Urban comprehensive traffic network toughness real-time prediction method combined with Internet of Things perception

Abstract

The invention discloses an urban comprehensive traffic network toughness real-time prediction method combined with Internet of Things perception, and relates to the technical field of Internet of Things, and the method comprises the steps: carrying out the information collection of a traffic key node of a target city, and obtaining the traffic monitoring information; according to the scene attribute information of the target city in the preset time zone, evaluating and obtaining the index weight of the toughness evaluation index; performing data sensitivity evaluation on the plurality of monitoring data types based on meteorological fluctuation characteristics and index weights of a target city in a preset time zone, and setting a to-be-analyzed data type set; performing data screening and extraction on the traffic monitoring information at the edge processing unit according to the to-be-analyzed data type set to obtain a to-be-analyzed monitoring information set; and inputting the to-be-analyzed monitoring information set into the traffic network toughness prediction channel, and outputting the predicted traffic network toughness index of the target city in the preset time zone. The problem that a traffic network toughness prediction method in the prior art cannot accurately predict traffic toughness in limited time is solved.

Description

Technical Field

[0001] This application relates to the field of Internet of Things (IoT) technology, specifically to a method for real-time prediction of the resilience of urban integrated transportation networks that incorporates IoT sensing. Background Technology

[0002] With urban development and population growth, urban transportation networks face increasing challenges, such as traffic congestion and frequent accidents. However, traditional methods for predicting traffic network resilience are computationally inefficient when dealing with massive, multi-source, and heterogeneous IoT data streams, making it difficult to accurately predict traffic resilience within a limited timeframe and thus failing to meet real-time requirements. Summary of the Invention

[0003] This application provides a method for real-time prediction of urban integrated transportation network resilience by combining IoT sensing, which solves the technical problem that existing transportation network resilience prediction methods cannot accurately predict traffic resilience within a limited time.

[0004] The technical solution to the above-mentioned technical problems in this application is as follows: Firstly, this application provides a method for real-time prediction of the resilience of urban integrated transportation networks combined with Internet of Things (IoT) sensing, the method comprising: IoT sensing devices are deployed at several key traffic nodes in the target city to collect traffic information and obtain traffic monitoring information. Each key traffic node includes an edge processing unit. Based on the scenario attribute information of the target city within the preset time zone, several indicator weights of several resilience assessment indicators are obtained. Based on the meteorological fluctuation characteristics of the target city within the preset time zone and several indicator weights, data sensitivity evaluation is performed on multiple monitoring data types, and a set of data types to be analyzed is set according to multiple data sensitivities. According to the data set to be analyzed, valuable data is filtered and extracted from the traffic monitoring information in several edge processing units to obtain the monitoring information set to be analyzed. The set of monitoring information to be analyzed is input into the traffic network resilience prediction channel, and the predicted traffic network resilience index of the target city in the preset time zone is output.

[0005] This application provides several technical solutions, which have at least the following technical effects or advantages: This application provides a real-time prediction method for the resilience of urban integrated transportation networks by combining IoT sensing. First, IoT sensing devices collect information at key traffic nodes, and then combine this information with scene attribute information to evaluate and obtain resilience assessment index weights. Next, meteorological fluctuation characteristics are considered to evaluate data sensitivity and filter data. Finally, the prediction results are output through a trained transportation network resilience prediction channel. This method effectively improves the computational efficiency of transportation network resilience prediction, can accurately predict traffic resilience within a limited time, better meets the real-time requirements of urban traffic, and provides strong support for the management and optimization of urban integrated transportation networks.

[0006] The above technical solution involves deploying IoT sensing devices at key urban traffic nodes to collect multi-dimensional data in real time, including traffic flow, road conditions, and facility operational status. Edge computing technology is used to preprocess the collected real-time data, identifying core indicators related to network resilience, such as road segment capacity, transfer efficiency, and emergency response speed. Furthermore, a deep learning prediction model integrating spatiotemporal features is constructed and trained to predict the resilience of the traffic network under disturbances such as sudden weather events and accidents. This provides traffic management departments with accurate early warnings and decision support, effectively improving the anti-interference and recovery efficiency of the urban transportation system. Attached Figure Description

[0007] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0008] Figure 1 This is a flowchart illustrating the real-time prediction method for the resilience of urban integrated transportation networks combined with IoT sensing provided in an embodiment of this application. Figure 2 This is a flowchart illustrating the process of obtaining the data set of data to be analyzed in the real-time prediction method for the resilience of urban integrated transportation networks combined with Internet of Things sensing, as provided in the embodiments of this application. Detailed Implementation

[0009] This application provides a real-time prediction method for the resilience of urban integrated transportation networks that incorporates Internet of Things (IoT) sensing, addressing the technical problem that existing transportation network resilience prediction methods cannot accurately predict traffic resilience within a limited timeframe.

