Traffic early warning management and control method, system and device for expressway

Abstract

The application discloses a traffic early warning management and control method for an expressway, comprising the following steps: based on multi-source traffic perception data of an expressway main line section, calculating the traffic flow parameters of the main line section, fusing the traffic flow parameters with structural defect parameters of the main line section, and obtaining comprehensive operation state parameters of the main line section; the structural defect parameters represent geometric structure defects of the expressway that exist in design and affect traffic fluency; inputting a time sequence of the comprehensive operation state parameters into a prediction model, obtaining traffic stability features representing the smoothness degree of traffic flow operation according to index prediction values; determining the early warning level corresponding to the main line section based on at least the traffic stability features; and based on the early warning level, performing dynamic management and control on target objects associated with the main line section. The technical effects of fine perception, intelligent early warning and differentiated linkage management and control of the traffic flow of the expressway are achieved, and the traffic efficiency and operation stability of the bottleneck section of the main line of the expressway are effectively improved.

Description

Technical Field

[0001] This application relates to the field of traffic control, specifically to a traffic early warning and control method, system, and equipment for expressways. Background Technology

[0002] The main line of urban expressways serves as the main artery for alleviating long-distance, rapid commuting traffic in the city. The lane change sections, weaving areas, entrances and exits, and interchange ramps on the main line are key bottlenecks in the main road network. They are also the core causes of congestion transmission and paralysis of the main road network during morning and evening rush hours, accidents, road construction, and other emergencies.

[0003] At present, the management and control of urban expressways in China suffers from the following defects: lack of operational analysis dimensions, poor universality of early warning models, weak adaptability of management and control strategies, and one-sided evaluation system. Summary of the Invention

[0004] In view of this, the embodiments of this application aim to provide a traffic early warning and control method, system and device for expressways, so as to at least solve the above-mentioned technical problems.

[0005] To achieve the above objectives, the technical solution of this application is implemented as follows: According to one aspect of the embodiments of this application, a traffic early warning and control method for expressways is provided, the method comprising: Based on multi-source traffic perception data of the mainline section of the expressway, the traffic flow parameters of the mainline section are calculated, and the traffic flow parameters include at least traffic flow density, traffic flow volume, and average speed. The traffic flow parameters are fused with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section; the structural defect parameters characterize the geometric structural defects that exist in the design of the expressway and affect traffic flow. The time series of the comprehensive operating state parameters is input into the prediction model, and traffic flow stability features characterizing the smoothness of traffic flow operation are obtained based on the prediction data. The warning level corresponding to the mainline section is determined at least based on the traffic flow stability characteristics. Based on the warning level, dynamic control is performed on the target objects associated with the mainline section. The target objects include at least one of the lanes and entrance ramps associated with the mainline section.

[0006] The method in the above scheme further includes: The effectiveness of the dynamic control is evaluated using multi-dimensional indicators, including at least one of the following: average road segment operating speed, traffic saturation, vehicle speed dispersion coefficient, number of traffic conflicts, and overall control operating cost. Based on the evaluation results, feedback optimization is performed on the warning level and / or the relevant parameters of the dynamic control.

[0007] In the above scheme, the multi-source traffic sensing data includes at least two of the following: checkpoint vehicle passage data, Internet traffic condition data, and mobile phone signaling data; The calculation of traffic flow parameters for the mainline section based on multi-source traffic sensing data from the expressway mainline cross-section includes: The number of vehicles passing through the main line section is counted based on the vehicle passage data at the checkpoint to obtain the traffic flow rate; Based on the internet traffic data and / or the checkpoint vehicle data, the average speed of the main line section is obtained; The traffic density is calculated based on the ratio of the traffic flow volume to the average speed.

[0008] In the above scheme, the step of fusing the traffic flow parameters with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section includes: The traffic flow parameters and the structural defect parameters are processed to obtain processed parameters, and each parameter in the processed parameters is mapped to the same numerical range. The processing parameters are weighted and summed to obtain the comprehensive operating status parameters.

[0009] In the above scheme, the structural defect parameters are obtained in the following way: Extract structural feature data of the location of the mainline section from the road topology profile of the expressway; Calculate the initial defect value based on the degree of deviation between the structural feature data and the preset standard value; The initial defect values ​​are normalized to obtain the structural defect parameters.

[0010] In the above scheme, the step of inputting the time series of the comprehensive operating state parameters into the prediction model and obtaining traffic flow stability characteristics characterizing the smoothness of traffic flow operation based on the prediction data includes: Based on a preset time window, a time sequence of comprehensive operating status parameters of the main section within the time window is constructed; The time series is input into the prediction model to extract the time series feature vector; The time-series feature vector is mapped to a preset interval to obtain traffic flow stability features that characterize the smoothness of traffic flow operation.

[0011] In the above scheme, determining the warning level corresponding to the mainline section based at least on the traffic flow stability characteristics includes: Calculate the current vehicle speed deviation based on the historical average speed of the main line section; The vehicle speed deviation is fused with the traffic flow stability characteristics to generate a comprehensive early warning index; The comprehensive early warning index and / or the traffic flow stability features are input into a trained multi-classifier to obtain the probability distribution of multiple early warning levels. The level corresponding to the highest probability in the probability distribution is taken as the warning level.

[0012] In the above scheme, the step of performing dynamic control on the target objects associated with the main line section based on the early warning level includes: The warning levels are progressively divided into Level 1, Level 2, Level 3 and Level 4, with Level 1 being the most severe. If the warning level is Level 4, no flow control will be implemented on the entrance ramp, and / or no speed limit control will be implemented on the lane; If the warning level is the third level, a first-level flow restriction control will be implemented on the entrance ramp, and / or a first-level speed limit control will be implemented on the lane; If the warning level is the second level, a second degree of flow restriction control is implemented on the entrance ramp, where the intensity of the first degree of flow restriction is greater than that of the second level; and / or a second level of speed limit control is implemented on the lane, where the reduction in the second level speed limit is greater than that of the first level speed limit; If the warning level is Level 1, flow interruption or extreme flow restriction control is implemented on the entrance ramp, and / or Level 3 speed limit control is implemented on the lane; the reduction range of the Level 3 speed limit is greater than that of the Level 2 speed limit.

[0013] In the above scheme, the step of optimizing the early warning level and / or the relevant parameters of the dynamic control based on the evaluation results includes: Based on the evaluation results, the diversion adjustment coefficients and / or weighted fusion weight coefficients of the traffic loss stabilization critical threshold and early warning level are reversed.

