Multi-level linkage ramp control method based on multi-source data fusion and related equipment

Abstract

The invention discloses a multi-level linkage ramp control method and related equipment based on multi-source data fusion, and the method comprises the steps: carrying out the clustering analysis of target multi-source traffic trajectory data, and obtaining travel trajectory feature data; on the basis of the travel track feature data, performing weight calibration on the on-bridge ramp of the expressway to obtain on-bridge ramp control priority data; performing short-time prediction on the current traffic flow data to obtain a short-time flow prediction result; performing congestion prediction according to the short-time flow prediction result and the dynamic bearing capacity of the expressway to obtain a congestion section prediction result; and according to the short-time flow prediction result, the congestion section prediction result and the on-bridge ramp management and control priority data, carrying out ramp control on the predicted on-bridge ramps of the expressway congestion section by adopting a layered management and control strategy. The method can achieve the precise dynamic sorting of the turn-off priorities of the ramps, balances the load of the road network, avoids the secondary congestion, improves the overall traffic efficiency of the road network, and can be widely applied to the technical field of traffic control.

Description

Technical Field

[0001] This application relates to the field of traffic control technology, and in particular to a multi-level linkage ramp control method and related equipment based on multi-source data fusion. Background Technology

[0002] Urban expressways, as the backbone of urban transportation networks, serve medium- and long-distance travel. However, with the surge in urban traffic volume, expressway congestion has become increasingly prominent, especially during peak hours or in the event of unforeseen incidents. Localized congestion can rapidly spread along the main expressway and easily propagate to the surface road network via ramps, creating cross-level traffic congestion. Traditional ramp control methods, relying mainly on timed signal control, independent ramp adjustments, and manual intervention, suffer from response lags and insufficient coordination, lacking real-time responsiveness to the dynamic characteristics of traffic flow and failing to balance the collaborative needs of expressways and surface traffic. Furthermore, while breakthroughs in IoT, big data, and AI technologies are driving intelligent transportation systems towards data-driven and dynamic collaboration, these technologies still have significant shortcomings in areas such as the coordinated optimization of elevated and surface traffic, the suppression of cross-level congestion propagation, and the resolution of the conflict between short-distance travel and the core functions of expressways, hindering further improvements in the overall efficiency of the road network.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The embodiments of this application aim to at least partially solve one of the technical problems in the related art. Therefore, the main objective of the embodiments of this application is to propose a multi-level linkage ramp control method and related equipment based on multi-source data fusion, which can achieve accurate dynamic sorting of ramp closure priorities, balance road network load, avoid secondary congestion, and improve the overall traffic efficiency of the road network.

[0005] To achieve the above objectives, one aspect of this application proposes a multi-level coordinated ramp control method based on multi-source data fusion, the method comprising the following steps: Acquire target multi-source traffic trajectory data for expressways; Cluster analysis is performed on the target multi-source traffic trajectory data to obtain travel trajectory feature data; Based on the travel trajectory feature data, the on-ramp of the expressway is weighted and labeled to obtain the on-ramp control priority data. Short-time traffic flow prediction is performed on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-time traffic flow prediction results; Based on the short-term traffic flow prediction results and the dynamic carrying capacity of the expressway, congestion prediction is performed on the expressway to obtain the congestion section prediction results. Based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramps of the predicted congested sections of the expressway in order to suppress the spread of congestion on the expressway.

[0006] To achieve the above objectives, another aspect of this application proposes a multi-level linkage ramp control device based on multi-source data fusion, the device comprising the following modules: The multi-source traffic data acquisition module is used to acquire target multi-source traffic trajectory data for expressways; The travel feature clustering analysis module is used to perform clustering analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; The on-ramp weight calibration module is used to calibrate the weights of the on-ramp of the expressway based on the travel trajectory feature data, so as to obtain on-ramp control priority data. The short-term traffic flow prediction module is used to perform short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain the short-term traffic flow prediction result. The congestion section prediction module is used to predict the congestion of the expressway based on the short-term traffic flow prediction results and the dynamic carrying capacity of the expressway, and obtain the congestion section prediction results. The on-ramp control module is used to control the on-ramp of the predicted expressway congestion section based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, using a hierarchical control strategy to suppress the spread of congestion on the expressway.

[0007] To achieve the above objectives, another aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method.

[0008] To achieve the above objectives, another aspect of the embodiments of this application proposes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0009] To achieve the above objectives, another aspect of this application provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0010] The embodiments of this application include at least the following beneficial effects: This application provides a multi-level linkage ramp control method and related equipment based on multi-source data fusion. This scheme obtains target multi-source traffic trajectory data of the expressway; performs cluster analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; based on the travel trajectory feature data, weights are assigned to the on-ramp of the expressway to obtain on-ramp control priority data; short-term prediction is performed on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-term flow prediction results; based on the short-term flow prediction results and the dynamic carrying capacity of the expressway, congestion prediction is performed on the expressway to obtain congestion section prediction results; based on the short-term flow prediction results, congestion section prediction results, and on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of the predicted congested section of the expressway to suppress the spread of congestion on the expressway. This application embodiment integrates multi-source traffic trajectory data and combines travel trajectory clustering analysis to accurately identify travel characteristics, thereby improving the scientific nature and pertinence of subsequent ramp control strategies. By weighting the on-ramp of expressways, priority data for on-ramp control is obtained, enabling precise dynamic sorting of ramp closure priorities. This solves the problem of rigid strategies caused by traditional static thresholds and avoids the erroneous closure or inefficient control of high-weight ramps. Based on short-term traffic flow prediction results, congestion section prediction results, and on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of predicted congested sections of expressways. This balances the road network load, avoids secondary congestion, suppresses the spread of congestion on expressways, and improves the overall traffic efficiency of the road network. Attached Figure Description

[0011] Figure 1 This is a flowchart of the steps of the multi-level linkage ramp control method based on multi-source data fusion provided in the embodiments of this application; Figure 2 This is a schematic diagram of the workflow of the multi-level linkage ramp control method based on multi-source data fusion provided in the embodiments of this application; Figure 3 This is a schematic diagram of the structure of the multi-level linkage ramp control device based on multi-source data fusion provided in the embodiments of this application; Figure 4 This is a schematic diagram of the hardware structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0012] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of this application and are not intended to limit it. In the following description, when referring to the accompanying drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with those of this application; they are merely examples of apparatuses and methods consistent with some aspects of the embodiments of this application as detailed in the appended claims.

[0013] It is understood that the terms “first,” “second,” etc., used in this application may be used herein to describe various concepts, but unless otherwise stated, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of this application, first information may also be referred to as second information, and similarly, second information may also be referred to as first information. Depending on the context, the words “if,” “when,” or “in response to a determination” as used herein may be interpreted as “when…” or “when…” or “in response to a determination.”

[0014] As used in this application, the terms "at least one", "multiple", "each", "any", etc., "at least one" includes one, two or more, "multiple" includes two or more, "each" refers to each of the corresponding multiples, and "any" refers to any one of the multiples.

[0015] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0016] Before providing a detailed description of the embodiments of this application, some of the nouns and terms involved in the embodiments of this application will be explained first. The nouns and terms involved in the embodiments of this application are subject to the following interpretations.

[0017] (1) Ramp control: A traffic management method applied to the entrance ramps of expressways. By installing traffic lights on the ramps, the speed at which vehicles are allowed to enter the main line is dynamically adjusted according to the real-time traffic conditions of the main line of the expressway (such as flow rate, speed, occupancy rate, etc.). Its main purpose is to prevent or alleviate congestion on the main line of the expressway caused by the merging of traffic. By controlling the traffic flow entering the main line, the smooth and efficient operation of the main line traffic flow is ensured.

[0018] (2) High-level linkage: refers to the integrated monitoring, analysis and collaborative management of the traffic operation status of the elevated expressway system and the ground road system below or directly connected to it in traffic management or control. Through data sharing and strategy collaboration, the overall traffic efficiency of the elevated and ground road networks is optimized to achieve the balance and smooth flow of the overall traffic flow of the urban three-dimensional road network.

[0019] (3) Multi-level linkage: refers to the vertical coordination between elevated roads and ground roads through ramp traffic lights and ground traffic lights, and the coordination between different ramps in a region.

[0020] Currently, urban expressways serve as the backbone of urban transportation networks, providing services for medium- and long-distance travel. However, with the surge in urban traffic volume, expressway congestion has become increasingly prominent, especially during peak hours or in the event of unforeseen incidents. Localized congestion can rapidly spread along the main expressway and easily extend to the surface road network via ramps, creating multi-level traffic congestion. Traditional ramp control methods, primarily relying on timed signal control, independent ramp adjustments, and manual intervention, suffer from issues such as response lag and insufficient coordination. They lack the ability to respond in real-time to the dynamic characteristics of traffic flow and struggle to balance the coordinated needs of expressways and surface traffic. With breakthroughs in IoT, big data, and AI technologies, intelligent transportation systems are gradually evolving towards data-driven and dynamic collaboration. The fusion of multi-source data (such as checkpoint, electronic police, and internet data) provides a technological foundation for accurately perceiving traffic flow characteristics and predicting congestion trends. The introduction of technologies such as adaptive signal control and dynamic ramp adjustment has significantly improved the management efficiency of local road sections. However, these technologies still have significant shortcomings in areas such as the coordinated optimization of elevated and ground traffic, the suppression of congestion propagation across levels, and the resolution of the conflict between short-distance travel and the core functions of expressways, which restricts the further improvement of the overall efficiency of the road network.