[0010] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0011] In the description of this application, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the stated features. In the description of this application, "multiple" means two or more, unless otherwise explicitly specified.

[0012] In the description of this application, the term "for example" is used to mean "used as an example, illustration, or description." Any embodiment described as "for example" in this application is not necessarily to be construed as being more preferred or advantageous than other embodiments. The following description is provided to enable any person skilled in the art to make and use this application. Details are set forth in the following description for purposes of explanation. It should be understood that those skilled in the art will recognize that this application can be made without using these specific details. In other instances, well-known structures and processes will not be described in detail to avoid unnecessarily obscuring the description of this application. Therefore, this application is not intended to be limited to the embodiments shown, but is consistent with the broadest scope of the principles and features disclosed in this application.

[0013] Example 1, as Figure 1 As shown in the embodiments of this application, a method for real-time prediction of the resilience of urban integrated transportation networks combined with Internet of Things (IoT) sensing is provided, including: S10: Deploy IoT sensing devices at several key traffic nodes in the target city to collect traffic information and obtain traffic monitoring information. Each key traffic node includes an edge processing unit. In this embodiment, IoT sensing devices are first deployed at several key traffic nodes in the target city to collect traffic information and obtain traffic monitoring data. Key traffic nodes are important hubs in the urban traffic network, such as major intersections, bridges, tunnels, and traffic hubs. The traffic conditions at key traffic nodes affect the operation of the entire urban traffic network. Each key traffic node contains an edge processing unit.

[0014] The edge processing unit performs preliminary processing and analysis on the collected raw data, reducing the amount of data transmission and improving data processing efficiency. At the same time, it can also reduce latency and risks during data transmission to a certain extent.

[0015] By deploying IoT sensing devices at key traffic nodes, comprehensive and real-time traffic monitoring information can be collected, providing a rich data foundation for subsequent traffic network resilience prediction.

[0016] Specifically, step S10 of this application includes: Deploy IoT sensing devices at several key traffic nodes in the target city to build a traffic monitoring network. The IoT sensing devices include at least high-definition cameras, LiDAR, RFID readers, weather sensors, and road condition sensors. The traffic monitoring network continuously monitors the traffic status of several key traffic nodes and obtains traffic monitoring information, which includes basic traffic flow data sequences, traffic status data sequences, environmental status data sequences, and vehicle characteristic data sequences within a historical time range.

[0017] In this embodiment of the application, IoT sensing devices are deployed at several key traffic nodes in the target city to build a traffic monitoring network. The IoT sensing devices should include at least high-definition cameras, LiDAR, RFID readers, weather sensors, and road condition sensors.

[0018] High-definition cameras capture images and videos of traffic scenes to identify vehicles, pedestrians, traffic signs, etc.; LiDAR obtains real-time distance and shape information of surrounding objects to help judge traffic conditions; RFID readers can identify and collect information from vehicles equipped with electronic tags; meteorological sensors collect meteorological data such as temperature, humidity, wind speed, and precipitation; road surface condition sensors can detect road surface friction, water accumulation, etc., which helps assess driving safety.

[0019] For example, RFID readers are installed at intersections to monitor vehicle flow, speed, and traffic congestion; meteorological sensors and road surface condition sensors are deployed on bridges, tunnels, and roads to detect the health status of traffic facilities and equipment and environmental parameters; and high-definition cameras are installed at road intersections and traffic hubs to obtain information such as the entry and exit times of traffic vehicles and the number of passengers getting on and off vehicles.

