Transportation operation monitoring, early warning and decision analysis method based on vehicle-road collaboration

Abstract

This invention discloses a method for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation, belonging to the field of intelligent transportation technology. This invention achieves real-time collection of dynamic traffic elements through a multi-dimensional perception network of vehicles, roadsides, and the environment, as well as V2X communication. It ensures the consistency of spatiotemporal benchmarks for multi-source data, constructs a semantic network of "road-vehicle-environment-event," and, combined with hierarchical feature extraction, achieves multi-dimensional identification of abnormal events such as accidents, congestion, and severe weather, improving identification accuracy. It can predict the propagation path of abnormal events, possesses global road network situation prediction capabilities, generates differentiated early warnings based on a comprehensive risk index, and triggers multi-level responses such as traffic signal adjustments and route planning, significantly shortening response time, optimizing emergency response efficiency, and establishing a real-time feedback mechanism for decision execution effects. System performance is continuously optimized with data accumulation, avoiding the shortcomings of big data platforms lacking intelligent decision-making closed loops.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, and in particular to a method for monitoring, early warning and decision analysis of transportation operations based on vehicle-road cooperation. Background Technology

[0002] With the rapid development of intelligent transportation systems, traditional transportation operation monitoring and decision-making technologies face numerous challenges. For example, Chinese patent CN117727183B discloses a method and system for autonomous driving safety early warning combined with vehicle-road cooperation. The method includes: reading real-time driving data of the target, then reading a driving-related state dataset to obtain a visual model of the target driving, then performing static vehicle-road cooperative safety detection to obtain the static cooperative safety detection result, then performing dynamic vehicle-road cooperative safety detection to obtain the dynamic cooperative safety detection result, generating a safety analysis report, and sending it to a safety early warning device for safety warning. This addresses the problem of existing methods over-reliance on historical data, which increases the difficulty of real-time warning, may lead to a decrease in warning accuracy, and cannot adaptively adjust to changes in the scenario, thus limiting the real-time effectiveness of the warning. Based on the dynamic cooperative safety detection result, it determines in real time whether a vehicle has safety risks or violations and issues a warning, improving the real-time performance and accuracy of the data.

[0003] While the aforementioned patents improve data real-time performance and early warning accuracy, the following problems still exist:

[0004] 1. Existing transportation systems are limited to data on the vehicles themselves and their surrounding local environment, lacking a comprehensive understanding of road infrastructure, macro-level road network conditions, and environmental parameters, making it difficult to construct a complete picture of traffic operation status.

[0005] 2. The risk assessment only targets the safety risks or violations of individual vehicles, and does not identify abnormal events such as accidents, congestion, and severe weather from multiple dimensions, nor can it predict the propagation path and comprehensive risks of events in the road network.

[0006] 3. Without establishing semantic representations and dynamic models of traffic conditions, it is impossible to quantify the impact of environmental factors on the traffic system, making it difficult to support refined traffic management and decision-making. Summary of the Invention

[0007] The purpose of this invention is to provide a method for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation. By utilizing four-dimensional semantic networks and hierarchical feature extraction technology, combined with graph neural networks and time-series data mining, it can accurately identify abnormal events and predict their propagation paths, construct a comprehensive risk assessment model, improve the intelligent analysis capability of complex traffic scenarios, and achieve efficient linkage between early warning, decision-making, and execution, thereby solving the problems mentioned in the background technology.

[0008] To achieve the above objectives, the present invention provides the following technical solution:

[0009] Methods for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation include:

[0010] Multidimensional sensing network construction: Construct a multidimensional sensing network to acquire vehicle status, road environment and infrastructure data in real time, and fuse the collected vehicle status, road environment and infrastructure data to generate multidimensional spatial data;

[0011] Dynamic traffic state analysis: hierarchical feature extraction is performed on the fused multidimensional spatial data to construct a dynamic traffic state model, and the degree of deviation of traffic flow parameters from the normal distribution is observed in real time. Combined with historical observation data and real-time observation values, abnormal events are identified in multiple dimensions.

[0012] Comprehensive risk prediction and assessment: Based on the identification results, correlation analysis is performed on the identified abnormal events to predict the propagation path. According to the degree of harm and propagation risk of the abnormal events, the comprehensive risk index of the abnormal events is calculated.

[0013] Early warning generation and decision-making: Based on the comprehensive risk index, a multi-level response mechanism is triggered, and corresponding early warning information is generated. A decision-making plan is established, and the safety of the decision-making plan is verified. The decision-making instructions are pushed to vehicles and various infrastructures through a two-way communication link.

[0014] Furthermore, the construction of multidimensional sensing networks also includes:

[0015] The vehicle-mounted V2X communication components are used to build bidirectional communication links between vehicles and between vehicles and infrastructure, and a connection is established with the traffic cloud platform based on the V2X communication components.

[0016] The vehicles send vehicle status data packets to adjacent vehicles at a preset period based on a two-way communication link, and each roadside device sends event information in real time, forming a dynamic communication network between vehicles and between vehicles and each infrastructure.

[0017] Time synchronization between vehicles and between vehicles and infrastructure is determined based on a clock synchronization mechanism.

[0018] The vehicle acquires monitoring information of neighboring vehicles based on a two-way communication link, including the real-time location, movement status and intention information of neighboring vehicles, and also collects traffic signal status and road environment data actively uploaded by roadside infrastructure.

[0019] Furthermore, dynamic traffic state analysis also includes constructing a four-dimensional semantic network topology:

[0020] Using physical road entities as topological foundation nodes, the structural skeleton of the road network topology graph is constructed based on topological attributes and historical traffic characteristics.