[0014] According to a second aspect of this application, a traffic early warning and control system for expressways is provided, the system comprising: The calculation unit is used to calculate the traffic flow parameters of the mainline section of the expressway based on multi-source traffic sensing data. The traffic flow parameters include at least traffic density, traffic volume, and average speed. It is also used to fuse the traffic flow parameters with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section. The structural defect parameters characterize the geometric structural defects that are inherent in the design of the expressway and affect traffic flow. The input unit is used to input the time series of the comprehensive operating state parameters into the prediction model, and obtain the traffic flow stability characteristics that characterize the smoothness of traffic flow operation based on the prediction data. A determining unit is configured to determine the warning level corresponding to the mainline section based at least on the traffic flow stability characteristics; The control unit is used to perform dynamic control on the target objects associated with the mainline section based on the warning level. The target objects include at least one of the lanes and entrance ramps associated with the mainline section.

[0015] According to a third aspect of this application, a traffic early warning and control device for expressways is provided, the device comprising: Memory is used to store computer programs that can run on a processor; The processor is configured to execute, while running the computer program, a traffic early warning and control method for expressways as described above.

[0016] This application provides a traffic early warning and control system and equipment for expressways, which is an intelligent hierarchical control scheme that integrates static structural defects of the road and dynamic traffic flow parameters. The scheme is implemented through the following technical means: First, it calculates the traffic flow parameters of the mainline section based on multi-source traffic perception data, breaking through the limitations of a single data source and achieving accurate perception of the traffic state of the section; second, it integrates the traffic flow parameters with structural defect parameters characterizing road geometric defects to obtain comprehensive operational state parameters that comprehensively reflect the actual carrying capacity of the road, overcoming the shortcomings of existing technologies that only focus on dynamic traffic parameters while ignoring inherent road bottlenecks; third, it uses a predictive model to make short-term predictions of key parameters and establishes future short-term traffic flow stability characteristics based on the predicted values, achieving early identification of traffic flow from stable operation to instability; finally, it determines the early warning level based on the traffic flow stability characteristics and performs dynamic control on the lanes and / or entrance ramps associated with the mainline section according to the early warning level, achieving accurate matching between control intensity and real-time traffic conditions. Through the above solution, this application solves the problems of single operational judgment dimension, poor universality of early warning model, and disconnect between control measures and actual road carrying capacity in the existing technology. It achieves the technical effects of refined perception, intelligent early warning and differentiated linkage control of expressway traffic flow, effectively improving the traffic efficiency and operational stability of bottleneck sections of the expressway mainline. Attached Figure Description

[0017] Figure 1 This is a schematic diagram illustrating the implementation process of the traffic early warning and control method for expressways in this application; Figure 2 This is a schematic diagram of the structural composition of the traffic early warning and control system for expressways in this application; Figure 3This is a schematic diagram of the structural composition of the traffic early warning and control equipment for expressways in this application. Detailed Implementation

[0018] The technical solution of this application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0019] The specific technical features described in the various embodiments in the detailed implementation can be combined in various ways without contradiction. For example, different implementation methods can be formed by combining different specific technical features. In order to avoid unnecessary repetition, the various possible combinations of the specific technical features in this application will not be described separately.

[0020] It should be noted that the terms "first," "second," and "third" used in the embodiments of this application are merely to distinguish similar objects and do not represent a specific order of objects. It is understood that "first," "second," and "third" can be interchanged in a specific order or sequence where permitted. It should be understood that the objects distinguished by "first," "second," and "third" can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in an order other than those illustrated or described herein.

[0021] Figure 1 This is a schematic diagram illustrating the flow of the traffic early warning and control method for expressways in this application. This method can be applied to a traffic early warning and control system for expressways. This system can be installed on traffic control equipment, which may be computer equipment, laptop equipment, server equipment, etc. Figure 1 As shown, the method includes: Step 101: Based on the multi-source traffic perception data of the mainline section of the expressway, calculate the traffic flow parameters of the mainline section. The traffic flow parameters include at least traffic flow density, traffic flow volume, and average speed. Here, multi-source traffic sensing data refers to raw data from different sources that reflect the state of traffic operation. This includes, but is not limited to, the following three categories: 1) Checkpoint vehicle data: Records the time, lane, license plate, etc. of vehicles passing through a specific section, which is used to accurately count the flow of traffic at the section.

[0022] 2) Internet traffic data: Real-time traffic speed and congestion status of roads from navigation maps (such as Gaode and Baidu), supplementing vehicle speed information between sections and across the entire area.

[0023] 3) Mobile signaling data: Anonymous location data generated by communication between user mobile phones and base stations, used to extract travel trajectories and traffic demands over a wide range and long distance.

[0024] In this application, when calculating the traffic flow parameters of the mainline section based on multi-source traffic sensing data of the mainline section of the expressway, the multi-source traffic sensing data includes at least two of the following: checkpoint vehicle passage data, Internet traffic data, and mobile phone signaling data. Specifically, the number of vehicles passing through the mainline section can be counted based on the checkpoint vehicle passage data to obtain the traffic flow rate; the average speed of the mainline section can be obtained based on the Internet traffic data and / or the checkpoint vehicle passage data; and the traffic density can be calculated based on the ratio of the traffic flow rate to the average speed.

[0025] The formula for calculating the traffic flow density at the main line section is as follows: (1) in, Let m be the traffic density at a certain moment on the main line section. Real-time traffic flow at mainline section m at a certain moment; The average running speed of the main line section m at a certain moment.

[0026] The formula for calculating the historical average vehicle speed of the main line section during the same period is as follows: (2) in, The historical average vehicle speed at section m of the main line during a certain period; The total historical traffic flow sample size is the main line. This represents the historical vehicle speed for the same period.

[0027] Based on the multi-source traffic perception data of this application, the traffic flow parameters of the main line section can be calculated, which can overcome the shortcomings of single-type data in terms of spatial accuracy, coverage and sample size.

[0028] Step 102: The traffic flow parameters are fused with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section; the structural defect parameters characterize the geometric structural defects that exist in the design of the expressway and affect traffic flow. Here, the structural defect parameter includes, but is not limited to, at least one of the following: the magnitude of lane number abrupt changes, the degree of deviation of the entrance / exit spacing from the standard value, and the degree of deviation of the weaving zone length from the standard value. This structural defect parameter can be obtained through the following methods: Extract structural feature data of the location of the mainline section from the road topology profile of the expressway; Calculate the initial defect value based on the degree of deviation between the structural feature data and the preset standard value; The initial defect values ​​are normalized to obtain the structural defect parameters.

[0029] In this application, the traffic flow parameters and the structural defect parameters can be dimensionlessly processed to obtain processing parameters, and each parameter in the processing parameters is mapped to the same numerical range; then the processing parameters are weighted and summed to obtain the comprehensive operating status parameters.

[0030] Here, dimensionless processing is achieved through methods including but not limited to extreme value normalization, Z-score standardization, and fractional scaling standardization.

[0031] The following example uses extreme value normalization: We can first perform unified extreme value normalization on the four types of heterogeneous index data: traffic density, traffic flow, average speed, and structural defect parameters, so that each parameter is mapped to the dimensionless interval [0, 1].