[0021] Currently, expressway ramp control technologies mainly fall into two categories: First, single-point demand control, which dynamically adjusts ramp release rates by real-time monitoring of downstream main road occupancy and traffic flow. While this method can alleviate congestion locally, it fails to consider the OD (Origin-Destination) path characteristics and regional road network load balance proposed in this application, thus easily leading to secondary congestion caused by traffic flow shifting. Second, collaborative control, which links multiple ramps based on historical traffic data or simple rules. However, due to the lack of high-precision short-term prediction and dynamic weighting proposed in this application, it is difficult to achieve accurate priority ranking, and it does not incorporate the ground traffic overflow risk prediction proposed in this application, resulting in the failure of high-ground linkage. Meanwhile, current technologies rely solely on static weights or empirical rules for ramp closure strategies, failing to incorporate dynamic factors such as travel distance and alternative routes proposed in this application. This leads to the accidental closure or inefficient management of critical ramps. In terms of elevated-ground linkage control, elevated and ground control strategies are relatively isolated, resulting in weak ability to suppress congestion propagation across levels. The signal timing at off-ramp connection intersections is rigid and cannot adapt to fluctuations in turning traffic flow. Ground overflow control relies on fixed thresholds and lacks the real-time data-driven hierarchical response mechanism proposed in this application.

[0022] For example, the solution of related technology one is as follows: Based on traffic flow detection equipment such as microwave and coil, real-time monitoring of mainline traffic flow, ramp queue length and other data is performed, and signal control thresholds are dynamically set and ramp signal light timing is adaptively adjusted; when the queue of vehicles on the ramp overflows to the ground intersection, the signal timing of key ground intersections is adjusted in conjunction to reduce the merging of traffic flow from the direction of the ramp; at the same time, real-time traffic conditions and control information are pushed to the navigation platform and guidance screen to guide vehicles to detour; finally, the effectiveness of the solution is tracked, evaluated and iteratively optimized through simulation tools. The disadvantages of related technology 1 are: (1) Insufficient data fusion, relying on single detection equipment such as microwave and coil, resulting in limited accuracy of OD path analysis, unable to accurately identify short, medium and long distance travel characteristics, affecting the scientific nature of ramp control weight calibration; (2) Static threshold decision, ramp signal control is based on fixed flow and queue length thresholds, weight allocation is rigid, and it is easy to mistakenly close high-weight ramps or manage them inefficiently; (3) Single high-ground coordination measures, ground linkage only intercepts and prevents overflow through the connection of bridge ramps with signal intersections, and the ability to suppress cross-level congestion is insufficient; (4) Delayed guidance information, road condition push relies on fixed threshold triggering, detour suggestions lack foresight, and cannot effectively guide vehicles to actively avoid congestion.

[0023] The solution of related technology 2 is to aggregate four types of data: traffic police own data, internally shared data, externally perceived data, and data obtained from the Internet. By using models such as event perception and main line traffic efficiency prediction, LSTM (Long Short-Term Memory) time series congestion prediction, real-time adaptive signal control, and self-learning optimization, traffic flow can be optimized to reduce the impact of sudden events. The disadvantages of related technology 2 are: (1) Insufficient utilization of OD features. Although multi-source data is integrated, it is difficult to distinguish the different impacts of different travel distances on congestion, resulting in insufficient targeting of control strategies; (2) Insufficient consideration of congestion propagation. It does not combine the spatial diffusion pattern of congestion (such as the traffic flow transmission relationship between upstream and downstream sections), resulting in the inability to predict the range of sections affected by congestion, which affects the accuracy of control strategies; (3) Lack of hierarchical control mechanism. The ramp control does not distinguish between single-point and regional strategies and lacks the priority dynamic sorting mechanism proposed in this application. Important ramps are at risk of being closed by mistake, and the risk of secondary congestion is high; (4) Weak high-ground linkage diversion. It does not design dynamic diversion technologies such as dynamic grouping light control and variable lanes for off-ramp proposed in this application. The traffic diversion efficiency under the bridge is low, and the risk of cross-level congestion spread still exists.

[0024] In view of this, this application provides a multi-level linkage ramp control method and related equipment based on multi-source data fusion. This scheme acquires target multi-source traffic trajectory data of the expressway; performs cluster analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; assigns weights to the on-ramp of the expressway based on the travel trajectory feature data to obtain on-ramp control priority data; performs short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-term flow prediction results; predicts congestion on the expressway based on the short-term flow prediction results and the dynamic carrying capacity of the expressway to obtain congestion section prediction results; and uses a hierarchical control strategy to control the on-ramp of the predicted congested sections of the expressway based on the short-term flow prediction results, the congestion section prediction results, and the on-ramp control priority data to suppress the spread of congestion on the expressway. This application embodiment integrates multi-source traffic trajectory data and combines travel trajectory clustering analysis to accurately identify travel characteristics, thereby improving the scientific nature and pertinence of subsequent ramp control strategies. By weighting the on-ramp of expressways, priority data for on-ramp control is obtained, enabling precise dynamic sorting of ramp closure priorities. This solves the problem of rigid strategies caused by traditional static thresholds and avoids the erroneous closure or inefficient control of high-weight ramps. Based on short-term traffic flow prediction results, congestion section prediction results, and on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of predicted congested sections of expressways. This balances the road network load, avoids secondary congestion, suppresses the spread of congestion on expressways, and improves the overall traffic efficiency of the road network.

[0025] The multi-level linked ramp control method based on multi-source data fusion provided in this application relates to the field of traffic control technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, smart speaker, smartwatch, or vehicle terminal, but is not limited to these. The server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network. The software can be an application implementing the multi-level linked ramp control method based on multi-source data fusion, but is not limited to the above forms.

[0026] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0027] Please see Figure 1 , Figure 1 This is an optional flowchart of a multi-level linkage ramp control method based on multi-source data fusion provided in an embodiment of this application. Figure 1 The method may include, but is not limited to, steps S101 to S106.

[0028] Step S101: Obtain target multi-source traffic trajectory data for the expressway; Among them, target multi-source traffic trajectory data refers to target data obtained after preprocessing such as data cleaning.

[0029] In practice, data related to the main road and on / off ramps of the expressway are collected based on checkpoint / electronic police equipment and Internet data. The multi-source data mainly includes: expressway network coordinates, on / off ramp point coordinates, road segment connection relationships, traffic zone boundaries (road network topology data), vehicle ID, vehicle passage timestamp, location coordinates, lane number (checkpoint / electronic police data), vehicle ID, GPS (Global Positioning System) point sequence, timestamp, speed, and direction (Internet data).

[0030] After obtaining the initial multi-source traffic trajectory data, the vehicle passage time at the checkpoint is aligned with the GPS trajectory points, missing trajectory segments are interpolated and the abnormal data such as excessive dwell time, abnormal speed, and path jump are removed to obtain the cleaned OD trajectory data (i.e., target multi-source traffic trajectory data). The OD trajectory data includes the start and end coordinates, path point sequence, and road network topology information of each trajectory.

[0031] Step S102: Perform cluster analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; In some embodiments, step S102 may include: performing feature extraction processing on the target multi-source traffic trajectory data to obtain travel feature data to be classified; wherein, the travel feature data to be classified includes vehicle travel distance and the number of vehicles crossing regions; inputting the travel feature data to be classified into a travel trajectory feature classification model to output initial travel feature category data; performing logical rule matching on the initial travel feature category data based on travel trajectory feature classification rules; if the initial travel feature category data matches the travel trajectory feature classification rules normally, then the initial travel feature category data is used as travel trajectory feature data; if the initial travel feature category data does not match the travel trajectory feature classification rules normally, then the initial travel feature category data is corrected according to the travel trajectory feature classification rules to obtain travel trajectory feature data; wherein, the travel trajectory feature data includes short-distance travel features, medium-distance travel features, and long-distance travel features.

[0032] Among them, vehicle travel distance refers to the actual travel distance calculated based on road network topology; vehicle cross-district number refers to the number of cross-districts calculated based on the traffic zone numbers of the origin and destination of the trajectory.

[0033] Optionally, an example of a travel trajectory feature classification rule is as follows: Travel origin-destination (OD) is divided into short-distance, medium-distance, and long-distance categories based on travel distance. Short-distance refers to a travel distance ≤ 2km, with the origin and destination within the same traffic zone; medium-distance refers to a travel distance ≤ 8km, spanning 1-2 zones; and long-distance refers to a travel distance > 8km. It should be noted that the specific values ​​in the classification rules can be set according to actual circumstances, and this embodiment does not impose any restrictions on this.

[0034] In the specific implementation, the dynamic classification process is divided into 4 steps: (1) Feature calculation: For each OD trajectory to be classified, the travel distance and number of cross-regions are extracted; (2) Preliminary classification: The features (travel distance and number of cross-regions) are input into the K-NN model and the preliminary categories are output. The preliminary categories refer to the three categories of short distance, medium distance and long distance labels; (3) Rule verification: Logical rule matching is performed on the output preliminary categories. If the preliminary categories conflict with the classification rules, the preliminary categories are corrected. The data correction refers to the secondary rule verification of the output preliminary categories (i.e., short distance, medium distance and long distance labels) to determine the accuracy of the classification labels. If the preliminary categories conflict with the classification rules, they are reclassified to ensure that the final generated classification labels are correct; (4) Result output: The final classification labels are generated.

[0035] Step S103: Based on the travel trajectory feature data, the on-ramp of the expressway is weighted to obtain on-ramp control priority data. In some embodiments, step S103 may include: performing traffic flow statistical analysis on the current traffic flow data of the expressway based on travel trajectory feature data and target multi-source traffic trajectory data to obtain traffic flow statistical analysis results; and using a dynamic weight calibration formula to calibrate the weights of the on-ramp of the expressway based on travel trajectory feature data, traffic flow statistical analysis results, vehicle travel distance and alternative route data to obtain on-ramp control priority data.