[0021] The real-time trajectory of the vehicle is mapped as a dynamic edge, and the vehicle position transfer within adjacent time slices is used as the connection relationship of the edge. Based on the edge attribute behavior characteristics, an edge set is constructed.

[0022] Environmental parameters are used as weighting factors in the graph structure, mapped to corresponding road nodes and vehicle behavior edges according to the spatiotemporal grid, and the identified abnormal events are embedded in the four-dimensional semantic network topology.

[0023] Simultaneously, the relationships between nodes in the four-dimensional semantic network topology are learned to construct a semantic representation of traffic status.

[0024] Furthermore, the hierarchical feature extraction for dynamic traffic state analysis also includes:

[0025] The road network topology map is spatially discretized. Based on the coordinate range of road physical entities and lane division, the road network is divided into spatiotemporal grid units. The basic traffic flow features in each spatiotemporal grid unit are extracted to form a spatial grid traffic state vector.

[0026] Extract the edge attributes formed by vehicle position transfers within adjacent time slices to generate a feature sequence characterizing the temporal changes in vehicle movement. At the same time, combine historical traffic data to identify the periodic patterns and sudden change characteristics of traffic flow and generate a temporal dynamic feature vector.

[0027] The spatial grid traffic state vector and the temporal dynamic feature vector are weighted and fused to form a composite feature representation, which is then embedded into the four-dimensional semantic network topology.

[0028] Furthermore, a correlation analysis is conducted based on the comprehensive risk prediction and assessment, specifically including:

[0029] Historical abnormal event data is acquired, and combined with multidimensional spatial data and environmental parameters in the four-dimensional semantic network topology, the propagation patterns of different types of abnormal events under different environmental parameters are extracted, and a propagation rule knowledge base is constructed.

[0030] The identified abnormal events are used as key nodes. Based on the relationship between nodes and edge attributes in the four-dimensional semantic network topology, the potentially affected road nodes and adjacent event nodes are selected based on the propagation rule knowledge base, and an event evolution tree is constructed.

[0031] The tree node attributes include event type, spatiotemporal location, current impact level, and severity. The tree node connection relationship attributes include event propagation direction, propagation probability, and time delay threshold.

[0032] Furthermore, predicting the transmission path specifically includes:

[0033] Using critical nodes of abnormal events as target nodes, multi-path sampling is performed on the connection relationship attributes in the event evolution tree to generate at least one propagation path. During each sampling, the corresponding branch path is selected based on the propagation probability distribution of the current target node.

[0034] By combining the propagation time distribution of similar historical events, the time sequence of each target node affected in each propagation path is obtained, and a set of propagation paths is generated;

[0035] The propagation probability distribution information of each propagation path in the propagation path set is mapped to spatiotemporal grid cells, and the probability of being affected in each grid cell in the future time period is accumulated and calculated to generate a risk probability cloud map for the future time period.

[0036] Identify key hub nodes in the propagation path set, calculate the risk contribution of each key hub node in the propagation process, and prioritize each key hub node based on its risk contribution.

[0037] Furthermore, key hub nodes within the propagation path set are identified, specifically including:

[0038] Extract the node attributes and topology information of each road node in the propagation path set in the four-dimensional semantic network topology;

[0039] Based on node attributes and topology information, the basic traffic characteristics of each road node and the traffic transmission sensitivity between road nodes are obtained, and the traffic weighting coefficient and traffic transmission weighting coefficient are determined.

[0040] Based on the preset law enforcement management method information corresponding to each road node, the law enforcement management weight of the road node is determined according to the management level corresponding to the law enforcement management method.

[0041] The road topology weight value is obtained by comprehensively calculating the law enforcement management weight, traffic weight coefficient and traffic transmission weight coefficient. The road topology weight value is compared with the preset weight threshold, and road nodes with a weight value higher than the preset weight threshold are selected as key hub nodes.

[0042] Furthermore, a multi-level response mechanism includes:

[0043] Establish a mapping rule between early warning levels and response strategies. For example, if the early warning level is low risk, push a text warning to surrounding vehicles and remind them to drive cautiously via in-vehicle navigation; if the early warning level is medium risk, simultaneously trigger roadside traffic light timing optimization and send route suggestions to logistics fleets; if the early warning level is high risk, initiate cross-regional emergency response, push event details and handling plans to the traffic management center, and simultaneously enforce speed limit instructions through two-way communication links.

[0044] Furthermore, before fusing the collected vehicle status, road environment, and infrastructure data, the process also includes: denoising the image data included in the road environment data;

[0045] The noise reduction of image data in road environment data includes:

[0046] Take any image from the road environment data as the image to be denoised;

[0047] The image to be denoised is processed into a grayscale image.

[0048] Take any pixel in the grayscale image as the first pixel; determine the target area with the first pixel as the center and a preset distance as the radius;

[0049] Calculate the difference between the first pixel and other pixels in the target area, and obtain several difference values; compare the several difference values ​​with a preset difference threshold, and mark the pixels with differences greater than or equal to the preset difference threshold as abnormal;

[0050] Iterate through all pixels in the grayscale image and count the number of outlier markers for each pixel.

[0051] The number of abnormal markers is compared with a preset abnormal marker threshold. Pixels with an abnormal marker number greater than or equal to the preset abnormal marker threshold are designated as first abnormal pixels, thus obtaining a number of first abnormal pixels.

[0052] Edge detection is performed on grayscale images to determine edge pixels in the grayscale images;

[0053] Subtracting the edge pixels from the first abnormal pixels yields a number of second abnormal pixels;

[0054] The dispersion of each second abnormal pixel is calculated and compared with a preset dispersion threshold. Pixels with dispersion greater than or equal to the preset dispersion threshold are designated as third abnormal pixels, resulting in a number of third abnormal pixels.