[0032] in, The original input value; For normalized variables; output the normalized traffic flow rate uniformly. Normalized average velocity Normalized traffic density Normalized structural defect parameters The subscript 'm' represents the main section number, the subscript 't' represents the time, and the superscript '*' indicates that the parameter has been dimensionless (i.e., processed parameter). After these processed parameters are mapped to the same numerical range, they can be directly weighted and summed to obtain the comprehensive operating status parameters.

[0033] In this application, the above processing parameters are weighted and summed to obtain the comprehensive operating state parameters of the main section at time t. The methods for determining the weighting coefficients include, but are not limited to: Empirical Value Method: This method directly sets the weights of each parameter based on engineering practice experience. In a specific example of this embodiment, considering that average speed best reflects traffic operation quality, and that the structural defect parameter is a static correction term, the weights can be set as follows: traffic flow weight 0.22, speed weight 0.38, density weight 0.25, and structural defect weight 0.15. This set of weights has been verified in measured data from multiple urban expressways and can effectively reflect the overall operational status of the cross-section.

[0034] Subjective weighting in the Analytic Hierarchy Process (AHP): Experts in the field of traffic engineering compare each parameter pairwise to determine the importance of traffic operation status, construct a judgment matrix, and calculate the subjective weight of each parameter using the eigenvalue method.

[0035] Entropy-based objective weighting: Based on the degree of variation of each parameter in historical data, the objective weight of each parameter is determined using information entropy. The greater the degree of variation of a parameter, the smaller its information entropy, and the greater its assigned weight.

[0036] Combination weighting method: In order to take into account both subjective experience and objective data patterns, the subjective weights of AHP and the objective weights of entropy method can be linearly combined or multiplied to obtain the final combined weights.

[0037] The comprehensive operating status parameters obtained after weighted summation S m,t This is a dimensionless value, and its magnitude comprehensively reflects the traffic operation quality of that section at the current moment. The larger the value, the worse the overall operation of the section and the more congested it is; the smaller the value, the better the operation.

[0038] The calculation method for determining the operating status of multi-source data coupling in the main line is as follows:

[0039] in, The main line section comprehensive operating status value (i.e., comprehensive operating status parameter); , , , These are the coupling weighting coefficients; Normalization coefficients for abrupt changes in mainline lanes and defects in entrance / exit layout. Weight determination can be achieved through a combination of subjective weighting using the AHP (Analog-Philosophy of Things) method and objective weighting using the entropy method, or it can be determined using empirical values. =0.22, =0.38, =0.25, =0.15 Based on the dimensionless state value of the coupled output, and combined with the critical threshold of the traffic flow measured in engineering, the mainline operation status is calibrated. Taking a section of a city's expressway mainline with a sudden change in the number of lanes as an example, at a certain moment, the measured traffic flow was 1500 vehicles / hour / lane, the average speed was 45 km / h, and the traffic density was 33 vehicles / km. Due to the change in the number of lanes from four to three, the structural defect coefficient of this section, after normalization, was 0.6. After extreme value normalization, each parameter was mapped to the [0,1] interval, and weighted according to empirical weights to obtain the comprehensive operating status parameters of this section at the current moment. When this parameter exceeds a preset threshold, the subsequent early warning and classification mechanism can be triggered.

[0040] By incorporating the inherent structural defects of roads into the condition assessment system through the above integration method, even if the traffic flow parameters of sections with design bottlenecks have not yet reached the traditional congestion standard, they can still obtain higher comprehensive operating condition parameter values ​​due to the existence of structural defects. This allows them to be identified and responded to by the early warning system earlier, thereby achieving proactive and preventive control of bottleneck road sections.

[0041] Step 103: Input the time series of the comprehensive operating state parameters into the prediction model, and obtain the traffic flow stability characteristics that characterize the smoothness of traffic flow operation based on the prediction data; Here, a time series of comprehensive operating status parameters of the main line section within the time window can be constructed based on a preset time window; the time series can be input into a prediction model to extract time series feature vectors; the time series feature vectors can be mapped to a preset interval to obtain traffic flow stability features that characterize the smoothness of traffic flow operation.

[0042] For example: Set a preset time window length T, which determines the number of continuous state parameters analyzed by the model in each step. The value of T needs to balance information sufficiency and real-time response, and is typically between 10 and 30 time steps (for example, if the data sampling interval is 1 minute, then T=10 represents analyzing the state evolution over the past 10 minutes). For each mainline section m, at each time t, extract T consecutive comprehensive operating state parameter values ​​from time t-T+1 to time t in chronological order to construct the input sequence for that time step:

[0043] This sequence contains the complete dynamic change process of traffic status at the cross section within the most recent time window, providing a temporal context for the model to analyze the evolution of traffic flow.

[0044] Then, the aforementioned time-series sequence is input into the prediction model, which automatically extracts time-series feature vectors containing information about traffic flow stability. This prediction model can be any neural network model capable of processing time-series data, such as a recurrent neural network (RNN), a long short-term memory network (LSTM), a gated recurrent unit network (GRU), a temporal convolutional network (TCN), or a Transformer model based on an attention mechanism. Among these, LSTM is particularly suitable for modeling traffic flow sequences with significant temporal correlation and delay effects because it can effectively capture long-distance temporal dependencies and alleviate the gradient vanishing problem. This embodiment preferably uses LSTM as the prediction model. LSTM uses a gating mechanism to control the flow and retention of information; its internal structure mainly consists of a forget gate, an input gate, and an output gate.

[0045] Forget gate: Determines which old information to discard from the cell's previous state. It uses the current input as its basis. S m,τ and the hidden state of the previous time step h τ-1 The input is the forgetting ratio in the range [0,1], which is output by the Sigmoid function.

[0046] Input gate: Determines which new information to store in the cell state. It also takes the current input and the previous hidden state as input, uses the sigmoid function to determine which values ​​to update, and uses the tanh function to generate candidate vectors of new information.

[0047] Cell state update: Combining the results of the forgetting gate and the input gate, some information from the old cell state is forgotten, and new candidate information is added to form the updated cell state. cτ This mechanism enables LSTM to retain important information for a long time while forgetting irrelevant information.

[0048] Output gate: Determines which hidden state to output based on the current cell state. It uses the sigmoid function to determine which parts of the cell state need to be output, then processes the cell state through the tanh function and multiplies it with the output gate result to obtain the current hidden state. h τ .

[0049] In this application, a Long Short-Term Memory (LSTM) network model is used to extract the temporal evolution features of traffic flow. The gating calculation and feature output are shown in the following formula:

[0050]

[0051]

[0052]

[0053]

[0054]

[0055] in: : The current timing step; -1: Step to the previous historical moment; , , These are the input gate, forget gate, and output gate, respectively. It is in the cellular state.