[0036] In the specific implementation, firstly, by comparing the OD trajectory data with the coordinates of the on-ramps and off-ramps, the relationship between the OD trajectory and the on-ramps and off-ramps is fitted. Combined with the obtained OD trajectory classification labels, the overall traffic flow and the proportion of short / medium / long travel distance traffic flow served by each on-ramps and off-ramps (i.e., traffic flow statistical analysis results) are obtained, which lays a data foundation for the subsequent work of determining the control weight of the ramps. Then, based on the OD path classification results, combined with multiple factors such as travel distance, traffic flow, and alternative routes, the weight of the on-ramps is determined through normalization processing. The priority sequence for closing each on-ramps in the event of congestion on the main expressway and the key protection level of each off-ramps are determined.

[0037] The specific details of the influencing factors are as follows: (1) Average travel distance served by the ramp. The average travel distance of vehicles served by the ramp. The larger the average distance, the more the ramp is functionally inclined to serve long-distance travel, which is in line with the overall service value orientation of the expressway. It should be given priority protection and should be closed as much as possible.

[0038] (2) Total traffic flow of the ramp. The total traffic flow from the ramp into the main line of the expressway. The larger the traffic flow, the higher the functional status it plays. However, on the other hand, the impact on the main line will be more obvious when the main line is congested. Further analysis of its benefits is needed, including factors such as the proportion of short-distance trips and whether there are alternative routes.

[0039] (3) Short-distance travel ratio. The higher the proportion of short-distance travel served by the ramp, the more short-distance travel the ramp serves. When congestion occurs on the main line of the expressway, it should be closed first to avoid excessive short-distance travel traffic occupying the expressway capacity that mainly serves medium and long-distance travel.

[0040] (4) Are there alternative routes? That is, are there alternative routes parallel to the expressway in the overall direction of travel? If there are alternative routes near the ramp, they are allowed to be closed first; if there are no alternative routes, they should be closed as much as possible.

[0041] Step S104: Perform short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-term traffic flow prediction results; In some embodiments, step S104 may include: performing a preliminary prediction on the historical traffic flow data of the previous time step using a Kalman filter model to generate a preliminary traffic flow prediction result; optimizing the preliminary traffic flow prediction result based on the current traffic flow data of the current time step using a Kalman filter model to obtain a current traffic flow prediction result; performing traffic flow prediction on the expressway for a target time period based on the state transition matrix using a Kalman filter model to obtain a target traffic flow prediction result; wherein, the target time period is the target continuous time step after the current time step; and constructing a short-term traffic flow prediction result based on the current traffic flow prediction result and the target traffic flow prediction result.

[0042] Optionally, the short-term traffic forecast result is composed of the single-step traffic forecast result at the current moment (i.e., the current traffic forecast result) and the multi-time-step traffic forecast result at future moments (i.e., the target traffic forecast result). Among them, the single-step traffic forecast result at the current moment is used to characterize the accuracy of the forecast, and the multi-time-step traffic forecast result at future moments is used to characterize the forecast capability.

[0043] Among them, the multi-time step flow prediction results for future times are obtained based on the single-step flow prediction results for the current time. That is, future predictions are made on the basis of ensuring the reliability of predictions by correcting and verifying the predicted and observed values ​​at the current time.

[0044] In practice, based on the traffic flow data of the main line of the expressway and the on / off ramps obtained from the electronic police / checkpoint equipment, the Kalman filter model is used to gradually optimize the traffic flow prediction value by combining historical prediction and real-time observation data through state equations and observation equations. This allows for short-term traffic flow prediction of the expressway. Then, based on the actual traffic capacity of the expressway, the congestion events and the range of congested sections are further predicted.

[0045] Specifically, firstly, based on the optimal state estimate of the expressway traffic flow at the previous moment, the prior state and uncertainty of the expressway traffic flow at the current moment are deduced. That is, the evolution law of expressway traffic flow and rate of change is described using the state transition matrix, and the error of the model itself is quantified by combining the process noise covariance to generate preliminary prediction results, providing a benchmark for subsequent observation fusion. Then, the prediction results are optimized by fusing real-time expressway observation data, and the credibility of the Kalman gain dynamic trade-off model prediction with the observation data is calculated. Subsequently, the prior state is corrected using the observation residuals to reduce the uncertainty of the state estimate. Through iterative adjustment, the predicted value of expressway traffic flow gradually approaches the actual traffic state. Finally, by continuously applying the state transition matrix, the single-step prediction is extended to multiple future time steps to predict the traffic flow changes of the expressway in the next 5-30 minutes, providing a data foundation for subsequent regional ramp control and high-level coordination.

[0046] Step S105: Based on the short-term traffic flow prediction results and the dynamic carrying capacity of the expressway, perform congestion prediction on the expressway to obtain congestion section prediction results. In some embodiments, step S105 may include: calculating the theoretical capacity of the expressway based on the actual road design parameters of the expressway; correcting the theoretical capacity based on the real-time correction factor coefficient to obtain the dynamic carrying capacity; quantifying the short-term traffic flow prediction results and the dynamic carrying capacity using a logistic regression model to obtain the congestion probability prediction results; classifying the congestion level of the expressway based on the ratio of the short-term predicted traffic flow to the dynamic carrying capacity in the short-term traffic flow prediction results, combined with the average speed index of the mainline section of the expressway, to obtain the congestion level classification results; and simulating the congestion propagation range and speed of the expressway using a sampling cell transmission model based on the congestion probability prediction results and the congestion level classification results to obtain the congestion section prediction results.

[0047] Dynamic carrying capacity can also be referred to as dynamic traffic capacity.

[0048] In practice, based on short-term traffic flow forecasts, combined with the dynamic carrying capacity of each section of the expressway and external environmental factors, a probabilistic assessment and quantification of congestion risk are achieved, providing a basis for decision-making in subsequent precise management and control.

[0049] Specifically, firstly, based on the actual road design parameters of the expressway (number of lanes, gradient, curvature), the theoretical traffic capacity of each section of the expressway is calculated. Dynamic traffic capacity is calculated by combining real-time correction factor coefficients based on weather, event impacts, and other factors. First, the real-time capacity limit of the road is quantified; then, combining predicted traffic flow and dynamic capacity, Logistic regression is used to quantify the probability of congestion. First, the flow-capacity ratio is transformed into an intuitive risk probability to support tiered response. Second, based on the flow-capacity ratio and the average speed index of the main expressway section, multi-dimensional judgment rules are formulated to classify congestion levels. Finally, based on the probability of congestion occurrence and the congestion level, the Cell Transmission Model (CTM) is used to simulate the propagation range and speed of congestion along the expressway, predict the sections that may be affected by congestion in the future, predict the spatiotemporal spread trend of congestion, and guide the dynamic adjustment of the control scope.

[0050] Step S106: Based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of the predicted congestion section of the expressway in order to suppress the spread of congestion on the expressway.

[0051] Optionally, the tiered control strategy includes a single-point priority control strategy and a regional coordinated control strategy.

[0052] In some embodiments, step S106 may include: based on short-term traffic flow prediction results and congestion section prediction results, using a single-point priority control strategy to dynamically adjust the ramp closure status and adjustment rate corresponding to a single on-ramp adjacent to the upstream of the congested section of the expressway; based on short-term traffic flow prediction results, congestion section prediction results, and on-ramp control priority data, using a regional coordinated control strategy to dynamically and collaboratively adjust the ramp closure status and adjustment rate corresponding to several on-ramps upstream of the congested section of the expressway.

[0053] In practice, based on short-term traffic flow and congestion forecasts, a hierarchical control strategy of prioritizing single-point control and regional coordinated regulation is adopted for expressway sections that may experience congestion. This strategy dynamically adjusts the closure status and adjustment rate of on-ramp to achieve congestion suppression and road network balance.

[0054] Specifically, (1) Single-point priority control: For the upstream adjacent on-ramp of the congested section, the closure status and adjustment rate are dynamically adjusted based on the forecast to quickly block the incremental traffic flow in the congested section and prevent the congestion from continuing to worsen. (2) Regional coordinated control: For multiple on-ramp of the congested section, the closure status and adjustment rate are dynamically adjusted based on the forecast and control weight. The low-weight and close-distance ramps are closed first, and the medium-priority ramps are adjusted linearly according to the traffic overload ratio. The weight and distance factors jointly constrain the adjustment range to balance the pressure on the road network and avoid the spread of congestion to cause large-scale congestion.

[0055] In some embodiments, after step S106, the method may further include: relieving congestion at off-ramp connection points in congested sections of the expressway according to a dynamic grouping traffic light control strategy and dynamic variable lane technology; relieving congestion at upstream signalized intersections on the ground according to a dynamic overflow prevention control strategy after the on-ramp in the congested section of the expressway is closed; and relieving congestion at downstream signalized intersections on the ground according to a high-bandwidth green wave relief strategy after the on-ramp in the congested section of the expressway is closed.

[0056] In practice, the specific details of achieving high-altitude-ground linkage are as follows: For traffic management at off-ramp connection points: Traffic management strategies are implemented for adjacent off-ramp connection points downstream of congested sections; simultaneously, for long-distance off-ramp points with higher priority requiring focused attention, corresponding traffic management strategies are implemented based on their actual operational status. Specifically: (1) Dynamic grouping of traffic lights: For ground-level connecting intersections where left and right turn lanes are placed outside due to the connection of off-ramp, grouping of traffic lights is adopted. By grouping the traffic flow of the intersection according to direction or function, the signal timing scheme is dynamically optimized. The signal phase is used to split conflicting traffic flows in time, avoiding weaving conflicts caused by behaviors such as crossing lanes, and ensuring efficient diversion of off-ramp traffic flow and overall traffic efficiency and safety.