[0055] Several abnormal third pixels are deleted to obtain a noise-reduced grayscale image;

[0056] By iterating through all images in the road environment data, the denoised road environment data is obtained.

[0057] Furthermore, based on the severity and spread risk of the abnormal event, a comprehensive risk index for the abnormal event is calculated, including steps 1-2:

[0058] Step 1: Obtain the average risk value of vehicle a in historical driving data and the propagation risk value of abnormal events on the driving path of vehicle a. Based on the average risk value, the propagation risk value and the expected risk value of abnormal events, determine the comprehensive risk index of abnormal events on vehicle a.

[0059] ;

[0060] in, This represents the overall risk index of the abnormal event to vehicle a; n represents the total number of risk factors in the abnormal event. This represents the average risk value of vehicle a in historical driving data; This represents the risk value of the abnormal event propagating along the travel path of vehicle a. This represents the risk level value of the i-th type of risk factor among n types of risk factors; Let represent the probability density function of the occurrence of risk of the i-th type of risk factor; It represents the cumulative probability of the occurrence of the i-th type of risk factor within the next unit time interval starting from time t; This represents the total cumulative probability of the occurrence of the i-th type of risk factor over all future time periods, starting from time t.

[0061] Step 2: Determine the comprehensive risk index of the abnormal event based on the comprehensive risk index of the abnormal event for vehicle a and the total number of vehicles in the driving path of vehicle a.

[0062] ;

[0063] in, A comprehensive risk index representing abnormal events; This represents the total number of vehicles in the path that vehicle a is traveling on.

[0064] Compared with the prior art, the beneficial effects of the present invention are:

[0065] By leveraging a multi-dimensional perception network encompassing vehicles, roadsides, and the environment, along with V2X communication, real-time collection of dynamic traffic elements is achieved. This ensures a unified spatiotemporal benchmark for multi-source data, constructs a semantic network of "road-vehicle-environment-event," and, combined with hierarchical feature extraction, enables multi-dimensional identification of abnormal events such as accidents, congestion, and severe weather. This improves identification accuracy, predicts the propagation path of abnormal events, and provides global road network situation prediction capabilities. Based on a comprehensive risk index, differentiated early warnings are generated, triggering multi-level responses such as traffic signal adjustments and route planning, significantly shortening response time, optimizing emergency response efficiency, and establishing a real-time feedback mechanism for decision-making execution effects. System performance is continuously optimized with data accumulation, avoiding the shortcomings of big data platforms that lack intelligent decision-making loops. Attached Figure Description

[0066] Figure 1 This is a flowchart of the transportation operation monitoring, early warning, and decision analysis method of the present invention. Detailed Implementation

[0067] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments of the present invention without making any creative efforts shall fall within the scope of protection of the present invention.

[0068] To address the technical challenges of existing technologies being limited to specific vehicles and their surroundings, lacking multi-dimensional risk assessment and propagation prediction, lacking differentiated response and cross-regional collaboration in decision-making mechanisms, and failing to establish semantic representations of traffic conditions, thus hindering the construction of a complete situational awareness map and the support for refined management, please refer to [link to relevant documentation]. Figure 1 This embodiment provides the following technical solution:

[0069] Methods for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation include:

[0070] Multi-dimensional perception network construction: Construct a multi-dimensional perception network covering vehicle, roadside, and environment to acquire vehicle status, road environment, and infrastructure data in real time. Then, integrate the collected vehicle status, road environment, and infrastructure data to generate multi-dimensional spatial data containing timestamps, thereby achieving real-time acquisition of dynamic traffic elements and unified spatiotemporal reference.

[0071] Dynamic traffic state analysis: Layered feature extraction is performed on the fused multidimensional spatial data to construct a dynamic traffic state model, and the degree of deviation of traffic flow parameters from the normal distribution is observed in real time. Combining historical observation data and real-time observation values, abnormal events such as accidents, congestion, and severe weather are identified from multiple dimensions.

[0072] Comprehensive risk prediction and assessment: Based on the identification results, correlation analysis is performed on the identified abnormal events to predict the propagation path. According to the degree of harm and propagation risk of the abnormal events, the comprehensive risk index of the abnormal events is calculated.

[0073] Early warning generation and decision-making: Based on the comprehensive risk index, a multi-level response mechanism is triggered, and corresponding early warning information containing event type, scope of impact and handling suggestions is generated. A multi-objective optimization model including traffic efficiency, safety risk and energy consumption cost is constructed. A multi-agent reinforcement learning (MARL) algorithm is used to establish differentiated decision-making schemes, and the safety of the decision-making schemes is verified. A real-time feedback mechanism for decision execution effect is established to realize the collaborative update of cross-regional decision-making models, improve global adaptability, and push decision instructions to vehicles and various infrastructures through a two-way communication link.

[0074] In this embodiment, the multi-level response mechanism includes:

[0075] Establish a mapping rule between early warning levels and response strategies. For low-risk warnings, push text warnings (such as "slippery road ahead") to surrounding vehicles and remind them to drive cautiously via in-vehicle navigation. For medium-risk warnings, simultaneously trigger roadside traffic light timing optimization (such as extending green lights to alleviate congestion) and send route suggestions to logistics fleets. For high-risk warnings, initiate cross-regional emergency response, push event details and handling plans to the traffic management center, and simultaneously enforce speed limit instructions (such as a speed limit of 40 km / h for the accident section) through two-way communication links.

