiple neighborhood matrices of different levels according to the free flow speed between the sections corresponding to the information release device and the constructed directed graph; The highway network directed graph convolution operator acquisition module 104 is used to calculate the convolution operator of the constructed directed graph at each time in the set time period according to the neighborhood matrix of all orders, the free flow reachability matrix and the traffic status of the section where each information release device is located at each time in the set time period; The highway network traffic status acquisition module 105 is used to input the convolution operator of the constructed directed graph at each moment in the set time period into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period; The highway network collaborative control module 106 is used to determine the final regional road network collaborative control decision-making means under the regional road network collaborative control boundary conditions based on the traffic status of each information release device on the road section at each moment within the set time period and the traffic status of each information release device on the road section at each moment within the predicted time period.

[0062] The embodiment of the present application also discloses a computer-readable storage medium.

[0063] Specifically, the computer-readable storage medium stores a computer program that can be loaded by a processor and executes the above-mentioned highway network-level collaborative active control method. The computer-readable storage medium includes, for example: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and other media that can store program codes.

[0064] The embodiment of the present application also discloses a computer device.

[0065] Specifically, the computer device includes a memory and a processor, and the memory stores a computer program that can be loaded by the processor and execute the above-mentioned highway network-level collaborative active control method.

[0066] The above are all preferred embodiments of the present application, and are not intended to limit the protection scope of the present application. Any feature disclosed in this specification (including the abstract and drawings), unless otherwise stated, can be replaced by other equivalent or alternative features with similar purposes. That is, unless otherwise stated, each feature is only an example of a series of equivalent or similar features.

Claims

1. A highway network-level collaborative active control method, characterized in that: include: Obtaining the locations of different detectors in the expressway network and the traffic data of the sections where the detectors are collected, and calculating key indicators of the expressway area operation status; the key indicators of the expressway area operation status include the average vehicle speed of the road network, the road network congestion and the road network travel time; Based on the key indicators of the regional operation status of expressways, the fuzzy C-means clustering method is collected to determine the coordinated control boundary between the expressway network operation status and the regional network; Obtain the location information of the highway network information release device, and construct a directed graph based on the device location information, and determine the free flow reachability matrix of the directed graph and multiple neighborhood matrices of different levels based on the free flow speeds between the road sections corresponding to the information release device and the constructed directed graph; According to the neighborhood matrix of all orders, the free flow reachability matrix and the traffic status of the section where each information release device is located at each time in the set time period, the convolution operator of the constructed directed graph at each time in the set time period is calculated; The convolution operator of the constructed directed graph at each moment in the set time period is input into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period; According to the traffic status of each information release device on the road section at each moment in the set time period and the traffic status of each information release device on the road section at each moment in the predicted time period, under the boundary conditions of the regional road network collaborative control, the final regional road network collaborative control decision-making means is determined.

2. The highway network-level collaborative active control method according to claim 1 is characterized in that: The acquisition of the locations of different detectors in the expressway network and the traffic data of the sections where the detectors collect data, and the calculation of key indicators of the expressway area operation status include: Use different detectors to collect traffic data on the section, including: section flow, section speed between two adjacent gantries, traffic volume, single vehicle speed and average speed; According to the collected traffic data, the key indicators of the highway area operation status are calculated; among them, the average vehicle speed calculation formula of the road network is: Where v is the average vehicle speed of the road network; v i is the average vehicle speed of the i-th section of the regional road network; l i is the mileage of the i-th section of the regional road network; q i is the traffic volume of the i-th section of the regional road network; According to the average speed of the road network, the road network speed threshold range under the preset highway section congestion conditions is compared to determine the highway section congestion, including: unblocked, basically unblocked, light congestion, heavy congestion and severe congestion; the road network congestion is calculated based on the highway section congestion, and the calculation formula is: Among them, F N is the road network congestion; L is the total mileage of the road network; d i is the total mileage of road section i with moderate congestion; w j is the mileage of j in the severely congested section; m is the number of sections in moderate congestion; n is the number of sections in severe congestion; Calculate the road network travel time reliability index, the formula is: Among them, TTK is the road network travel time reliability index; m is the number of road sections; VKT i is the vehicle-kilometers, which is equal to the number of cars passing through section i during the statistical period multiplied by the length of section i; TBI i is the travel time reliability index of section i, calculated by: Among them, TBI i is the travel time reliability index; t 0.90，i is the 90% percentile travel time of road section i under congestion conditions, that is, the travel time taken by travelers to pass through road section i under 90% probability conditions; is the average travel time required to pass through road section i under congestion conditions; is the average travel time required to pass through road section i under ideal conditions; wherein, the congestion condition can be determined by comparing the calculated road network congestion with the preset road network congestion classification standard to determine whether the congestion classification level is medium or above.