[0057] 1) Equipment deployment: Radar-guided cameras are deployed at the corresponding entrances of the off-ramp connecting intersections to collect real-time data on traffic flow, queue length, and conflict points for each lane at the corresponding entrances; full-screen lights are installed in the straight lanes of the off-ramp, and independent arrow lights are installed in the left-turn and right-turn outer lanes, with separate control from the straight lane signals.

[0058] 2) Control strategy: Divide the light groups according to the traffic flow conflict points, assign non-conflicting directions to the same phase, set up arrow lights in high-flow turning lanes and control them independently; based on the collected real-time traffic data, when the traffic flow is dense when going off the bridge, dynamically extend the green light time for going straight off the bridge, compress the time of low-flow phase, and improve the ability and efficiency of quickly clearing the traffic flow when going off the bridge.

[0059] (2) Dynamic variable lanes: For off-ramp connection intersections where traffic flow turning characteristics fluctuate significantly at different times, dynamic variable lanes are implemented to balance the spatial and temporal resource allocation of the intersection, ensuring efficient diversion of off-ramp traffic and overall traffic efficiency.

[0060] 1) Equipment deployment: Radar-based surveillance equipment is deployed at the corresponding entrances of the off-ramp connecting intersections to collect traffic flow, occupancy, and queuing information for each lane at the corresponding entrance; lane signs are set at the starting point of the guide lane lines at the corresponding entrances to indicate the current driving direction of the lanes, and electronic signs are set for variable guide lanes; lane driving direction signs should preferably be set at the standard section of the entrance lanes to indicate the driving direction of the lanes, and electronic arrow signs are used for variable guide lanes.

[0061] 2) Control Strategy: Based on the collected dynamic traffic data, the dynamic change characteristics of traffic flow at each turning point of the off-ramp connection intersection are identified. The electronic indicator signs of the variable guidance lanes are dynamically adjusted according to the traffic flow changes. The attributes of the entrance lanes are adjusted in accordance with the changes in traffic characteristics to quickly relieve off-ramp traffic flow and avoid the waste of time and space resources of the entrance lanes, which would lead to low efficiency in relieving off-ramp traffic flow and cause congestion and queues to spread to the bridge, affecting the normal operation of the main road traffic flow.

[0062] For overflow control at upstream intersections during ramp closures: To address potential vehicle queuing and congestion caused by ramp closures, a dynamic overflow control strategy is implemented at upstream signalized intersections on the ground level to prevent vehicle overflow from spreading and causing even larger-scale congestion. Specifically: (1) Equipment deployment: Lightning surveillance equipment is uniformly deployed at the upstream signal control intersections, and combined with the existing electronic police and checkpoint facilities at the intersections, plus video AI technology, to identify the traffic flow at each entrance of the upstream intersection, the queuing situation at each entrance, the overflow situation at the intersection, and the queuing situation of vehicles at the exit lane in the overflow direction.

[0063] (2) Control strategies: 1) Evacuation strategy: When the overflow exit vehicles have not queued to the overflow risk position, the evacuation plan is selected according to the queuing situation of each entrance to maximize the evacuation of the queuing traffic flow; 2) Flow control strategy: When the overflow exit vehicles queue to the overflow risk position, the flow control plan is selected according to the queuing situation of each entrance to control the traffic flow merging into the overflow direction; 3) Interception strategy: When the overflow exit vehicles queue beyond the overflow risk position and approach the intersection, the immediate interception + interception plan is activated. The green light for merging into the overflow direction is cut off immediately in this cycle, and the interception plan is activated in the next cycle.

[0064] For the high-bandwidth green wave diversion at downstream intersections after ramp closures: When the closure of on-ramp leads to increased traffic flow and saturation on the ground level under the bridge, a high-bandwidth green wave diversion is implemented at the downstream signalized intersections to improve the main traffic capacity under the bridge and quickly alleviate traffic congestion. Specifically: (1) Equipment deployment: Lightning vision equipment is uniformly deployed at downstream signal control intersections, in conjunction with existing electronic police and checkpoint facilities at the intersections, plus video AI technology, to identify the traffic flow, queuing situation at each entrance of the downstream intersection, and vehicle arrival situation.

[0065] (2) Control strategy: When the ramp is closed, the downstream signal control intersection under the bridge will simultaneously open the green wave plan with a large bandwidth for the main direction. Based on the real-time traffic detection data of the downstream intersection, the green light duration of the non-coordinated direction will be dynamically compressed and given to the coordinated direction, increasing the green wave bandwidth of the coordinated direction of the under-bridge passage, improving the overall traffic capacity of the under-bridge passage, and receiving and quickly diverting ground traffic.

[0066] Furthermore, this application embodiment can also guide drivers to make decisions in advance by pushing real-time traffic conditions, route suggestions, and control instructions, thereby preventing the spread of congestion at ramps and connecting intersections and improving the overall road network operating efficiency. Specifically: (1) Data-driven dynamic guidance: integrate historical patterns, real-time status and future predictions to accurately judge the congestion trend of ramps and related road sections, form dynamic guidance decision support, generate guidance strategies in advance based on prediction results, release detour suggestions in a timely manner, or release control instructions in real time according to emergencies, to ensure the foresight and timeliness of guidance information and avoid the control lag caused by passive response.

[0067] (2) Graded induction: 1) Ramp access status prompts: Dynamic information boards are set up at the ramp entrances to display ramp status and detour suggestions. When ramp closure control is implemented, vehicles are prompted not to enter, and vehicles are guided to divert.

[0068] 2) Route diversion guidance: Dynamic information boards are set up at key nodes of the road network to display information such as ramp access status and traffic congestion prompts, assisting drivers in adjusting their driving strategies midway.

[0069] 3) In-vehicle and mobile terminal push: By communicating with in-vehicle terminals and mobile device navigation platforms, road network-level dynamic information is pushed before or during the journey to guide drivers to plan routes in advance and avoid high-risk areas.

[0070] Steps S101 to S106 as shown in the embodiments of this application involve: acquiring target multi-source traffic trajectory data of the expressway; performing cluster analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; assigning weights to on-ramp ramps of the expressway based on the travel trajectory feature data to obtain on-ramp ramp control priority data; performing short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-term flow prediction results; predicting congestion on the expressway based on the short-term flow prediction results and the dynamic carrying capacity of the expressway to obtain congestion section prediction results; and using a hierarchical control strategy to control the on-ramp ramps of the predicted congested sections of the expressway, based on the short-term flow prediction results, the congestion section prediction results, and the on-ramp ramp control priority data, in order to suppress the spread of congestion on the expressway. This application embodiment integrates multi-source traffic trajectory data and combines travel trajectory clustering analysis to accurately identify travel characteristics, thereby improving the scientific nature and pertinence of subsequent ramp control strategies. By weighting the on-ramp of expressways, priority data for on-ramp control is obtained, enabling precise dynamic sorting of ramp closure priorities. This solves the problem of rigid strategies caused by traditional static thresholds and avoids the erroneous closure or inefficient control of high-weight ramps. Based on short-term traffic flow prediction results, congestion section prediction results, and on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of predicted congested sections of expressways. This balances the road network load, avoids secondary congestion, suppresses the spread of congestion on expressways, and improves the overall traffic efficiency of the road network.

[0071] To explain in detail the principle of the technical solution of the present invention, the overall process of the present invention will be described below with reference to some specific embodiments. It is easy to understand that the following is an explanation of the technical principle of the present invention and should not be regarded as a limitation of the present invention.

[0072] The multi-level linkage ramp control method based on multi-source data fusion provided in this application uses multi-source data (checkpoints, electronic police, and internet trajectories) as a foundation. First, it identifies short, medium, and long-distance travel characteristics through OD trajectory clustering analysis. Then, it combines dynamic weighting to determine ramp control priorities. Next, it supports predictive decision-making through short-term traffic flow prediction and probabilistic assessment of congestion risk. Finally, it adopts a hierarchical control strategy to strengthen the core function of expressways in serving medium and long-distance travel. It dynamically balances the road network load through single-point ramp adjustment and regional collaborative linkage, suppressing the spread of congestion on the main expressway. Furthermore, it integrates dynamic grouping traffic light control and dynamic variable lane technology to ensure efficient traffic flow at off-ramp connection intersections. It implements dynamic overflow prevention control at upstream intersections of closed ramps and large-bandwidth green wave traffic flow at downstream intersections, improving the overall traffic capacity of ground-level channels and forming a three-level collaborative chain of "elevated-ramp-ground" linkage to prevent congestion from spreading across levels. In addition, it simultaneously establishes a data-driven dynamic and hierarchical guidance mechanism, releasing real-time traffic conditions and detour suggestions through multiple channels to guide vehicles to actively avoid congestion. This application establishes a closed-loop management system of "perception-prediction-control-guidance" to form a ramp control method that links regional and high-level areas, providing systematic technical support for urban traffic congestion management.

[0073] Please see Figure 2 , Figure 2 This is a schematic diagram of the workflow of the multi-level linkage ramp control method based on multi-source data fusion provided in the embodiments of this application, as follows: Figure 2 As shown, the specific implementation process of the multi-level linkage ramp control method based on multi-source data fusion is as follows (steps 1 to 8): Step 1: Distribution of travel characteristics and OD trajectory clustering analysis on expressways; This study systematically analyzes urban expressway traffic flow using multi-source data from the internet, electronic police systems, and checkpoints. Regional origin-destination (OD) tracing technology is employed to analyze the regional-level distribution characteristics of expressway traffic flow. The K-Nearest Neighbors (K-NN) model is used to classify urban expressway OD trajectories, identifying short, medium, and long-distance travel paths. The classification results are then combined to analyze the spatiotemporal distribution characteristics of traffic flow within each distance range, providing data support for expressway management and optimization. Step 1 includes the following steps 1.1 to 1.3: Step 1.1, Analysis of travel distribution characteristics: (a) Data Acquisition and Cleaning: (1) Data collection: Based on checkpoint / electronic police equipment and Internet data, collect relevant data of the main road and on- and off-ramps of the expressway. The multi-source data mainly includes: expressway network coordinates, on- and off-ramps location coordinates, road segment connection relationship, traffic zone boundary (road network topology data), vehicle ID, vehicle passage timestamp, location coordinates, lane number (checkpoint / electronic police data), vehicle ID, GPS point sequence, timestamp, speed, and direction (Internet data).