[0076] In this embodiment, the construction of a multi-dimensional perception network realizes the fusion of multi-source data and spatiotemporal unification, constructs a vehicle-road-environment multi-dimensional fusion perception system, and enables dynamic traffic state analysis to identify abnormal events from multiple dimensions. Hierarchical feature extraction and multi-dimensional anomaly identification models enable parallel detection of accidents, congestion, and severe weather. Comprehensive risk prediction and assessment can accurately predict propagation paths and risks, solving the problem of one-sidedness in traditional risk assessment. It can predict the expansion path of the impact range of events. The multi-level response mechanism in early warning generation and decision-making can take differentiated measures according to the risk level. The application of multi-objective optimization models and MARL algorithms makes decision-making more scientific, and cross-regional collaborative updates improve global adaptability.

[0077] In this embodiment, the construction of the multidimensional sensing network further includes:

[0078] The vehicle-mounted V2X communication components are used to build bidirectional communication links between vehicles and between vehicles and infrastructure, and a connection is established with the traffic cloud platform based on the V2X communication components.

[0079] The vehicles send vehicle status data packets to adjacent vehicles at a preset period based on a two-way communication link. These packets include "beacon frames" containing information such as position, speed, and heading angle. Each roadside device sends traffic signal status, road construction / congestion, and other event information in real time, forming a dynamic communication network between vehicles and between vehicles and infrastructure.

[0080] Based on the IEEE 1588 clock synchronization protocol or the GPS / BeiDou satellite clock synchronization mechanism, time synchronization between vehicles and between vehicles and infrastructure is determined, achieving nanosecond-level time synchronization of all devices.

[0081] The vehicle acquires monitoring information of neighboring vehicles based on a two-way communication link, including the real-time location, movement status and intention information of neighboring vehicles, and also collects traffic signal status and road environment data actively uploaded by roadside infrastructure.

[0082] In this embodiment, dynamic traffic state analysis also includes constructing a four-dimensional semantic network topology of "road-vehicle-environment-event":

[0083] Using physical road entities (such as road segments and intersections) as the basic nodes of the topology, the structural framework of the road network topology map is constructed based on topological attributes (coordinate range, number of lanes, speed limit) and historical traffic characteristics (average daily flow and speed distribution).

[0084] The real-time trajectory of vehicles is mapped as dynamic edges, and the position transfer of vehicles in adjacent time slices is used as the connection relationship of the edges. Based on the behavioral features of edge attributes, such as speed, acceleration, and steering intention, an edge set reflecting the dynamic interaction of traffic flow is constructed.

[0085] Environmental parameters (precipitation intensity, visibility, and road surface temperature) are used as weighting factors in the graph structure and mapped to corresponding road nodes and vehicle behavior edges according to the spatiotemporal grid. This enables the quantitative representation of the impact of environmental factors on traffic conditions. Identified abnormal events are embedded in the four-dimensional semantic network topology, and node attributes include event type, scope of impact, and timeliness parameters.

[0086] Meanwhile, the semantic representation of traffic status is constructed by learning the relationships between nodes in the four-dimensional semantic network topology through graph neural networks (GNNs).

[0087] In this embodiment, a dynamic communication network is constructed based on V2X communication components. Nanosecond-level time synchronization is achieved by combining IEEE 1588 or satellite clock synchronization, overcoming the traditional communication delay and time deviation problems, significantly enhancing data real-time performance and collaboration. The four-dimensional semantic network topology of "road-vehicle-environment-event" is visualized to present traffic status, which can more accurately depict the dynamic interaction of traffic flow and the influence of environmental factors, more comprehensively identify abnormal events, improve the accuracy of understanding complex traffic scenarios, reduce the prediction error of event propagation paths, and provide a more intuitive and accurate semantic representation for risk assessment and decision-making.

[0088] In this embodiment, the hierarchical feature extraction for dynamic traffic state analysis further includes:

[0089] The road network topology map is spatially discretized. Based on the coordinate range of road physical entities and lane division, the road network is divided into spatiotemporal grid units containing attributes such as geographic coordinates, speed limits, and number of lanes. The basic traffic flow features (flow rate, occupancy rate, and average speed) in each spatiotemporal grid unit are extracted to form a spatial grid traffic state vector.

[0090] Extract the edge attributes formed by vehicle position transfers within adjacent time slices to generate a feature sequence characterizing the temporal changes in vehicle movement. At the same time, combine historical traffic data to identify the periodic patterns and sudden change characteristics of traffic flow and generate a temporal dynamic feature vector.

[0091] The spatial grid traffic state vector and the temporal dynamic feature vector are weighted and fused to form a composite feature representation that includes spatiotemporal context, highlighting the feature influence of key spatiotemporal areas (such as the real-time impact weight of accident road sections on the surrounding road network), and embedding the four-dimensional semantic network topology.

[0092] In this embodiment, the spatiotemporal grid unit division and feature fusion strategy discretizes the road network space and combines vehicle motion temporal characteristics and historical data patterns to construct a composite feature representation containing spatiotemporal context. This accurately captures the impact of key areas and embeds a four-dimensional semantic network topology, enabling deep fusion analysis of spatial distribution and temporal dynamic changes. This refines the granularity of traffic flow basic feature extraction to the lane level. Combined with historical data to identify patterns, it can significantly predict sudden changes in traffic flow in advance. The composite feature representation effectively improves the accuracy of the impact weight of key areas such as accident sections, enhancing the timeliness and accuracy of traffic state analysis.

[0093] In this embodiment, the comprehensive risk prediction and assessment involves correlation analysis, specifically including:

[0094] Acquire historical abnormal event data (such as the propagation patterns of the past 1000 accidents), combine it with multidimensional spatial data and environmental parameters in the four-dimensional semantic network topology, and use time-series data mining algorithms to extract the propagation patterns of different types of abnormal events (accidents, congestion, severe weather) under different environmental parameters (such as rainy days and nights), and construct a propagation rule knowledge base. The rule knowledge base includes the mapping relationship between event type, environmental conditions and propagation path.