[0056] For the Sigmoid activation function:

[0057] It is the hyperbolic tangent activation function; This is a hidden state.

[0058] , Let be the weight matrix, where, The previous hidden state The corresponding weight matrix; Current traffic input characteristics The corresponding weight matrix.

[0059] b is the bias term.

[0060] To multiply element by element, the values ​​at the same position in the matrix are multiplied one by one.

[0061] The above process is performed sequentially for each time step τ (from t-T+1 to t) within the time window, ultimately obtaining the hidden state at the end of the window at time t. h t The hidden state vector is the extracted temporal feature vector. F m,t It encodes all the temporal evolution patterns of the comprehensive operating status parameters within the time window, and is a deep representation of the traffic flow characteristics.

[0062] In this application, the formula for taking the hidden state at the final time of the window as the temporal feature vector is as follows:

[0063] Then, the above time-series feature vectors F m,t The feature vector is transformed into a scalar value through a mapping network, and this value is made to fall within a preset interval, serving as a traffic flow stability feature characterizing the smoothness of traffic flow. In a specific implementation, the feature vector can first be linearly transformed through a fully connected layer, and then the output value can be compressed to the [0,1] interval through a Sigmoid activation function to obtain the traffic flow stability score. Score m,t : ( + )

[0064] in, Original time-series feature vector; The first-level linear transformation weight matrix matches the input features in dimension. ; The first layer of bias terms corrects the linear output baseline offset and prevents the neuron output from always being 0; This is the output weight vector for the second layer.

[0065] For the Sigmoid activation function:

[0066] in, The closer the value is to 1, the more stable the traffic flow is in the time evolution; the closer the value is to 0, the more unstable the traffic flow is, that is, it is heading towards congestion or has already entered an unstable state.

[0067] Through the above steps, deep stability features can be automatically extracted from multi-dimensional comprehensive state parameter sequences, overcoming the shortcomings of existing technologies that rely on a single speed threshold or manual experience rules to determine congestion, and significantly improving the sensitivity and accuracy of early warning of traffic instability.

[0068] Step 104: Determine the warning level corresponding to the mainline section based at least on the traffic flow stability characteristics; Here, the current vehicle speed deviation can be calculated based on the historical average speed of the main line section; then, the vehicle speed deviation is fused with the traffic flow stability features to generate a comprehensive warning index; the comprehensive warning index and / or the traffic flow stability features are input into a trained multi-classifier to obtain a probability distribution of multiple warning levels; the level corresponding to the highest probability in the probability distribution is taken as the warning level.

[0069] The formula for calculating the current vehicle speed deviation is as follows:

[0070] In the formula, the historical average vehicle speed of the mainline section m is used. The measured vehicle speed V at the current moment m,t Calculate the current vehicle speed deviation t Dev m,t This deviation indicates the degree to which the current vehicle speed deviates from historical normal levels; the greater the deviation, the more abnormal the current traffic conditions.

[0071] The historical average vehicle speed for the same period can be obtained by grouping historical data according to weekday attributes (weekdays / holidays) and time periods (such as every 5 minutes or every 15 minutes). For example, to calculate the vehicle speed deviation at 8:30 am on a certain Tuesday, the historical average vehicle speed for all Tuesdays at 8:30 am in the past few weeks is taken as the benchmark.

[0072] By integrating vehicle speed deviation with traffic flow stability characteristics, a comprehensive early warning index is generated. W m,t The traffic flow stability characteristic is the traffic flow stability score obtained in the preceding steps. Score m,t Its value range is [0,1], and the closer the value is to 1, the smoother the traffic flow. To maintain consistency with the deviation in direction (i.e., the larger the value, the more warning is needed), the stability score is directionally reversed, that is, (1- Score m,t This characterizes the degree of vehicle instability.

[0073] The fusion can be achieved using linear weighting or nonlinear combination methods. In a specific implementation, the comprehensive early warning index can be calculated using the following formula:

[0074] in, The comprehensive early warning index is defined by μ and ε, which are fusion coefficients that can be set by fitting historical data or through experience. It characterizes the stability of vehicle movement.

[0075] The comprehensive early warning index and / or the traffic flow stability feature are used as classification features and input into a trained multi-classifier. The output is a probability distribution corresponding to multiple early warning levels, and the level with the highest probability is taken as the early warning level of the section at the current moment.

[0076] In a preferred embodiment, traffic flow stability scores are... Score m,t The input feature vector is directly used as input and mapped to the probabilities of four warning levels through a Softmax multi-classifier.

[0077] For example, a combined feature function can be constructed by integrating traffic flow stability and anomaly confidence using the following formula:

[0078] Output the probability of the four control levels using the Softmax multi-class classification model:

[0079] k=1,2,3,4 in, The probability values ​​are for control levels one, two, three, and four, respectively. The final warning level is determined by... k =

[0080] In this application, the four warning levels are arranged from low to high severity as follows: Level 4 - Traffic flow is stable and no control is required; Level 3 - Traffic flow fluctuates slightly and requires attention but no mandatory intervention is needed; Level 2 - Traffic flow continues to deteriorate and tends to become congested, requiring proactive intervention; Level 1 - Traffic flow is severely unstable or has formed a large-scale congestion, requiring the strictest control measures.

[0081] In this application, the early warning judgment is based on the traffic flow stability features extracted by time-series deep learning, which can capture the dynamic evolution process of traffic flow from stability to instability and achieve early warning of "knowing before congestion occurs". Secondly, it integrates historical comparison information such as vehicle speed deviation, taking into account both the absolute state and the relative anomaly. In addition, it adopts a multi-classifier to output level probability and determines the warning level in a probability maximization manner, replacing the hard judgment of fixed thresholds, and has better adaptability to different road sections and time periods.

[0082] Step 105: Based on the warning level, perform dynamic control on the target objects associated with the mainline section. The target objects include at least one of the lanes and entrance ramps associated with the mainline section.

[0083] Here, the lanes associated with the mainline section refer to the traffic lanes upstream of the mainline section or at the section location. Lane speed limit control dynamically adjusts the maximum permissible speed of lanes in that section through roadside facilities such as variable speed limit signs and lane indicators to smooth traffic flow and delay or alleviate congestion.

[0084] The entrance ramp associated with the mainline section refers to the entrance ramp located upstream of that section, through which vehicles merge into the mainline. Entrance ramp flow control dynamically adjusts the traffic volume merging into the mainline using ramp traffic lights to match the remaining capacity of the mainline.

[0085] The "dynamic" aspect described in this application is reflected in two aspects: first, the triggering and lifting of control measures are automatically executed based on the real-time warning level, rather than based on a fixed schedule; second, the intensity of control measures increases with the progression of the warning level, achieving a refined match with traffic conditions.