3. The highway network-level collaborative active control method according to claim 1 is characterized in that: The method of collecting fuzzy C-means clustering method based on the key indicators of the highway regional operation status to determine the highway network operation status and the regional network collaborative control boundary includes: The fuzzy C-means clustering method is used to perform cluster analysis on the road network operation status to obtain the highway network operation status results; the highway network operation dynamics are divided into five states: smooth, basically smooth, slightly congested, moderately congested, and severely congested, which correspond to the matching preset indexes A, B, C, D, and E respectively; The regional road network collaborative control boundary is determined based on the preset index matched with the results of the highway network operation status; if the preset index is matched between [D, E], it indicates that the overall operation status of the road network is not good, and the regional road network collaborative control strategy is determined as follows: temporarily opening the emergency lane, the main line variable speed limit strategy, and the entrance ramp control strategy are adopted, but the induced diversion strategy is not adopted; if the preset index is matched between C, it indicates that the overall operation status of the road network is general, and the regional road network collaborative control strategy is determined as follows: main / ramp control strategy and induced diversion strategy are adopted at nodes and roads with remaining traffic capacity; if the preset index is matched between [A, B], it indicates that the overall operation status of the road network is good, and the regional road network collaborative control strategy is determined as follows: in the event of sudden congestion on some sections, an induced diversion strategy is adopted.

4. The highway network-level collaborative active control method according to claim 3 is characterized in that: The method of determining the final regional road network collaborative control decision means according to the traffic status of each information release device at each time in the set time period and the traffic status of each information release device at each time in the predicted time period under the regional road network collaborative control boundary conditions includes: Determine the regional road network coordinated control time, including: Assuming that all vehicles can immediately obtain the guidance information at the guidance starting point, we introduce the characterization of the traveler's choice of road section a ij The lowest expected speed tolerance v tol According to the traffic status of each information release device at each time in the set time period and the traffic status of each information release device at each time in the predicted time period, it is determined that within the t0 period, section a ij The queue length of section a does not overflow; during the period t0+Δt, ij The queue overflows to the upstream section a i-1，i To ensure that the vehicle enters section a i-1，i After receiving the induction information, the vehicle has enough time to leave the congested section before the road is blocked. The length of the induction cycle is defined as follows: Among them, l i-1，i is the length of the road section; is the tolerable speed v tol The order function of ; For section a i-1，i Tolerable travel time at tolerable speed; As the induced cycle interval increases, the total travel time of travelers shows a trend of first decreasing and then increasing; the induced cycle interval is t low When H ≥ t low When the overall performance is better than H<t low , then the induction time needs to meet the following conditions: Determine the spatial scope of regional road network coordinated control, including: Determine the induced space distance L space , there exists u such that: but: Pre-control space length L space After the determination, the node where the queuing event occurs is set as the induction key point, and the key point is used as the starting point for classification. The nodes are classified in sequence along the upstream direction. The point set composed of all pre-controlled nodes is N r , N r As the action point of traffic control and the selection point of the starting point of the induced path; Determine the final regional road network collaborative control decision-making means, including: If the regional network collaborative control strategy is determined to be the temporary opening of emergency lanes, main line variable speed limit strategy, and entrance ramp control strategy based on the index matching the highway network operation status results, but the induced diversion strategy is not adopted, then the final regional network collaborative control decision means is to select the ramp or main line active control strategy for separate control; If the regional road network collaborative control strategy is determined to be the main / ramp control strategy and the induced diversion strategy at the nodes and roads with the remaining traffic capacity according to the index matched by the highway network operation status results, then the final regional road network collaborative control decision means is determined to be the selection of ramp or main line active control strategy for separate control or the implementation of the induced diversion strategy according to the determined road network collaborative control time and road network collaborative control space range; If the regional road network collaborative control strategy is determined according to the index matching the highway network operation status results to adopt an induced diversion strategy in the event of sudden congestion on some sections, then the final regional road network collaborative control decision-making means is to implement the induced diversion strategy based on the determined road network collaborative control time and road network collaborative control spatial range.