[0095] The identified abnormal events are used as key nodes. Based on the relationship between nodes and edge attributes in the four-dimensional semantic network topology, the potentially affected road nodes and adjacent event nodes are selected based on the propagation rule knowledge base, and an event evolution tree is constructed.

[0096] The tree node attributes include event type, spatiotemporal location, current impact level, and severity. The tree node connection relationship attributes include event propagation direction, propagation probability, and time delay threshold (e.g., the probability of congestion propagating from road segment A to road segment B is 0.7, and the delay time is 5 minutes).

[0097] In this embodiment, the initial propagation probability of each edge is calculated by an association rule mining algorithm, and the time delay in the edge attribute is determined by combining the average propagation time of historical data statistics. At the same time, the current road network traffic state parameters (such as real-time traffic flow and vehicle speed distribution) and the basic traffic flow characteristics in the spatiotemporal grid unit are used as constraints to prune the branches of the evolution tree and remove low-probability propagation paths.

[0098] In this embodiment, predicting the propagation path specifically includes:

[0099] Using critical nodes of abnormal events as target nodes, multi-path sampling is performed on the connection relationship attributes in the event evolution tree to generate at least one propagation path. During each sampling, the corresponding branch path is selected based on the propagation probability distribution of the current target node.

[0100] By combining the propagation time distribution of similar historical events, a time series interpolation algorithm is used to obtain the time series of each target node in each propagation path that is affected, generating a set of propagation paths containing spatiotemporal dimensions;

[0101] The propagation probability distribution information of each propagation path in the propagation path set is mapped to spatiotemporal grid cells. The probability of each grid cell being affected in the future time period is accumulated and calculated to generate a risk probability cloud map for the future time period, which intuitively shows the probability of different regions being affected at different time points.

[0102] Identify key hub nodes in the propagation path set, calculate the risk contribution of key hub nodes in the propagation process (e.g., if a node is blocked, it will cause a 40% decrease in the road network capacity), prioritize each key hub node based on its risk contribution, and treat them as priority targets for subsequent response.

[0103] In this embodiment, identifying key hub nodes in the propagation path set specifically includes:

[0104] Extract the node attributes and topology information of each road node in the propagation path set in the four-dimensional semantic network topology;

[0105] In this embodiment, traffic flow corresponding to key time nodes is obtained from road node attributes. The key time nodes include the time when the abnormal event occurs and the period when similar events occur frequently in the past.

[0106] In this embodiment, topology information is extracted, including the number of road nodes that are associated with the road node, and the percentage change in traffic volume of associated road nodes when the traffic volume of the road node increases by 1%. This data is obtained by analyzing the historical propagation data of edge attributes in the event evolution tree.

[0107] Based on node attributes and topology information, the basic traffic characteristics of each road node and the traffic transmission sensitivity between road nodes are obtained, and the traffic weighting coefficient and traffic transmission weighting coefficient are determined.

[0108] In this embodiment, the average traffic flow of each road node at key time nodes is used as the basic traffic flow feature. Based on the percentage change in traffic flow of each associated road node, the average percentage change of associated road nodes corresponding to the traffic flow change of the road node is calculated to quantify the traffic flow transmission sensitivity between road nodes.

[0109] In this embodiment, the average traffic flow of each road node is compared with a preset traffic flow reference value (such as the average traffic flow of the road network or the design traffic flow of the road segment) to obtain a traffic flow weight coefficient, which reflects the relative importance of the traffic flow of the road node.

[0110] In this embodiment, the average percentage change of the associated road nodes of each road node is normalized to obtain the traffic transmission weight coefficient, which characterizes the ability of the road node to spread its influence on the surrounding road network.

[0111] Based on the preset law enforcement management information (such as daily law enforcement frequency and emergency response priority) for each road node, the law enforcement management weight of the road node is determined by a mapping function based on the management level corresponding to the law enforcement management method. Road nodes with high law enforcement priority are given higher weights, reflecting the correlation between management resource investment and the importance of road nodes.

[0112] The road topology weight value is obtained by comprehensively calculating the law enforcement management weight, traffic weight coefficient and traffic transmission weight coefficient. The road topology weight value is compared with the preset weight threshold, and road nodes with a weight value higher than the preset weight threshold are selected as key hub nodes.

[0113] In this embodiment, based on a four-dimensional semantic network and combining historical data with real-time parameters, a knowledge base of abnormal event propagation rules is established using a time-series mining algorithm. This effectively improves the accuracy of abnormal event propagation pattern recognition. Through event evolution trees, multi-path sampling, and pruning algorithms, dynamic generation and accurate prediction of propagation paths are achieved, significantly improving prediction efficiency. The system can intuitively display the spatiotemporal distribution of risks, providing visualized decision support for traffic management. Based on multi-dimensional weight coefficients, key hub nodes are evaluated, and the risk contribution of road nodes is quantified and ranked, providing a basis for accurate early warning and priority handling. This improves the efficiency of emergency resource allocation, increases the traffic capacity guarantee rate in key areas, effectively reduces the overall impact of abnormal events on the traffic system, and significantly enhances traffic risk prevention and control capabilities.

[0114] In this embodiment, before fusing the collected vehicle status, road environment, and infrastructure data, the method further includes: denoising the image data included in the road environment data;

[0115] The noise reduction of image data in road environment data includes:

[0116] Take any image from the road environment data as the image to be denoised;

[0117] The image to be denoised is processed into a grayscale image.