[0086] In this application, the warning levels are divided into four levels in order of increasing severity: Level 4, Level 3, Level 2, and Level 1, wherein Level 1 is the most severe. If the warning level is Level 4, it indicates that the mainline traffic flow is in good condition, with overall operating parameters and traffic flow stability characteristics within the normal range, and the mainline's carrying capacity is sufficient. In this case: no flow restriction control is implemented on the entrance ramps, allowing vehicles to freely merge into the mainline; and / or no speed limit control is implemented on the lanes, with the mainline lanes operating at the default speed limit (e.g., 80 km / h or the road's design speed limit).

[0087] If the warning level is Level 3, it indicates that the traffic flow stability characteristics show slight signs of instability, but are still within a controllable range. To prevent further deterioration, the merging traffic flow at the entrance is lightly constrained to avoid traffic flow superposition and disturbance to the main line. At this time, a first-level flow restriction control can be implemented on the entrance ramp, that is, a slight reduction in the merging traffic flow at the entrance. Specifically, this can be achieved by appropriately extending the red light duration and shortening the green light duration of the ramp traffic lights; and / or implementing a first-level speed limit control on the lane, that is, lowering the speed limit value of the main line-related lane by a small margin (e.g., from 80 km / h to 70 km / h), guiding upstream vehicles to decelerate smoothly in advance, reducing the speed difference and shock wave when reaching the bottleneck section.

[0088] The formula for Level 3 control is as follows:

[0089] in, The baseline flow rate is incorporated into the entrance ramp; For the flow of traffic entering the entrance ramp; This is the diversion adjustment coefficient. .

[0090]

[0091] in, Basic traffic capacity of a single mainline lane; n: Number of mainline lanes; Lane width correction factor; Large vehicle proportion correction factor; Overall reduction factor (0.9-0.95); The ratio of ramp traffic flow to mainline capacity is typically set at 0.10 to 0.12.

[0092] If the warning level is Level 2, it indicates that the traffic flow stability characteristics have shown a significant trend of instability, the comprehensive warning index has risen significantly, the risk of mainline congestion is high, and it is necessary to significantly reduce the merging traffic at the entrance. At this time, a second level of flow restriction control can be implemented on the entrance ramp, where the intensity of the first level of flow restriction is greater than that of the second level; and / or a second level of speed limit control can be implemented on the lane, where the reduction in the second level speed limit is greater than that in the first level speed limit; The formula for Level 2 control is as follows:

[0093] in: This is the diversion adjustment coefficient. .

[0094] If the warning level is Level 1, it indicates that the mainline traffic flow is severely unstable or has formed a large-scale congestion, and the strictest entry flow interruption or extreme flow restriction measures must be implemented. At this time, flow interruption or extreme flow restriction control can be implemented on the entry ramps. Flow interruption means that the traffic lights on the entry ramps remain red for an extended period, completely prohibiting vehicles from merging into the mainline from this point, and guiding through traffic flow to other routes. Extreme flow restriction means that only a very small number of vehicles are allowed to pass when there is a long backlog, with the merging volume into the mainline approaching zero. And / or Level 3 speed limit control can be implemented on the lanes, with the reduction in Level 3 speed limit being greater than that of Level 2 speed limit, that is, the speed limit value is reduced to the minimum allowable value (such as 40 km / h or 30 km / h). The purpose is not only to regulate the arrival rate, but also to ensure driving safety in congested areas and avoid secondary accidents such as rear-end collisions caused by excessive speed differences.

[0095] The formula for Level 1 control is as follows:

[0096] in: The baseline flow rate is incorporated into the entrance ramp; For the flow of traffic entering the entrance ramp; This is the diversion adjustment coefficient. .

[0097] The intensity of flow restriction and the extent of speed limit reduction corresponding to each of the above control levels can be determined in the following ways: For entrance ramp traffic restrictions, the restriction intensity can be calculated based on the comprehensive warning index and the preset diversion adjustment coefficient. The diversion adjustment coefficient is set in segments according to the warning level; the higher the level, the larger the coefficient and the higher the restriction intensity. The ramp merging flow after the restriction can be obtained by multiplying the ramp baseline merging flow (calibrated by parameters such as mainline capacity and number of lanes) by the reduction ratio related to the warning index and diversion coefficient.

[0098] For lane speed limits, the speed limit values ​​at each level can be determined by multiplying the road design speed limit by a progressively decreasing proportional coefficient, or by calibrating the actual effect of different speed limit values ​​on traffic flow stability based on historical data.

[0099] The above parameters can be calibrated through simulation verification and field testing before actual deployment, and continuously optimized and adjusted during operation according to the evaluation feedback mechanism described later.

[0100] In this application, after dynamic control is performed on the target objects associated with the main line section based on the warning level, the implementation effect of the dynamic control can be evaluated by multi-dimensional indicators. The multi-dimensional indicators include at least one of the following: average operating speed of the road segment, traffic saturation, vehicle speed dispersion coefficient, number of traffic conflicts, and comprehensive control operation cost. Based on the evaluation results, the warning level and / or the relevant parameters of the dynamic control are optimized.

[0101] Here, a comprehensive assessment of control measures can be conducted based on the fuzzy-TOPSIS comprehensive evaluation algorithm. The average operating speed of road segments is a positive indicator, while the rest are negative indicators.

[0102] The formula for standardizing the indicators is as follows: (Positive indicator) (Negative indicator) in, The maximum value of all samples for the j-th indicator; It is the minimum value of all samples for the j-th indicator.

[0103] Then, a weighted standardized decision matrix is ​​constructed, specifically by using a combination weighting method to obtain the weights of each indicator. The weights satisfy .

[0104] The formula for calculating the weighted matrix is: .

[0105] in, Represents the weight of the indicator; The standardized results of the indicators are as shown above.

[0106] A positive ideal solution represents the theoretically optimal control state, while a negative ideal solution represents the theoretically worst control state. The formula is as follows: ,

[0107] ,

[0108] in, Let be the positive ideal solution index, representing the optimal weighted value of the j-th index among all the schemes to be evaluated; The negative ideal solution index represents the worst weighted value of the j-th index among all the solutions to be evaluated. Distance to the ideal solution, spatial distance between the i-th scheme and the optimal control benchmark. The smaller the value, the closer the solution is to the optimal ideal state, and the better the control effect. Distance to the negative ideal solution, spatial distance between the i-th scheme and the worst control benchmark. The larger the value, the further the solution is from the worst-case scenario, and the better the control effect.

[0109] Overall relevance is the core indicator for evaluating the quality of a control plan, and the calculation formula is as follows:

[0110] in, Overall relevance .

[0111] Based on overall relevance The magnitude of the value determines the effectiveness of the control measures. The larger the value, the better the overall implementation effect of the corresponding control plan. The smaller the value, the more necessary it is to further optimize the control effect or adjust the corresponding parameters of the dynamic control.