5. The highway network-level collaborative active control method according to claim 4 is characterized in that: The tolerable speed acquisition includes: The weather conditions and holiday information of the regional road sections are obtained according to the highway network intelligence release equipment, and the preset tolerance speed is obtained based on the matching of the weather conditions and holiday information of the regional road sections; wherein, different tolerance speeds are preset for different weather conditions and holidays, and the preset tolerance speed corresponding to each combination of weather conditions and holidays is determined according to the average tolerance speed of travelers under the same historical combination of weather conditions and holidays.

6. The highway network-level collaborative active control method according to claim 2 is characterized in that: The acquisition of the road network vehicle speed threshold range under the preset highway section congestion condition includes: The weather conditions and holiday information of the regional road network are obtained according to the highway network intelligence release equipment, and the road network speed threshold range under the preset highway section congestion conditions is obtained based on the matching of the weather conditions and holiday information of the regional road network; among them, different weather conditions and holidays correspond to different highway section congestion conditions, and each combination of weather conditions and holidays corresponds to the road network speed threshold range under the preset highway section congestion conditions, which is determined according to the average speed of the road network of travelers under the historical combination of the same weather conditions and holidays and different speed limit conditions of the highway.

7. The highway network-level collaborative active control method according to claim 1 is characterized in that: The convolution operator of the constructed directed graph at each moment in the set time period is input into the neural network model, and the traffic status of the section where each information release device is located at each moment in the predicted time period is obtained using the DGC-GRU neural network model. The input of the DGC-GRU neural network model is the convolution operator of the constructed directed graph at each moment in the set time period, and the output of the model is the traffic status of each road section in the measured time period.

8. A highway network-level collaborative active control system, characterized in that: include: The highway area operation status indicator acquisition module is used to obtain the positions of different detectors in the highway network and the traffic data of the sections where the detectors collect, and calculate the key indicators of the highway area operation status; the key indicators of the highway area operation status include the average vehicle speed of the road network, the road network congestion and the road network travel time; The highway network operation status determination module is used to determine the highway network operation status and the regional network collaborative control boundary based on the key indicators of the highway regional operation status and the fuzzy C-means clustering method; The highway network directed graph and parameter acquisition module is used to obtain the location information of the highway network intelligence release equipment, and construct a directed graph based on the equipment location information, and determine the free flow reachability matrix of the directed graph and multiple neighborhood matrices of different levels based on the free flow speed between the sections corresponding to the intelligence release equipment and the constructed directed graph; The highway network directed graph convolution operator acquisition module is used to calculate the convolution operator of the constructed directed graph at each moment in the set time period according to the neighborhood matrix of all orders, the free flow reachability matrix and the traffic status of the section where each information release device is located at each moment in the set time period; The highway network traffic status acquisition module is used to input the convolution operator of the constructed directed graph at each moment in the set time period into the neural network model to obtain the traffic status of the section where each information release device is located at each moment in the predicted time period; The highway network collaborative control module is used to determine the final regional road network collaborative control decision-making means under the regional road network collaborative control boundary conditions based on the traffic status of each information release device on the road section at each moment within the set time period and the traffic status of each information release device on the road section at each moment within the predicted time period.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method according to any one of claims 1 to 7.

10. A computer device, characterized in that: The computer device comprises a memory, a processor and a program stored and executable on the memory, and the program implements the steps of the method according to any one of claims 1 to 7 when executed by the processor.

Images