[0118] Take any pixel in the grayscale image as the first pixel; determine the target area with the first pixel as the center and a preset distance as the radius;

[0119] Calculate the difference between the first pixel and other pixels in the target area, and obtain several difference values; compare the several difference values ​​with a preset difference threshold, and mark the pixels with differences greater than or equal to the preset difference threshold as abnormal;

[0120] Iterate through all pixels in the grayscale image and count the number of outlier markers for each pixel.

[0121] The number of abnormal markers is compared with a preset abnormal marker threshold. Pixels with an abnormal marker number greater than or equal to the preset abnormal marker threshold are designated as first abnormal pixels, thus obtaining a number of first abnormal pixels.

[0122] Edge detection is performed on grayscale images to determine edge pixels in the grayscale images;

[0123] Subtracting the edge pixels from the first abnormal pixels yields a number of second abnormal pixels;

[0124] The dispersion of each second abnormal pixel is calculated and compared with a preset dispersion threshold. Pixels with dispersion greater than or equal to the preset dispersion threshold are designated as third abnormal pixels, resulting in a number of third abnormal pixels.

[0125] Several abnormal third pixels are deleted to obtain a noise-reduced grayscale image;

[0126] By iterating through all images in the road environment data, the denoised road environment data is obtained.

[0127] In this embodiment, the grayscale difference between the central pixel and other pixels within a certain range is calculated, and pixels with differences exceeding a threshold are marked as abnormal. Essentially, this captures pixels where the grayscale value changes significantly within a local area. Unlike existing technologies that directly identify abnormal pixels, which implies a final judgment on the pixel's nature (normal / abnormal), if it is subsequently found that the threshold setting is unreasonable or the area range is inappropriate, pixels that have been "determined" as abnormal will permanently lose their original attributes (such as grayscale value, spatial location, etc.), making it difficult to retrospectively adjust. "Marking" only adds a temporary label (such as "abnormal to be verified") to the pixel, retaining all the information of the original pixel. It can be re-evaluated by adjusting the threshold, expanding / shrinking the target area, etc., avoiding irreversible errors caused by initial parameter errors.

[0128] In this embodiment, edge recognition of the grayscale image is performed specifically by using a pre-trained edge recognition model.

[0129] In this embodiment, a number of first abnormal pixels are subtracted from the edge pixels to obtain a number of second abnormal pixels. Specifically, pixels that are duplicates of the edge pixels among the first abnormal pixels are deleted, and the remaining pixels are used as the second abnormal pixels.

[0130] In this embodiment, the specific implementation method for calculating the discreteness of each second abnormal pixel is as follows: the minimum Euclidean distance between each second abnormal pixel and the edge pixel is taken as the discreteness of each second abnormal pixel; the mean of the discreteness of all second abnormal pixels is taken as the preset discreteness threshold.

[0131] The working principle and beneficial effects of the above technical solution are as follows: First, an image is randomly selected from the road environment data as the image to be denoised, and then grayscale processing is performed on it to convert the color image into a grayscale image. The purpose of this is to simplify the image data and facilitate subsequent pixel analysis, because grayscale images only contain brightness information, reducing data complexity; for each pixel in the grayscale image (first, one is randomly selected as the first pixel), the target area is determined with the first pixel as the center and a preset distance as the radius; the difference between the first pixel and other pixels in the target area is calculated, and abnormal pixels are marked by comparing it with a preset difference threshold; this process is based on the brightness difference of pixels in the local area to initially determine which pixels may be noise points, because under normal circumstances, the brightness difference between adjacent pixels will not be too large; all pixels in the entire grayscale image are traversed, and the number of abnormal markings for each pixel is counted. The number of anomaly markers is compared with a preset anomaly marker threshold to identify the first anomaly pixel. This step further filters out pixels that may be noise, because if a pixel is marked as anomaly in multiple local areas, it is more likely to be a noise point. Edge recognition is performed on the grayscale image to identify edge pixels. The first anomaly pixel is subtracted from the edge pixel to obtain the second anomaly pixel, because edge pixels themselves have large brightness variations and may be misjudged as noise points. This method can eliminate the interference of edge pixels. The dispersion of each second anomaly pixel is calculated, and the dispersion reflects the distribution characteristics of the pixel in the local area. The dispersion is compared with a preset dispersion threshold to identify and delete the third anomaly pixel, thus obtaining the denoised grayscale image. Finally, all images in the road environment data are traversed to complete the denoising of the entire road environment data image portion. Through a detailed pixel analysis and filtering process, noise points in the image can be identified and removed relatively accurately. Compared with some simple denoising methods, it considers multiple factors such as the differences of pixels in local areas, the accumulation of anomaly markers, and the special cases of edge pixels, so as to better preserve the effective information of the image while removing noise.

[0132] In this embodiment, a comprehensive risk index for the abnormal event is calculated based on its severity and spread risk, including steps 1-2:

[0133] Step 1: Obtain the average risk value of vehicle a in historical driving data and the propagation risk value of abnormal events on the driving path of vehicle a. Based on the average risk value, the propagation risk value and the expected risk value of abnormal events, determine the comprehensive risk index of abnormal events on vehicle a.

[0134] ;

[0135] in, This represents the overall risk index of the abnormal event to vehicle a; n represents the total number of risk factors in the abnormal event. This represents the average risk value of vehicle a in historical driving data; This represents the risk value of the abnormal event propagating along the travel path of vehicle a. This represents the risk level value of the i-th type of risk factor among n types of risk factors; Let represent the probability density function of the occurrence of risk of the i-th type of risk factor; It represents the cumulative probability of the occurrence of the i-th type of risk factor within the next unit time interval starting from time t; This represents the total cumulative probability of the occurrence of the i-th type of risk factor over all future time periods, starting from time t.