[0074] (2) Data cleaning: Align the vehicle passage time at the checkpoint with the GPS trajectory points, interpolate to complete the missing trajectory segments, and remove abnormal data such as excessive dwell time, abnormal speed, and path jump to obtain the cleaned OD trajectory data, which includes the start and end coordinates, path point sequence and road network topology information of each trajectory.

[0075] (II) Travel characteristics analysis: Statistical analysis of the proportion of travel volume in each time period, such as morning peak, evening peak, and off-peak, to summarize the characteristics of travel time distribution; based on the "origin-destination" coordinates, to understand the distribution of high-frequency OD pairs and summarize the characteristics of travel spatial distribution.

[0076] Specifically, the travel characteristic analysis here is an analysis of the travel characteristics of expressways in the target area (such as the entire city). The aim is to roughly determine the peak travel time and areas, and to grasp the overall distribution of urban traffic characteristics. This can provide macro-level spatiotemporal distribution characteristics for the optimal allocation of resources in actual traffic management.

[0077] Step 1.2, OD path classification: By establishing classification rules, the OD paths are classified using a rule-constrained K-NN model.

[0078] (I) Data Feature Definition and Classification Rules: (1) Feature definition: Travel distance refers to the actual driving distance calculated based on the road network topology; Number of cross-regional trips refers to the number of cross-regional trips calculated based on the traffic zone numbers of the starting and ending points of the trajectory.

[0079] (2) Classification rules: The origin-destination (OD) of travel is divided into short distance, medium distance, and long distance according to the travel distance. Among them, short distance means the travel distance is ≤2km and the origin and destination are in the same traffic zone; medium distance means 2km < travel distance ≤8km, spanning 1-2 zones; long distance means travel distance >8km. It should be noted that the specific values ​​in the classification rules can be set according to the actual situation, and this application embodiment does not limit this.

[0080] (II) Data Classification: Based on the classification rules, establish three categories of labels: short-range, medium-range, and long-range. Construct a K-NN classifier. The calculation expression for the K-NN classifier is as follows:

[0081] In the formula, ; ; This represents the classifier function, which represents the mapping relationship of the K-NN model. The input is the feature vector, and the output is the corresponding class label. This represents the feature vector, i.e., the feature data of the OD trajectory; express The dimensional real vector space represents the dimension of the input feature vector. (Based on the feature definition, it includes two dimensions: travel distance and number of cross-regional trips). Indicates category tags; This represents a set of category labels, namely short distance, medium distance, and long distance.

[0082] The dynamic classification process consists of four steps: (1) Feature calculation: For each OD trajectory to be classified, extract the travel distance and number of cross-regions; (2) Preliminary classification: Input the features (travel distance and number of cross-regions) into the K-NN model and output the preliminary categories, where the preliminary categories refer to the three categories of short distance, medium distance and long distance labels; (3) Rule verification: Perform logical rule matching on the output preliminary categories. If the preliminary categories conflict with the classification rules, then perform data correction on the preliminary categories. Data correction refers to performing secondary rule verification on the output preliminary categories (i.e., short distance, medium distance and long distance labels) to determine the accuracy of the classification labels. If the preliminary categories conflict with the classification rules, then reclassify to ensure that the final generated classification labels are correct; (4) Result output: Generate the final classification labels.

[0083] Step 1.3, Traffic Flow Statistics and Analysis: By comparing the OD trajectory data with the coordinates of the on-ramps and off-ramps, the relationship between the OD trajectory and the on-ramps and off-ramps is fitted. Combined with the obtained OD trajectory classification labels, the overall traffic flow and the proportion of short / medium / long travel distance traffic flow served by each on-ramps and off-ramps are obtained, which lays a data foundation for the subsequent on-ramps control weight calibration work.

[0084] Step 2, Ramp Control Weighting: Based on the OD path classification results, and considering factors such as travel distance, traffic flow, and alternative routes, the on-ramp is weighted through normalization processing. This determines the closure priority sequence for each on-ramp in the event of congestion on the main expressway, as well as the priority protection level for each off-ramp. Step 2 may include the content of steps 2.1 and 2.2: Step 2.1, Influencing Factors: (1) Average travel distance served by the ramp. The average travel distance of vehicles served by the ramp. The larger the average distance, the more the ramp is functionally inclined to serve long-distance travel, which is in line with the overall service value orientation of the expressway. It should be given priority protection and should be closed as much as possible.

[0085] (2) Total traffic flow of the ramp. The total traffic flow from the ramp into the main line of the expressway. The larger the traffic flow, the higher the functional status it plays. However, on the other hand, the impact on the main line will be more obvious when the main line is congested. Further analysis of its benefits is needed, including factors such as the proportion of short-distance trips and whether there are alternative routes.

[0086] (3) Short-distance travel ratio. The higher the proportion of short-distance travel served by the ramp, the more short-distance travel the ramp serves. When congestion occurs on the main line of the expressway, it should be closed first to avoid excessive short-distance travel traffic occupying the expressway capacity that mainly serves medium and long-distance travel.

[0087] (4) Are there alternative routes? That is, are there alternative routes parallel to the expressway in the overall direction of travel? If there are alternative routes near the ramp, they are allowed to be closed first; if there are no alternative routes, they should be closed as much as possible.

[0088] Step 2.2, Weight calibration:

[0089] In the formula, This indicates the ramp weight score. The higher the value, the more important the ramp is, and it should be kept open as much as possible. The lower the value, the more likely the ramp can be closed. The distance score represents the normalized value of the average travel distance served by the ramp. The flow score represents the normalized value of the total flow of the ramp. The short-distance penalty is represented by the normalized value of the proportion of short-distance trips. The larger the value, the stronger the penalty; This represents a bonus for alternative paths; a score of 1 is awarded if there is an alternative path and 0 is awarded if there is no alternative path, serving as a positive incentive. , , , All represent weighting coefficients, satisfying Suggestions Take 0.5, Take 0.2, Take 0.15, Take 0.15.

[0090] Step 3, Expressway Segment Division: Using the entrance / exit points of adjacent on / off ramps as segment boundaries, the mainline between adjacent ramp pairs (the previous exit ramp and the next entrance ramp) is defined as an independent segment along the driving direction. This facilitates the rapid identification of affected segments when expressway congestion occurs, enables statistical analysis of the "ramp-mainline" traffic flow correlation, and allows for the development of precise ramp control strategies.

[0091] The division of expressway sections facilitates subsequent steps such as expressway congestion prediction in step 5 and regional ramp management in step 6. Specifically, the expressway congestion prediction in step 5, including capacity calculation, congestion probability prediction, congestion severity classification, and congestion propagation prediction, is conducted on a segment-by-segment basis. Similarly, in the regional ramp management in step 6, the single-point priority control point refers to the adjacent on-ramp upstream of the congested segment. The same applies to regional coordinated control; its control points are selected from multiple on-ramp upstream of the congested segment.

[0092] Step 4, Short-term Traffic Flow Prediction for Expressways: Based on traffic flow data from the expressway mainline and on / off ramps obtained from electronic traffic enforcement / checkpoint equipment, a Kalman filter model is used. Through state equations and observation equations, combined with historical predictions and real-time observation data, the traffic flow prediction values ​​are gradually optimized to perform short-term traffic flow prediction for the expressway. Then, based on the actual capacity of the expressway, congestion events and the extent of congested sections are further predicted. Step 4 may include the content of steps 4.1 to 4.3: Step 4.1, Prediction Step: Based on the optimal state estimate of the expressway traffic flow at the previous moment, the prior state and its uncertainties of the expressway traffic flow at the current moment are calculated. That is, the evolution law of the expressway traffic flow and its rate of change is described using the state transition matrix, and the error of the model itself is quantified by combining the process noise covariance to generate preliminary prediction results, providing a benchmark for subsequent observation fusion. Among these, the prior state... Uncertainty in predicting the state and the state transition matrix The calculation formula is as follows:

[0093]

[0094]

[0095] In the formula, Indicates the first The prior state estimate at time t, i.e. the prediction result of the unfused observations; This represents the state transition matrix, which describes how the state evolves from the previous time step to the current time step; Indicates the first Posterior state estimation at time 1; Indicates the first The prior covariance matrix at time t represents the uncertainty of the predicted state; Indicates the first The posterior covariance matrix at time t; The process noise covariance matrix represents the uncertainty of the system model. Indicates the time step.

[0096] Step 4.2, Update Step: The prediction results are optimized by fusing real-time observation data of the expressway, and the reliability of the Kalman gain dynamic trade-off model prediction with the observation data is calculated. Then, the prior state is corrected using the observation residuals to reduce the uncertainty of the state estimation. Through iterative adjustments, the predicted expressway traffic flow gradually approximates the actual traffic state. Among these steps, the Kalman gain... Posterior state estimation Posterior covariance matrix The calculation formula is as follows:

[0097]

[0098]

[0099] In the formula, This represents the Kalman gain, which determines the weight of the observation on the state correction. Indicates the first Posterior state estimation at time 1; Represents the observation matrix; Indicates matrix transpose; This represents the observation noise covariance matrix, reflecting the sensor measurement error; Indicates the first The actual observed value at time; Indicates the first The posterior covariance matrix at time t represents the state uncertainty after fusion of observations; This represents the identity matrix, with the same dimensions as the state vector.