[0112] In this application, after each control plan is implemented, the overall effect of the plan is quantitatively evaluated through a multi-dimensional evaluation system to obtain an evaluation value characterizing the quality of control (such as the comprehensive closeness in fuzzy-TOPSIS evaluation). When the evaluation value is lower than the preset expected standard or shows a continuous downward trend, a parameter feedback optimization mechanism is triggered. This mechanism can be executed periodically according to a preset cycle (such as weekly or monthly) or immediately when the evaluation result triggers a threshold.

[0113] Specifically, based on the evaluation results, the diversion adjustment coefficients and / or weighted fusion weight coefficients of the traffic loss stabilization threshold and early warning level can be adjusted in reverse.

[0114] The critical threshold for traffic flow instability refers to the boundary value of comprehensive operational status parameters or traffic flow stability score used to determine whether traffic flow transitions from a stable to an unstable state. The setting of this threshold directly affects the sensitivity of early warnings: if the threshold is set too low, warnings will be triggered too late, making early intervention impossible; if the threshold is set too high, warnings will be triggered frequently, leading to unnecessary control interventions and wasted resources.

[0115] The diversion adjustment coefficient determines the reduction in merging traffic at entrance ramps under each level of warning. The rationality of the coefficient directly affects the actual effect of the control: if the coefficient is too small, the flow restriction will be insufficient and the congestion on the main line will not be alleviated; if the coefficient is too large, the flow restriction will be excessive, the vehicles at the entrance will be severely backed up, and the congestion may even be transferred to the auxiliary roads or surface roads.

[0116] The weighted fusion coefficients of the integrated operating status parameters (such as flow rate weight α, velocity weight β, density weight γ, and structural defect weight δ) determine the sensitivity of the state assessment to each sub-parameter. Reasonable weights should enable the integrated operating status parameters to most accurately distinguish different levels of congestion.

[0117] Taking a bottleneck section of a city expressway with a sudden change in the number of lanes as an example, the diversion adjustment coefficients were initially set at 0.10 for Level 3 control, 0.30 for Level 2, and 0.60 for Level 1. After one month of continuous operation, multi-dimensional evaluation results showed that in the control cases triggered by Level 2 warnings, the improvement in mainline saturation was only 5%, far below the expected target (15%), while the queuing lengths at the entrance ramps were all within the allowable range.

[0118] Based on the aforementioned feedback optimization mechanism, the system determined that the diversion adjustment coefficient for the second-level control was too low, and increased it from 0.30 to 0.35 by a preset step size (0.05). The results of the next evaluation cycle showed that the mainline saturation improved by 12%, while the queue length at the entrance ramps remained within a reasonable range. The system continued to slightly increase the coefficient to 0.38, and by the third cycle, the improvement had stabilized at the target value of 15%, indicating that parameter convergence was complete.

[0119] This application achieves dynamic adaptive correction of core parameters through closed-loop feedback: First, it uses multi-dimensional quantitative evaluation results as the driving force, replacing manual experience-based judgment, ensuring that correction decisions are scientifically based; second, it designs targeted optimization strategies for different parameters, ensuring that all parameters converge towards a better direction after correction; third, it gradually approaches the optimal configuration through multi-cycle iteration, avoiding system instability that may be caused by a single large adjustment. This mechanism enables the control system to self-evolve and continuously optimize over time, ensuring the effectiveness of control in long-term operation.

[0120] This application provides a traffic early warning and control method for expressways. It calculates traffic flow parameters at the mainline cross-section by fusing multi-source traffic sensing data and integrates these parameters with structural defect parameters characterizing road geometric defects to obtain operational state parameters that comprehensively reflect the road's actual carrying capacity. Based on the time series of these parameters, a predictive model is used to extract traffic flow stability characteristics, enabling early identification and graded early warning of traffic flow instability trends. According to the warning level, differentiated dynamic control is implemented for lanes and entrance ramps associated with the mainline cross-section, and adaptive iterative optimization of core parameters is driven by a multi-dimensional evaluation and closed-loop feedback mechanism. This solution overcomes industry bottlenecks in existing technologies, such as single data source, disconnect between state perception and road structural defects, delayed early warning, and fixed control parameters. It constructs a complete technical closed loop from accurate perception, intelligent early warning, coordinated control to iterative optimization, achieving refined perception, forward-looking early warning, differentiated coordinated control, and continuous adaptive optimization of traffic flow in bottleneck sections of expressways. This significantly improves the traffic efficiency and operational stability of expressway mainlines under both normal and emergency conditions.

[0121] Figure 2 This is a schematic diagram illustrating the structural composition of the traffic early warning and control system for expressways in this application, as shown below. Figure 2 As shown, the system includes: The calculation unit 201 is used to calculate the traffic flow parameters of the mainline section of the expressway based on multi-source traffic perception data, wherein the traffic flow parameters include at least traffic flow density, traffic flow volume, and average speed; and to fuse the traffic flow parameters with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section; wherein the structural defect parameters characterize the geometric structural defects that exist in the design of the expressway and affect traffic flow. Input unit 202 is used to input the time series of the comprehensive operating state parameters into the prediction model, and obtain traffic flow stability features that characterize the smoothness of traffic flow operation based on the prediction data; Determining unit 203 is used to determine the warning level corresponding to the mainline section based at least on the traffic flow stability characteristics; The control unit 204 is used to perform dynamic control on the target objects associated with the main line section based on the warning level. The target objects include at least one of the lanes and entrance ramps associated with the main line section.

[0122] In a preferred embodiment of this application, the system further includes: Evaluation unit 205 is used to evaluate the implementation effect of the dynamic control using multi-dimensional indicators, including at least one of the following: average road segment operating speed, traffic saturation, vehicle speed dispersion coefficient, number of traffic conflicts, and comprehensive control operation cost. The optimization unit 206 is used to perform feedback optimization on the warning level and / or the relevant parameters of the dynamic control based on the evaluation results. Specifically, it can reverse-correct the traffic loss stabilization threshold, the diversion adjustment coefficient of the warning level, and / or the weighted fusion weight coefficient based on the evaluation results.

[0123] In this application, the multi-source traffic sensing data includes at least two of the following: checkpoint vehicle passage data, internet traffic data, and mobile phone signaling data; the calculation unit 201 is specifically used to count the number of vehicles passing through the main line section based on the checkpoint vehicle passage data to obtain the traffic flow; to obtain the average speed of the main line section based on the internet traffic data and / or the checkpoint vehicle passage data; and to calculate the traffic density based on the ratio of the traffic flow to the average speed.

[0124] In a preferred embodiment of this application, the calculation unit 201 is further configured to process the traffic flow parameters and the structural defect parameters to obtain processing parameters, wherein each parameter in the processing parameters is mapped to the same numerical range; and to perform a weighted summation of the processing parameters to obtain the comprehensive operating status parameters.