[0136] Step 2: Determine the comprehensive risk index of the abnormal event based on the comprehensive risk index of the abnormal event for vehicle a and the total number of vehicles in the driving path of vehicle a.

[0137] ;

[0138] in, A comprehensive risk index representing abnormal events; This represents the total number of vehicles in the path that vehicle a is traveling on.

[0139] In this embodiment, the propagation probability distribution information of each propagation path in the propagation path set can determine the propagation risk value of the driving path of vehicle a.

[0140] The working principle and beneficial effects of the above technical solution are as follows: It comprehensively considers multiple factors, including the vehicle's historical risk value, the risk of abnormal events propagating along the travel path, the risk level of various risk factors, and the probability of various risk factors occurring. This multi-dimensional assessment can more comprehensively reflect the risk status of abnormal events and avoid the one-sidedness of single-factor assessment. By introducing the time factor through the time integral in the probability density function, the probability of risk factors occurring can be assessed according to different time intervals, thereby better adapting to the situation where risks change over time. First, the comprehensive risk index of a single vehicle is calculated, and then the comprehensive risk index (T) of the abnormal event is obtained through the comprehensive calculation of all vehicles. It can assess the risk of a single vehicle in an abnormal event and also grasp the risk impact of the abnormal event on the entire group of vehicles along the travel path, which helps to formulate targeted risk management strategies.

[0141] The above description is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any equivalent substitutions or modifications made by those skilled in the art within the scope of the technology disclosed in the present invention, based on the technical solution and inventive concept of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A method for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation, characterized in that, include: Multidimensional sensing network construction: Construct a multidimensional sensing network to acquire vehicle status, road environment and infrastructure data in real time, and fuse the collected vehicle status, road environment and infrastructure data to generate multidimensional spatial data; Before fusing the collected vehicle status, road environment and infrastructure data, the process also includes denoising the image data included in the road environment data, traversing all the images in the road environment data to obtain the denoised road environment data. Dynamic traffic state analysis: hierarchical feature extraction is performed on the fused multidimensional spatial data to construct a dynamic traffic state model, and the degree of deviation of traffic flow parameters from the normal distribution is observed in real time. Combined with historical observation data and real-time observation values, abnormal events are identified in multiple dimensions. Comprehensive risk prediction and assessment: Based on the identification results, correlation analysis is performed on the identified abnormal events to predict the propagation path. According to the degree of harm and propagation risk of the abnormal events, the comprehensive risk index of the abnormal events is calculated. Based on the severity and spread risk of the abnormal event, calculate the comprehensive risk index of the abnormal event, including steps 1-2: Step 1: Obtain the average risk value of vehicle a in historical driving data and the propagation risk value of abnormal events on the driving path of vehicle a. Based on the average risk value, the propagation risk value and the expected risk value of abnormal events, determine the comprehensive risk index of abnormal events on vehicle a. ； in, This represents the overall risk index of the abnormal event to vehicle a; n represents the total number of risk factors in the abnormal event. This represents the average risk value of vehicle a in historical driving data; This represents the risk value of the abnormal event propagating along the travel path of vehicle a. This represents the risk level value of the i-th type of risk factor among n types of risk factors; Let represent the probability density function of the occurrence of risk of the i-th type of risk factor; This represents the cumulative probability of the occurrence of the i-th type of risk factor within the next unit time interval starting from time t; This represents the total cumulative probability of the occurrence of the i-th type of risk factor over all future time periods, starting from time t. Step 2: Determine the comprehensive risk index of the abnormal event based on the comprehensive risk index of the abnormal event for vehicle a and the total number of vehicles in the driving path of vehicle a. ； in, A comprehensive risk index representing abnormal events; This represents the total number of vehicles along the path that vehicle a is traveling on; Early warning generation and decision-making: Based on the comprehensive risk index, a multi-level response mechanism is triggered, and corresponding early warning information is generated. A decision-making plan is established, and the safety of the decision-making plan is verified. The decision-making instructions are pushed to vehicles and various infrastructures through a two-way communication link.

2. The method for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation as described in claim 1, characterized in that, The construction of multidimensional sensing networks also includes: The vehicle-mounted V2X communication components are used to build bidirectional communication links between vehicles and between vehicles and infrastructure, and a connection is established with the traffic cloud platform based on the V2X communication components. The vehicles send vehicle status data packets to adjacent vehicles at a preset period based on a two-way communication link, and each roadside device sends event information in real time, forming a dynamic communication network between vehicles and between vehicles and each infrastructure. Time synchronization between vehicles and between vehicles and infrastructure is determined based on a clock synchronization mechanism. The vehicle acquires monitoring information of neighboring vehicles based on a two-way communication link, including the real-time location, movement status and intention information of neighboring vehicles, and also collects traffic signal status and road environment data actively uploaded by roadside infrastructure.

3. The method for monitoring, early warning, and decision analysis of transportation operations based on vehicle-road cooperation as described in claim 2, characterized in that, Dynamic traffic state analysis also includes constructing a four-dimensional semantic network topology: Using physical road entities as topological foundation nodes, the structural skeleton of the road network topology graph is constructed based on topological attributes and historical traffic characteristics. The real-time trajectory of the vehicle is mapped as a dynamic edge, and the vehicle position transfer within adjacent time slices is used as the connection relationship of the edge. Based on the edge attribute behavior characteristics, an edge set is constructed. Environmental parameters are used as weighting factors in the graph structure, mapped to corresponding road nodes and vehicle behavior edges according to the spatiotemporal grid, and the identified abnormal events are embedded in the four-dimensional semantic network topology. Simultaneously, the relationships between nodes in the four-dimensional semantic network topology are learned to construct a semantic representation of traffic status.