[0100] Step 4.3, Multi-Step Prediction Extension: By continuously applying the state transition matrix, the single-step prediction is extended to multiple future time steps to predict traffic flow changes on the expressway over the next 5-30 minutes, providing a data foundation for subsequent regional ramp management and high-level coordination. The calculation formula for the multi-step prediction state estimation is as follows:

[0101] In the formula, Indicates the first Multi-step predictive state estimation at time points; Indicates the first Posterior state estimation at time 1.

[0102] Step 5, Expressway Congestion Prediction: Based on short-term traffic flow prediction results, combined with the dynamic carrying capacity of each expressway segment and external environmental factors, a probabilistic assessment and quantification of congestion risk are achieved, providing a decision-making basis for subsequent precise management. Step 5 may include the content of steps 5.1 to 5.4: Step 5.1, Traffic Capacity Calculation: Based on the actual road design parameters of the expressway (number of lanes, gradient, curvature), calculate the theoretical traffic capacity of each section of the expressway. Dynamic traffic capacity is calculated by combining real-time correction factor coefficients based on weather, event impacts, and other factors. This quantifies the real-time capacity limit of roads. Among these measures is dynamic traffic capacity. The calculation formula is as follows:

[0103] In the formula, This represents the weather impact coefficient, which is taken as 0.8-0.9 under rainy or snowy weather. This represents the event impact coefficient, which is taken as 0.6-0.8 during construction or accidents.

[0104] Step 5.2, Congestion Probability Prediction: Combining predicted traffic flow with dynamic capacity, Logistic regression is used to quantify the probability of congestion occurring. This transforms the flow-capacity ratio into an intuitive risk probability, supporting tiered responses. Among these, the probability of congestion occurrence... The calculation formula is as follows:

[0105]

[0106] In the formula, This represents the probability of congestion occurring, with a value ranging from 0 to 1, indicating the likelihood that traffic flow exceeds capacity. The intercept term represents the logistic regression, which represents the initial log odds when the flow-capacity ratio R=0, and is fitted using historical congestion data; The slope term of the logistic regression represents the increment of the log odds for every unit increase in the flow-capacity ratio R, fitted using historical congestion data. It represents the flow-capacity ratio, which is the ratio of predicted flow to dynamic capacity.

[0107] Step 5.3, Congestion Level Classification: Based on the flow-capacity ratio and the average speed index of the expressway mainline section, a multi-dimensional judgment rule is established to classify congestion levels. The specific classification content is as follows: (1) Smooth flow: flow-capacity ratio <0.6, average speed >60km / h, sparse traffic flow, high traffic efficiency, no need for control; (2) Mild congestion: 0.6≤ <0.75, 45≤ When the speed is less than 60km / h, the traffic density increases and occasional deceleration occurs. Corresponding fine-tuning of the ramp adjustment rate and early warning guidance can be carried out. (3) Moderate congestion: 0.75≤ <1 or 30≤ When the speed is less than 45km / h, the traffic flow is significantly dense and there are frequent starts and stops. Measures such as closing ramps, significantly adjusting ramp adjustment rates, high-ground linkage control, and guiding detour routes can be implemented. (4) Severe congestion: ≥1 or When the speed is less than 30km / h, and the traffic flow is slow or stagnant, measures such as closing ramps, significantly adjusting ramp adjustment rates, high-level joint control, and guiding detour routes can be implemented.

[0108] Step 5.4, Congestion Propagation Prediction: Based on the probability of congestion occurrence and the congestion level, the Cell Transmission Model (CTM) is used to simulate the propagation range and speed of congestion along the expressway, predict the sections that may be affected by congestion in the future, predict the spatiotemporal spread trend of congestion, and guide the dynamic adjustment of the control range.

[0109]

[0110] In the formula, Indicates the first The output flow of a cell (segment unit), that is, the maximum number of vehicles that can leave the cell per unit time; This indicates the demand capacity of downstream cells, which is the maximum number of vehicles that the downstream road segment can receive in the current time period; This indicates the speed of congestion wave propagation (km / h). Indicates free-flow velocity (km / h); Congestion density (veh / km) refers to the density of vehicles when the road is completely blocked. Indicates the first The current vehicle density (veh / km) of each cell.

[0111] Step 6: Regional ramp management: Based on short-term traffic flow and congestion forecasts, for expressway sections that may experience congestion, a hierarchical management strategy combining single-point priority management and regional coordinated regulation is adopted to dynamically adjust the closure status and adjustment rate of on-ramp, thereby achieving congestion suppression and road network balance.

[0112] (a) Single-point priority control: For the upstream adjacent on-ramp of the congested section, based on the forecast, the closure status and regulation rate are dynamically adjusted to quickly block the incremental traffic flow in the congested section and prevent the congestion from continuing to worsen.

[0113]

[0114] In the formula, This indicates the dynamic adjustment of the release rate of the on-ramp and the predicted traffic flow. hour, This indicates that the ramp is fully open; At that time, the ramps are linearly adjusted. hour, This indicates that the ramp is completely closed.

[0115] (ii) Regional coordinated regulation: For multiple on-ramp ramps upstream of congested sections, the closure status and adjustment rate are dynamically adjusted in conjunction with the forecast and control weights. Ramps with low weights and close proximity are closed first, while medium-priority ramps are adjusted linearly according to the proportion of traffic overload. The weight and distance factors jointly constrain the adjustment range, balance the pressure on the road network, and avoid the spread of congestion to prevent large-scale congestion.

[0116]

[0117]

[0118] In the formula, This indicates the ramp weight score. The higher the value, the more important the ramp is, and it should be kept open as much as possible. The lower the value, the more likely the ramp can be closed. Indicates the distance (km) from the ramp to the congested section; This represents the baseline distance (default is 2km) and controls the distance attenuation intensity. Indicates the priority index for ramp control. The smaller the value, the more likely it should be turned off. This represents the threshold for controlling the opening of ramps, set to 0.7. Ramps exceeding this threshold should remain open. This represents the ramp control closure threshold, set to 0.3. A value below this threshold indicates that the ramp should be completely closed.

[0119] Step 7, High Ground Coordination. Step 7 may include the content of Steps 7.1 to 7.3: Step 7.1, Traffic Management at Off-Ramp Connections: Implement traffic management strategies for adjacent off-ramp connections downstream of congested sections; simultaneously, for long-distance off-ramps requiring priority and protection, implement corresponding traffic management strategies based on their actual operational status. Specifically: (a) Dynamic grouping of traffic lights: For ground-level intersections where left and right turn lanes are placed outside due to off-ramp connections, grouping of traffic lights is adopted. By grouping the traffic flow at the intersection according to direction or function, the signal timing scheme is dynamically optimized. The signal phase is used to split conflicting traffic flows in time, avoiding weaving conflicts caused by behaviors such as crossing lanes, and ensuring efficient diversion of off-ramp traffic flow and overall traffic efficiency and safety.

[0120] (1) Equipment deployment: Lightning surveillance equipment is deployed at the corresponding entrance of the off-ramp connecting intersection to collect real-time data on traffic flow, queue length and conflict points of each lane at the corresponding entrance; full-screen lights are set in the straight lane of the off-ramp, and independent arrow lights are set in the left-turn and right-turn external lanes, which are controlled separately from the straight lane signals.

[0121] (2) Control strategy: Divide the light groups according to the traffic flow conflict points, assign non-conflicting directions to the same phase, set up arrow lights in high-flow turning lanes and control them independently; based on the collected real-time traffic data, when the traffic flow is dense when going off the bridge, dynamically extend the green light time for going straight off the bridge, compress the low-flow phase time, and improve the ability and efficiency of quickly clearing the traffic flow when going off the bridge.

[0122] (ii) Dynamic variable lanes: For off-ramp connection intersections where traffic flow turning characteristics fluctuate significantly at different times, dynamic variable lanes are implemented to balance the allocation of temporal and spatial resources at the intersection, ensuring efficient diversion of off-ramp traffic and overall traffic efficiency.

[0123] (1) Equipment deployment: Radar-guided cameras are deployed at the corresponding entrances of the off-ramp connecting intersections to collect traffic flow, occupancy rate and queuing status of each lane at the corresponding entrance; lane signs are set at the starting point of the guide lane line at the corresponding entrance to indicate the current driving direction of the lane; electronic signs are set at variable guide lanes; lane driving direction signs should be set at the standard section of the entrance lane to indicate the driving direction of the lane; electronic arrow signs are used for variable guide lane arrow signs.

[0124] (2) Control strategy: Based on the collected dynamic traffic data, identify the dynamic change characteristics of traffic flow at each turning point of the off-ramp connection intersection. The electronic indicator of the variable guide lane dynamically adjusts the display status according to the traffic flow change and adjusts the attributes of the entrance lane in accordance with the traffic characteristic change to quickly relieve off-ramp traffic flow and avoid the waste of time and space resources of the entrance lane, which would lead to low efficiency of off-ramp traffic flow relief and congestion that spreads to the bridge and affects the normal operation of the main road traffic flow.

[0125] Step 7.2, Overflow control at upstream intersections of ramp closures: To prevent vehicle queuing and congestion that may be caused by the closure of on-ramp, a dynamic overflow control strategy is implemented at the upstream signalized intersections on the ground to avoid the overflow of vehicles spreading and causing larger-scale congestion.

[0126] (1) Equipment deployment: Lightning surveillance equipment is uniformly deployed at the upstream signal control intersections, and combined with the existing electronic police and checkpoint facilities at the intersections, plus video AI technology, to identify the traffic flow at each entrance of the upstream intersection, the queuing situation at each entrance, the overflow situation at the intersection, and the queuing situation of vehicles at the exit lane in the overflow direction.