[0125] In a preferred embodiment of this application, the system further includes: Extraction unit 207 is used to extract structural feature data of the location of the main line section from the road topology profile of the expressway; The calculation unit 201 is further configured to calculate an initial defect value based on the degree of deviation between the structural feature data and the preset standard value; and to normalize the initial defect value to obtain the structural defect parameters.

[0126] In a preferred embodiment of this application, the system further includes: Construction unit 208 is used to construct a time sequence of comprehensive operating status parameters of the main section within the time window based on a preset time window; The extraction unit 207 is also used to input the time series into the prediction model, extract the time series feature vector, and map the time series feature vector to a preset interval to obtain the traffic flow stability feature that characterizes the smoothness of traffic flow operation.

[0127] In a preferred embodiment of this application, the calculation unit 201 is further configured to calculate the current vehicle speed deviation based on the historical average speed of the main line section; fuse the vehicle speed deviation with the traffic flow stability features to generate a comprehensive warning index; input the comprehensive warning index and / or the traffic flow stability features into a trained multi-classifier through the input unit 202 to obtain a probability distribution of multiple warning levels; and then use the determination unit 203 to select the level corresponding to the highest probability in the probability distribution as the warning level.

[0128] In a preferred embodiment of this application, the warning levels include a first level, a second level, a third level, and a fourth level, which are progressively higher, wherein the first level is the most severe. If the warning level is the fourth level, the control unit 204 will not implement flow restriction control on the entrance ramp and / or will not implement speed limit control on the lane; If the warning level is the third level, the control unit 204 is used to perform a first level of flow restriction control on the entrance ramp and / or perform a first level of speed limit control on the lane; If the warning level is the second level, the control unit 204 is used to perform a second level of flow restriction control on the entrance ramp, where the first level of flow restriction intensity is greater than the second level; and / or to perform a second level of speed limit control on the lane, where the reduction of the second level speed limit is greater than the first level speed limit; If the warning level is the first level, the control unit 204 is used to perform flow interruption or extreme flow restriction control on the entrance ramp, and / or perform three-level speed limit control on the lane; the reduction range of the three-level speed limit is greater than that of the second-level speed limit.

[0129] It should be noted that the traffic early warning and control system for expressways provided in the above embodiments is different from the one described above. Figure 1 The traffic early warning and control methods provided for expressways are based on the same concept. The specific implementation process can be found in the above method implementation examples, and will not be repeated here.

[0130] This application provides a traffic early warning and control system for expressways. It calculates mainline cross-section traffic flow parameters through the fusion of multi-source traffic sensing data and integrates these parameters with structural defect parameters characterizing road geometric defects to obtain operational status parameters that comprehensively reflect the road's actual carrying capacity. Based on the time series of these parameters, a predictive model is used to extract traffic flow stability characteristics, enabling early identification and graded early warning of traffic flow instability trends. According to the warning level, differentiated dynamic control is implemented for lanes and entrance ramps associated with the mainline cross-section, and adaptive iterative optimization of core parameters is driven through multi-dimensional evaluation and a closed-loop feedback mechanism. This solution overcomes industry bottlenecks in existing technologies, such as single data source, disconnect between state perception and road structural defects, delayed early warning, and fixed control parameters. It constructs a complete technical closed loop from accurate perception, intelligent early warning, coordinated control to iterative optimization, achieving refined perception, forward-looking early warning, differentiated coordinated control, and continuous adaptive optimization of traffic flow in bottleneck sections of expressways. This significantly improves the traffic efficiency and operational stability of expressway mainlines under both normal and emergency conditions.

[0131] Figure 3 This is a schematic diagram of the structural composition of the traffic early warning and control equipment for expressways in this application, as shown below. Figure 3 As shown, the device 3000 includes at least one processor 3001 and a memory 3002 for storing a computer program that can run on the processor 3001. When the processor 3001 runs the computer program, it executes the traffic warning and control method for expressways described in the above embodiments of this application. The device 3000 also includes at least one network interface 3004 and a user interface 3003. The various components in the device 3000 are coupled together via a bus system 3005. It is understood that the bus system 3005 is used to implement communication between these components. In addition to a data bus, the bus system 3005 also includes a power bus, a control bus, and a status signal bus. However, for clarity, in… Figure 3 The general labeled all buses as Bus System 3005.

[0132] The user interface 3003 may include a monitor, keyboard, mouse, trackball, click wheel, buttons, touchpad, or touch screen.

[0133] It is understood that memory 3002 can be volatile memory or non-volatile memory, or both. Non-volatile memory can be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), ferromagnetic random access memory (FRAM), flash memory, magnetic surface memory, optical disc or CD-ROM (compact disc read-only memory); magnetic surface memory can be disk storage or magnetic tape storage. Volatile memory can be random access memory (RAM), which is used as an external cache. By way of example, but not limitation, many forms of RAM are available, such as Static Random Access Memory (SRAM), Synchronous Static Random Access Memory (SSRAM), Dynamic Random Access Memory (DRAM), Synchronous Dynamic Random Access Memory (SDRAM), Double Data Rate Synchronous Dynamic Random Access Memory (DDRSDRAM), Enhanced Synchronous Dynamic Random Access Memory (ESDRAM), SyncLink Dynamic Random Access Memory (SLDRAM), and Direct Rambus Random Access Memory (DRRAM).The memory 3002 described in the embodiments of this application is intended to include, but is not limited to, these and any other suitable types of memory.

[0134] In this embodiment, the memory 3002 is used to store various types of data to support the operation of the device 3000. Examples of this data include any computer program used to operate on the device 3000, such as the operating system 30021 and application programs 30022. The operating system 30021 contains various system programs, such as a framework layer, a core library layer, and a driver layer, used to implement various basic business functions and handle hardware-based tasks. The application program 30022 can contain various applications, such as... Figure 1 The system, media player, browser, etc. shown are used for handling and preventing traffic accidents involving intelligent connected vehicles and are used to implement various application services. The program implementing the method of the embodiments of this application can be included in application program 30022.

[0135] Processor 3001 may be an integrated circuit chip with signal processing capabilities. In implementation, each step of the above method can be completed by the integrated logic circuitry in the hardware of processor 3001 or by instructions in software form. The processor 3001 may be a general-purpose processor, a digital signal processor (DSP), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 3001 can implement or execute the methods, steps, and logic block diagrams disclosed in the embodiments of this application. A general-purpose processor may be a microprocessor or any conventional processor. The steps of the methods disclosed in the embodiments of this application can be directly manifested as execution by a hardware decoding processor, or execution by a combination of hardware and software modules in the decoding processor. The software modules may be located in a storage medium, specifically memory 3002. Processor 3001 reads information from memory 3002 and, in conjunction with its hardware, completes the steps of the aforementioned method.