4. The transportation operation monitoring, early warning, and decision analysis method based on vehicle-road cooperation as described in claim 3, characterized in that, The hierarchical feature extraction for dynamic traffic state analysis also includes: The road network topology map is spatially discretized. Based on the coordinate range of road physical entities and lane division, the road network is divided into spatiotemporal grid units. The basic traffic flow features in each spatiotemporal grid unit are extracted to form a spatial grid traffic state vector. Extract the edge attributes formed by vehicle position transfers within adjacent time slices to generate a feature sequence characterizing the temporal changes in vehicle movement. At the same time, combine historical traffic data to identify the periodic patterns and sudden change characteristics of traffic flow and generate a temporal dynamic feature vector. The spatial grid traffic state vector and the temporal dynamic feature vector are weighted and fused to form a composite feature representation, which is then embedded into the four-dimensional semantic network topology.

5. The transportation operation monitoring, early warning, and decision analysis method based on vehicle-road cooperation as described in claim 4, characterized in that, The comprehensive risk prediction and assessment involves correlation analysis, specifically including: Historical abnormal event data is acquired, and combined with multidimensional spatial data and environmental parameters in the four-dimensional semantic network topology, the propagation patterns of different types of abnormal events under different environmental parameters are extracted, and a propagation rule knowledge base is constructed. The identified abnormal events are used as key nodes. Based on the relationship between nodes and edge attributes in the four-dimensional semantic network topology, the potentially affected road nodes and adjacent event nodes are selected based on the propagation rule knowledge base, and an event evolution tree is constructed. The tree node attributes include event type, spatiotemporal location, current impact level, and severity. The tree node connection relationship attributes include event propagation direction, propagation probability, and time delay threshold.

6. The transportation operation monitoring, early warning, and decision analysis method based on vehicle-road cooperation as described in claim 5, characterized in that, Predicting propagation paths, specifically including: Using critical nodes of abnormal events as target nodes, multi-path sampling is performed on the connection relationship attributes in the event evolution tree to generate at least one propagation path. During each sampling, the corresponding branch path is selected based on the propagation probability distribution of the current target node. By combining the propagation time distribution of similar historical events, the time sequence of each target node affected in each propagation path is obtained, and a set of propagation paths is generated; The propagation probability distribution information of each propagation path in the propagation path set is mapped to spatiotemporal grid cells, and the probability of being affected in each grid cell in the future time period is accumulated and calculated to generate a risk probability cloud map for the future time period. Identify key hub nodes in the propagation path set, calculate the risk contribution of each key hub node in the propagation process, and prioritize each key hub node based on its risk contribution.

7. The transportation operation monitoring, early warning, and decision analysis method based on vehicle-road cooperation as described in claim 6, characterized in that, Identify key hub nodes in the propagation path set, specifically including: Extract the node attributes and topology information of each road node in the propagation path set in the four-dimensional semantic network topology; Based on node attributes and topology information, the basic traffic characteristics of each road node and the traffic transmission sensitivity between road nodes are obtained, and the traffic weighting coefficient and traffic transmission weighting coefficient are determined. Based on the preset law enforcement management method information corresponding to each road node, the law enforcement management weight of the road node is determined according to the management level corresponding to the law enforcement management method. The road topology weight value is obtained by comprehensively calculating the law enforcement management weight, traffic weight coefficient and traffic transmission weight coefficient. The road topology weight value is compared with the preset weight threshold, and road nodes with a weight value higher than the preset weight threshold are selected as key hub nodes.

8. The transportation operation monitoring, early warning, and decision analysis method based on vehicle-road cooperation as described in claim 7, characterized in that, A multi-level response mechanism, including: Establish a mapping rule between early warning levels and response strategies. For example, if the early warning level is low risk, push a text warning to surrounding vehicles and remind them to drive cautiously via in-vehicle navigation; if the early warning level is medium risk, simultaneously trigger roadside traffic light timing optimization and send route suggestions to logistics fleets; if the early warning level is high risk, initiate cross-regional emergency response, push event details and handling plans to the traffic management center, and simultaneously enforce speed limit instructions through two-way communication links.

9. The transportation operation monitoring, early warning, and decision analysis method based on vehicle-road cooperation as described in claim 8, characterized in that, The noise reduction of image data in road environment data includes: Take any image from the road environment data as the image to be denoised; The image to be denoised is processed into a grayscale image. Take any pixel in the grayscale image as the first pixel; determine the target area with the first pixel as the center and a preset distance as the radius; Calculate the difference between the first pixel and other pixels in the target area, and obtain several difference values; compare the several difference values ​​with a preset difference threshold, and mark the pixels with differences greater than or equal to the preset difference threshold as abnormal; Iterate through all pixels in the grayscale image and count the number of outlier markers for each pixel. The number of abnormal markers is compared with a preset abnormal marker threshold. Pixels with an abnormal marker number greater than or equal to the preset abnormal marker threshold are designated as first abnormal pixels, thus obtaining a number of first abnormal pixels. Edge detection is performed on grayscale images to determine edge pixels in the grayscale images; Subtracting the edge pixels from the first abnormal pixels yields a number of second abnormal pixels; The dispersion of each second abnormal pixel is calculated and compared with a preset dispersion threshold. Pixels with dispersion greater than or equal to the preset dispersion threshold are designated as third abnormal pixels, resulting in a number of third abnormal pixels. Several abnormal third pixels are deleted to obtain a noise-reduced grayscale image; By iterating through all images in the road environment data, the denoised road environment data is obtained.

Images