[0127] (2) Control strategy: 1) Traffic diversion strategy: When the overflow exit vehicles have not queued to the overflow risk position, select a diversion plan according to the queuing situation at each entrance to divert the queuing traffic to the greatest extent possible; 2) Traffic control strategy: When the queue of vehicles at the overflow exit reaches the overflow risk position, select a traffic control plan based on the queuing situation at each entrance to control the flow of vehicles merging into the overflow direction; 3) Traffic diversion strategy: When the queue of vehicles at the overflow exit exceeds the overflow risk position and approaches the intersection, the immediate traffic diversion + traffic diversion plan will be activated. In this cycle, the green light for vehicles merging into the overflow direction will be cut off immediately, and the traffic diversion plan will be activated in the next cycle.

[0128] Step 7.3, Large-bandwidth green wave diversion at downstream intersections after ramp closure: In response to the situation where the closure of the on-ramp leads to an increase in traffic flow and saturation on the ground section under the bridge, a large-bandwidth green wave diversion is implemented at the downstream signalized intersections to improve the main traffic capacity under the bridge and quickly divert traffic flow under the bridge.

[0129] (1) Equipment deployment: Lightning vision equipment is uniformly deployed at downstream signal control intersections, in conjunction with existing electronic police and checkpoint facilities at the intersections, plus video AI technology, to identify the traffic flow, queuing situation at each entrance of the downstream intersection, and vehicle arrival situation.

[0130] (2) Control strategy: When the ramp is closed, the downstream signal control intersection under the bridge will simultaneously open the green wave plan with a large bandwidth for the main direction. Based on the real-time traffic detection data of the downstream intersection, the green light duration of the non-coordinated direction will be dynamically compressed and given to the coordinated direction, increasing the green wave bandwidth of the coordinated direction of the under-bridge passage, improving the overall traffic capacity of the under-bridge passage, and receiving and quickly diverting ground traffic.

[0131] Step 8, Traffic Guidance Information Release: By pushing real-time road conditions, route suggestions and control instructions, drivers are guided to make decisions in advance, avoiding the spread of congestion at ramps and connecting intersections, and improving the overall road network operation efficiency.

[0132] (i) Data-driven dynamic guidance: Integrating historical patterns, real-time status and future predictions, accurately assessing the congestion trend of ramps and related road sections, forming dynamic guidance decision support, generating guidance strategies in advance based on prediction results, issuing detour suggestions in a timely manner, or issuing control instructions in real time according to emergencies, ensuring the foresight and timeliness of guidance information, and avoiding the control lag caused by passive response.

[0133] (II) Graded Induction: (1) Ramp access status prompt: A dynamic information board is set up at the entrance of the ramp to display the ramp status and detour suggestions. When the ramp is closed, the board will prompt vehicles to be prohibited from entering and guide vehicles to divert.

[0134] (2) Path diversion guidance: Dynamic information boards are set up at key nodes of the road network to release information such as ramp access status and traffic congestion prompts to help drivers adjust their driving strategies midway.

[0135] (3) Push to vehicle and mobile terminals: By communicating with the navigation platform of vehicle terminals and mobile devices, dynamic information of the road network is pushed before the vehicle travels or during the journey, guiding the driver to plan the route in advance and avoid high-risk areas.

[0136] It should be noted that this embodiment is only a brief illustrative description of the overall process of the multi-level linkage ramp control method based on multi-source data fusion. The detailed description of each step can be found in the relevant content of the foregoing embodiment, and will not be repeated here. It is understood that the present invention does not limit this.

[0137] This application embodiment acquires target multi-source traffic trajectory data of an expressway; performs cluster analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; based on the travel trajectory feature data, assigns weights to the on-ramp of the expressway to obtain on-ramp control priority data; performs short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-term flow prediction results; based on the short-term flow prediction results and the dynamic carrying capacity of the expressway, performs congestion prediction on the expressway to obtain congestion section prediction results; and based on the short-term flow prediction results, congestion section prediction results, and on-ramp control priority data, adopts a hierarchical control strategy to control the on-ramp of the predicted congested sections of the expressway to suppress the spread of congestion on the expressway. This application embodiment integrates multi-source traffic trajectory data and combines travel trajectory clustering analysis to accurately identify travel characteristics, thereby improving the scientific nature and pertinence of subsequent ramp control strategies. By weighting the on-ramp of expressways, priority data for on-ramp control is obtained, enabling precise dynamic sorting of ramp closure priorities. This solves the problem of rigid strategies caused by traditional static thresholds and avoids the erroneous closure or inefficient control of high-weight ramps. Based on short-term traffic flow prediction results, congestion section prediction results, and on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of predicted congested sections of expressways. This balances the road network load, avoids secondary congestion, suppresses the spread of congestion on expressways, and improves the overall traffic efficiency of the road network.

[0138] In summary, the technical problems to be solved by the multi-level linkage ramp control method based on multi-source data fusion provided in this application are as follows: (1) Solving the problem of low accuracy of OD analysis caused by insufficient fusion of multi-source data. This application integrates multi-source data such as checkpoint, electronic police, and Internet trajectory data, and combines OD trajectory clustering analysis to accurately identify short, medium and long distance travel characteristics, thereby improving the scientific nature and pertinence of ramp control strategies.

[0139] (2) Eliminate the rigidity of static threshold decision-making. This application constructs a weighting model based on factors such as dynamic travel distance, traffic flow, short-distance ratio and alternative routes to achieve accurate dynamic ranking of ramp closure priorities and avoid erroneous closure or inefficient management of high-weight ramps.

[0140] (3) Improve the problem of insufficient cross-level congestion suppression and weak off-ramp diversion capacity caused by the single high-level coordination measures. This application constructs a three-level coordinated diversion chain of "elevated road-ramp-ground": the off-ramp adopts dynamic grouping light control and variable lane technology to improve the diversion efficiency of connecting intersections; the ground implements upstream overflow prevention control and downstream large-bandwidth green wave to block the cross-level spread of congestion.

[0141] (4) Strengthen differentiated management driven by OD characteristics. This application improves the pertinence of management strategies by classifying OD routes in a refined manner to distinguish the differentiated impact of travel distance on congestion.

[0142] (5) Eliminate the risk of secondary congestion caused by the lack of a hierarchical management and control mechanism. This application designs a hierarchical strategy of single-point priority management and regional coordinated linkage, combined with a dynamic priority ranking mechanism, to balance the road network load and avoid secondary congestion.

[0143] (6) Optimize the lag of guidance information. This application integrates short-term prediction results to establish a data-driven dynamic hierarchical guidance mechanism, realize forward-looking road condition push and detour suggestions, and guide vehicles to actively avoid congestion.

[0144] The beneficial effects of the multi-level linkage ramp control method based on multi-source data fusion provided in this application are as follows: (1) Construct a ramp classification and control mechanism with dynamic weight calibration.

[0145] By integrating multiple factors such as travel distance, traffic volume, short-distance ratio, and alternative routes, a dynamic calculation model for ramp closure priority is established to solve the problem of rigid strategies caused by traditional static thresholds. This enables precise protection of high-value ramps (serving long-distance travel) and rapid closure of inefficient ramps (mainly for short distances).

[0146] (2) Form a closed loop of high-level linkage and evacuation.

[0147] 1) Enhanced traffic flow management at off-bridge junctions: Dynamic grouping and traffic light control are used to separate conflicting traffic flows, and variable lanes are adapted to turning requirements in real time to improve traffic efficiency at off-bridge junctions. 2) Ground overflow blocking: By coordinating the upstream three-level overflow prevention strategy (dredging / controlling / intercepting) with the downstream large-bandwidth green wave, the risk of cross-level congestion spread caused by ramp closure is reduced.

[0148] (3) Achieve precise decision-making driven by OD features.

[0149] Based on the identification of short / medium / long-distance travel routes and the analysis of traffic flow spatiotemporal distribution characteristics, differentiated ramp control strategies are formulated to ensure that expressway resources are tilted towards medium and long-distance travel.

[0150] (4) Design a hierarchical and coordinated regional load balancing system.

[0151] By employing a hierarchical control strategy that combines rapid response at a single point with coordinated operation across multiple ramps in a region, the saturation of each node in the road network can be balanced to avoid secondary congestion.

[0152] (5) Enhance the adaptive control capability of the diversion equipment.

[0153] By leveraging radar-visual fusion sensing technology to dynamically optimize signal timing and lane functions, the problem of delayed response from traditional equipment is solved, ensuring efficient traffic flow off the bridge.

[0154] (6) Enhance the foresight and timeliness of guiding information.

[0155] By integrating short-term traffic flow forecasts and congestion probability assessments, a data-driven dynamic hierarchical guidance mechanism is established. Detour suggestions are then released in a tiered manner through channels such as vehicle terminals and roadside information boards to guide vehicles to proactively avoid congested areas.

[0156] The key points of the multi-level linkage ramp control method based on multi-source data fusion provided in this application are as follows: (1) Ramp control weight calibration model and parameter design; (2) Design of short-term traffic flow prediction and congestion prediction models and parameters for expressways; (3) Design of high-altitude linkage closed-loop diversion logic and implementation equipment design.

[0157] In addition, the information guidance scheme of the multi-level linkage ramp control method based on multi-source data fusion provided in this application may also include: future vehicle-mounted warning prompts can replace the current guidance display screen prompts: for the implementation of the current guidance information release, in the future when vehicle-road cooperative technology is mature, the vehicle-mounted warning system can replace the current guidance display screen prompts to provide warning reminders and dynamic guidance to drivers.