[0136] In an exemplary embodiment, the device 3000 may be implemented by one or more application-specific integrated circuits (ASICs), DSPs, programmable logic devices (PLDs), complex programmable logic devices (CPLDs), field-programmable gate arrays (FPGAs), general-purpose processors, controllers, microcontrollers (MCUs), microprocessors, or other electronic components to perform the aforementioned method.

[0137] In an exemplary embodiment, this application also provides a computer-readable storage medium, such as a memory 3002 including a computer program, which can be executed by a processor 3001 of device 3000 to complete the steps described in the aforementioned method. The computer-readable storage medium may be a memory such as FRAM, ROM, PROM, EPROM, EEPROM, Flash Memory, magnetic surface memory, optical disc, or CD-ROM; it may also be various devices including one or any combination of the above-mentioned memories, such as AR devices, CR devices, VR devices, MR devices, etc.

[0138] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, performs the traffic warning and control method for expressways as described in the above embodiments of this application.

[0139] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. Furthermore, the features disclosed in the several method or device embodiments provided in this application can be arbitrarily combined without conflict to obtain new method or device embodiments.

[0140] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A traffic early warning and control method for expressways, characterized in that, The method includes: Based on multi-source traffic perception data of the mainline section of the expressway, the traffic flow parameters of the mainline section are calculated, and the traffic flow parameters include at least traffic flow density, traffic flow volume, and average speed. The traffic flow parameters are fused with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section; the structural defect parameters characterize the geometric structural defects that exist in the design of the expressway and affect traffic flow. The time series of the comprehensive operating state parameters is input into the prediction model, and traffic flow stability features characterizing the smoothness of traffic flow operation are obtained based on the prediction data. The warning level corresponding to the mainline section is determined at least based on the traffic flow stability characteristics. Based on the warning level, dynamic control is performed on the target objects associated with the mainline section. The target objects include at least one of the lanes and entrance ramps associated with the mainline section.

2. The method according to claim 1, characterized in that, The method further includes: The effectiveness of the dynamic control is evaluated using multi-dimensional indicators, including at least one of the following: average road segment operating speed, traffic saturation, vehicle speed dispersion coefficient, number of traffic conflicts, and overall control operating cost. Based on the evaluation results, feedback optimization is performed on the warning level and / or the relevant parameters of the dynamic control.

3. The method according to claim 1, characterized in that, The multi-source traffic sensing data includes at least two of the following: checkpoint vehicle passage data, internet traffic data, and mobile phone signaling data; The calculation of traffic flow parameters for the mainline section based on multi-source traffic sensing data from the expressway mainline cross-section includes: The number of vehicles passing through the main line section is counted based on the vehicle passage data at the checkpoint to obtain the traffic flow rate; Based on the internet traffic data and / or the checkpoint vehicle data, the average speed of the main line section is obtained; The traffic density is calculated based on the ratio of the traffic flow volume to the average speed.

4. The method according to claim 1, characterized in that, The process of fusing the traffic flow parameters with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section includes: The traffic flow parameters and the structural defect parameters are processed to obtain processed parameters, and each parameter in the processed parameters is mapped to the same numerical range. The processing parameters are weighted and summed to obtain the comprehensive operating status parameters.

5. The method according to claim 1, characterized in that, The structural defect parameters are obtained in the following way: Extract structural feature data of the location of the mainline section from the road topology profile of the expressway; Calculate the initial defect value based on the degree of deviation between the structural feature data and the preset standard value; The initial defect values ​​are normalized to obtain the structural defect parameters.

6. The method according to claim 1, characterized in that, The step of inputting the time series of the comprehensive operating state parameters into the prediction model, and obtaining traffic flow stability characteristics characterizing the smoothness of traffic flow operation based on the prediction data, includes: Based on a preset time window, a time sequence of comprehensive operating status parameters of the main section within the time window is constructed; The time series is input into the prediction model to extract the time series feature vector; The time-series feature vector is mapped to a preset interval to obtain traffic flow stability features that characterize the smoothness of traffic flow operation.

7. The method according to claim 1, characterized in that, The determination of the warning level corresponding to the mainline section based at least on the traffic flow stability characteristics includes: Calculate the current vehicle speed deviation based on the historical average speed of the main line section; The vehicle speed deviation is fused with the traffic flow stability characteristics to generate a comprehensive early warning index; The comprehensive early warning index and / or the traffic flow stability features are input into a trained multi-classifier to obtain the probability distribution of multiple early warning levels. The level corresponding to the highest probability in the probability distribution is taken as the warning level.

8. The method according to claim 1, characterized in that, The dynamic control of the target objects associated with the main section based on the warning level includes: The warning levels are progressively divided into Level 1, Level 2, Level 3 and Level 4, with Level 1 being the most severe. If the warning level is Level 4, no flow control will be implemented on the entrance ramp, and / or no speed limit control will be implemented on the lane; If the warning level is the third level, a first-level flow restriction control will be implemented on the entrance ramp, and / or a first-level speed limit control will be implemented on the lane; If the warning level is the second level, a second degree of flow restriction control is implemented on the entrance ramp, where the intensity of the first degree of flow restriction is greater than that of the second level; and / or a second level of speed limit control is implemented on the lane, where the reduction in the second level speed limit is greater than that of the first level speed limit; If the warning level is Level 1, flow interruption or extreme flow restriction control is implemented on the entrance ramp, and / or Level 3 speed limit control is implemented on the lane; the reduction range of the Level 3 speed limit is greater than that of the Level 2 speed limit.

9. The method according to claim 2, characterized in that, The step of optimizing the early warning level and / or the relevant parameters of the dynamic control based on the evaluation results includes: Based on the evaluation results, the diversion adjustment coefficients and / or weighted fusion weight coefficients of the traffic loss stabilization critical threshold and early warning level are reversed.

10. A traffic early warning and control system for expressways, characterized in that, The system includes: The calculation unit is used to calculate the traffic flow parameters of the mainline section of the expressway based on multi-source traffic sensing data. The traffic flow parameters include at least traffic density, traffic volume, and average speed. It is also used to fuse the traffic flow parameters with the structural defect parameters of the mainline section to obtain the comprehensive operating status parameters of the mainline section. The structural defect parameters characterize the geometric structural defects that are inherent in the design of the expressway and affect traffic flow. The input unit is used to input the time series of the comprehensive operating state parameters into the prediction model, and obtain the traffic flow stability characteristics that characterize the smoothness of traffic flow operation based on the prediction data. A determining unit is configured to determine the warning level corresponding to the mainline section based at least on the traffic flow stability characteristics; The control unit is used to perform dynamic control on the target objects associated with the mainline section based on the warning level. The target objects include at least one of the lanes and entrance ramps associated with the mainline section.

11. A traffic early warning and control device for expressways, characterized in that, The device includes: Memory is used to store computer programs that can run on a processor; The processor is configured to execute, while running the computer program, a traffic early warning and control method for expressways as described in any one of claims 1 to 9.

Images