[0158] Please see Figure 3 This application also provides a multi-level linkage ramp control device 300 based on multi-source data fusion, which can implement the above-mentioned method. The device includes the following modules: The multi-source traffic data acquisition module 301 is used to acquire target multi-source traffic trajectory data of the expressway. The travel feature clustering analysis module 302 is used to perform clustering analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; The on-ramp weight calibration module 303 is used to calibrate the weights of the on-ramp of the expressway based on the travel trajectory feature data to obtain on-ramp control priority data. The short-term traffic flow prediction module 304 is used to perform short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain the short-term traffic flow prediction result. The congestion section prediction module 305 is used to predict the congestion of the expressway based on the short-term traffic flow prediction result and the dynamic carrying capacity of the expressway, and obtain the congestion section prediction result. The on-ramp control module 306 is used to control the on-ramp of the predicted expressway congestion section based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, using a hierarchical control strategy to suppress the spread of congestion on the expressway.

[0159] It is understood that the content of the above method embodiments is applicable to the present device embodiments. The specific functions implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0160] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the above-described method. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0161] It is understood that the content of the above method embodiments is applicable to this device embodiment. The specific functions implemented by this device embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0162] Please see Figure 4 , Figure 4 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes: The processor 401 can be implemented using a general-purpose CPU (Central Processing Unit), microprocessor, application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application. The memory 402 can be implemented as a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 402 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 402 and is called and executed by the processor 401 using the methods described in the embodiments of this application. Input / output interface 403 is used to implement information input and output; The communication interface 404 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, Ethernet cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.). Bus 405 transmits information between various components of the device (e.g., processor 401, memory 402, input / output interface 403, and communication interface 404); The processor 401, memory 402, input / output interface 403 and communication interface 404 are connected to each other within the device via bus 405.

[0163] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described method.

[0164] It is understood that the content of the above method embodiments is applicable to this storage medium embodiment. The specific functions implemented in this storage medium embodiment are the same as those in the above method embodiments, and the beneficial effects achieved are also the same as those achieved in the above method embodiments.

[0165] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described method.

[0166] It is understood that the content of the above method embodiments is applicable to the embodiments of this program product. The specific functions implemented by the embodiments of this program product are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0167] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0168] The multi-level linkage ramp control method, device, electronic device, storage medium, and program product based on multi-source data fusion provided in this application embodiment acquires target multi-source traffic trajectory data of an expressway; performs cluster analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; assigns weights to on-ramp ramps of the expressway based on the travel trajectory feature data to obtain on-ramp ramp control priority data; performs short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-term flow prediction results; predicts congestion on the expressway based on the short-term flow prediction results and the dynamic carrying capacity of the expressway to obtain congestion section prediction results; and uses a hierarchical control strategy to control the on-ramp ramps of the predicted congested sections of the expressway based on the short-term flow prediction results, the congestion section prediction results, and the on-ramp ramp control priority data to suppress the spread of congestion on the expressway. This application embodiment integrates multi-source traffic trajectory data and combines travel trajectory clustering analysis to accurately identify travel characteristics, thereby improving the scientific nature and pertinence of subsequent ramp control strategies. By weighting the on-ramp of expressways, priority data for on-ramp control is obtained, enabling precise dynamic sorting of ramp closure priorities. This solves the problem of rigid strategies caused by traditional static thresholds and avoids the erroneous closure or inefficient control of high-weight ramps. Based on short-term traffic flow prediction results, congestion section prediction results, and on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramp of predicted congested sections of expressways. This balances the road network load, avoids secondary congestion, suppresses the spread of congestion on expressways, and improves the overall traffic efficiency of the road network.

[0169] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A multi-level linkage ramp control method based on multi-source data fusion, characterized in that, The method includes the following steps: Acquire target multi-source traffic trajectory data for expressways; Cluster analysis is performed on the target multi-source traffic trajectory data to obtain travel trajectory feature data; Based on the travel trajectory feature data, the on-ramp of the expressway is weighted and labeled to obtain the on-ramp control priority data. Short-time traffic flow prediction is performed on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-time traffic flow prediction results; Based on the short-term traffic flow prediction results and the dynamic carrying capacity of the expressway, congestion prediction is performed on the expressway to obtain the congestion section prediction results. Based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, a hierarchical control strategy is adopted to control the on-ramps of the predicted congested sections of the expressway in order to suppress the spread of congestion on the expressway.

2. The method according to claim 1, characterized in that, The clustering analysis of the target multi-source traffic trajectory data to obtain travel trajectory feature data includes: Feature extraction processing is performed on the target multi-source traffic trajectory data to obtain travel feature data to be classified; wherein, the travel feature data to be classified includes vehicle travel distance and the number of vehicles crossing regions; Input the travel feature data to be classified into the travel trajectory feature classification model, and output the initial travel feature category data; Based on the travel trajectory feature classification rules, logical rule matching is performed on the initial travel feature category data; If the initial travel feature category data matches the travel trajectory feature classification rule correctly, then the initial travel feature category data will be used as the travel trajectory feature data. If the initial travel feature category data does not match the travel trajectory feature classification rule, the initial travel feature category data is corrected according to the travel trajectory feature classification rule to obtain the travel trajectory feature data; wherein, the travel trajectory feature data includes short-distance travel features, medium-distance travel features and long-distance travel features.

3. The method according to claim 1, characterized in that, The step of assigning weights to the on-ramp of the expressway based on the travel trajectory feature data to obtain on-ramp control priority data includes: Based on the travel trajectory feature data and the target multi-source traffic trajectory data, traffic flow statistics analysis is performed on the current traffic flow data of the expressway to obtain traffic flow statistics analysis results; Based on the travel trajectory feature data, the traffic flow statistical analysis results, vehicle travel distance and alternative route data, a dynamic weight calibration formula is used to calibrate the weights of the on-ramp of the expressway to obtain the control priority data of the on-ramp.

4. The method according to claim 1, characterized in that, The step of performing short-time prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain short-time traffic flow prediction results includes: Preliminary traffic flow prediction results are generated by using the Kalman filter model to make preliminary predictions based on the historical traffic flow data of the previous moment. The Kalman filter model is used to optimize the preliminary traffic flow prediction result based on the current traffic flow data at the current moment, so as to obtain the current traffic flow prediction result. The Kalman filter model, based on the state transition matrix, is used to predict the traffic flow of the expressway during the target time period, thereby obtaining the target traffic flow prediction result; wherein, the target time period is the target continuous time step after the current moment; The short-term traffic prediction result is constructed based on the current traffic prediction result and the target traffic prediction result.

5. The method according to claim 1, characterized in that, The step of predicting congestion on the expressway based on the short-term traffic flow prediction results and the expressway's dynamic carrying capacity, to obtain congestion section prediction results, includes: The theoretical traffic capacity of the expressway is calculated based on the actual road design parameters of the expressway. The theoretical capacity is corrected based on the real-time correction factor coefficient to obtain the dynamic carrying capacity; The short-term traffic flow prediction results and the dynamic carrying capacity are quantified using a logistic regression model to obtain the congestion probability prediction results; Based on the ratio of the short-term predicted flow rate to the dynamic carrying capacity in the short-term flow prediction results, and combined with the average speed index of the main line section of the expressway, the congestion level of the expressway is classified, and the congestion level classification results are obtained. Based on the congestion probability prediction results and the congestion severity classification results, the sampling cell transmission model simulates the congestion propagation range and speed of the expressway to obtain the congestion section prediction results.

6. The method according to claim 1, characterized in that, The hierarchical control strategy includes a single-point priority control strategy and a regional coordinated control strategy. The step of using the hierarchical control strategy to control the on-ramp of the predicted congested expressway section, based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, includes: Based on the short-term traffic flow prediction results and the congestion section prediction results, the single-point priority control strategy is used to dynamically adjust the ramp closure status and adjustment rate of a single on-ramp adjacent to the upstream of the congestion section of the expressway. Based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, the regional collaborative control strategy is used to dynamically adjust the ramp closure status and adjustment rate of several on-ramps upstream of the congested section of the expressway.

7. The method according to claim 1, characterized in that, After implementing ramp control on the predicted congestion sections of the expressway using a hierarchical control strategy based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data to suppress the spread of congestion on the expressway, the method further includes: Based on the dynamic grouping traffic light control strategy and dynamic variable lane technology, the off-ramp connection intersections of the congested sections of the expressway are treated to alleviate congestion. After the on-ramp of the expressway in the congested section is closed, overflow prevention control is carried out on the upstream signalized intersection on the ground according to the dynamic overflow prevention control strategy; After the on-ramp of the expressway in the congested section is closed, the downstream signalized intersections on the ground are relieved according to the high-bandwidth green wave relief strategy.

8. A multi-level linkage ramp control device based on multi-source data fusion, characterized in that, The device includes the following modules: The multi-source traffic data acquisition module is used to acquire target multi-source traffic trajectory data for expressways; The travel feature clustering analysis module is used to perform clustering analysis on the target multi-source traffic trajectory data to obtain travel trajectory feature data; The on-ramp weight calibration module is used to calibrate the weights of the on-ramp of the expressway based on the travel trajectory feature data, so as to obtain on-ramp control priority data. The short-term traffic flow prediction module is used to perform short-term prediction on the current traffic flow data in the target multi-source traffic trajectory data to obtain the short-term traffic flow prediction result. The congestion section prediction module is used to predict the congestion of the expressway based on the short-term traffic flow prediction results and the dynamic carrying capacity of the expressway, and obtain the congestion section prediction results. The on-ramp control module is used to control the on-ramp of the predicted expressway congestion section based on the short-term traffic flow prediction results, the congestion section prediction results, and the on-ramp control priority data, using a hierarchical control strategy to suppress the spread of congestion on the expressway.

9. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 7.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

Images