[0020] By continuously monitoring the traffic conditions at several key traffic nodes, the acquired traffic monitoring information encompasses multiple aspects. For example, the basic traffic flow data sequence within a historical timeframe records fundamental information such as traffic volume in each direction, average vehicle speed, time occupancy, and vehicle density, reflecting the basic traffic flow situation; the traffic status data sequence includes congestion levels, queue lengths, and automatic identification and alarms for traffic accidents and violations, reflecting the operational status of the traffic network; the environmental status data sequence covers information such as real-time precipitation, wind speed, visibility, road surface temperature, and conditions such as dryness / wetness, icing, and water accumulation; and the vehicle characteristic data sequence includes vehicle type classification, license plate number recognition, and continuous trajectory tracking.

[0021] By using the above information collection methods, real-time traffic information of key traffic nodes in the target city can be obtained, providing comprehensive and accurate data support for subsequent analysis and prediction.

[0022] S20: Based on the scene attribute information of the target city within the preset time zone, evaluate and obtain the weights of several resilience assessment indicators. In this embodiment, the weights of several resilience assessment indicators are evaluated based on the scenario attribute information of the target city within a preset time zone. The scenario attribute information covers various situations that may occur in the target city within the preset time zone, such as differences in traffic flow between weekdays and weekends, travel peaks during special holidays, and population movement during large-scale events. The importance of each resilience assessment indicator varies depending on the scenario.

[0023] Specifically, step S20 of this application includes: Obtain scene attribute information of a target city within a preset time zone, wherein the scene attribute information includes time, day type, season, and climate conditions; Several resilience assessment indicators were obtained, including the stability of traffic facilities and equipment, traffic network connectivity, total travel time loss, effective redundant path ratio, speed maintenance rate, average recovery time after an accident and planned time index. Based on historical traffic monitoring data of the target city, the relative importance of the several resilience assessment indicators is evaluated according to the scenario attribute information, and the weights of several indicators are output.

[0024] In this embodiment, scene attribute information of a target city within a preset time zone is obtained, including time, day type, season, and climate conditions. Scene attribute information can have varying degrees of impact on the resilience of the urban transportation network. By analyzing and evaluating the scene attribute information, the importance of different resilience assessment indicators under different scenarios is determined.

[0025] Traffic facility stability refers to the ability of traffic infrastructure and related equipment to maintain normal operation under various conditions, such as the integrity and reliability of roads, bridges, and traffic lights; traffic network connectivity measures the tightness of connection and accessibility between nodes in an urban traffic network; total travel time loss reflects the increase in vehicle travel time caused by traffic congestion or emergencies; effective redundancy ratio reflects the abundance of alternative routes available in the traffic network; speed maintenance rate indicates the proportion of vehicles that can maintain their expected speed during travel; mean time to recovery after an accident measures the average time required for traffic to return to normal operation after an accident; and the planned time index reflects how close travelers' actual travel time is to their planned travel time.

[0026] Furthermore, the stability of transportation facilities and equipment reflects the aging degree of transportation facilities and equipment and the impact of their failure probability on the resilience of the transportation network; the connectivity of the transportation network reflects the overall efficiency of the network and the convenience of residents' travel; the total travel time loss reflects the degree of impact of traffic congestion on travel time, and the larger the total travel time loss value, the lower the operating efficiency of the transportation network; the effective redundant path ratio reflects the alternative paths available to the transportation network when facing local failures or congestion, and the higher the ratio, the stronger the reliability of the transportation network; the speed maintenance rate reflects the ability of vehicles to maintain normal speed during travel, and can intuitively reflect the smoothness of traffic; the average accident recovery time measures the ability of the transportation network to resume normal operation after an accident, and the shorter the time, the stronger the recovery ability of the transportation network; the planned time index comprehensively considers the reliability and stability of travel time.

[0027] For example, during morning and evening rush hours, traffic congestion is more likely to occur due to high traffic volume, so total travel time loss and speed maintenance rate are more important; since severe weather increases the probability of traffic accidents, such as heavy rain and snow, the average recovery time and planned time index of accidents have greater weights.

[0028] Based on historical traffic monitoring data of the target city, the relative importance of resilience assessment indicators is evaluated according to scenario attribute information. For example, the analytic hierarchy process (AHP) is used to output the weights of several indicators.

[0029] For example, during weekday morning rush hour, the weight of the total travel time loss indicator is 0.3, the effective redundant path ratio is 0.2, the speed maintenance rate is 0.25, the average accident recovery time is 0.15, and the planned time index is 0.1. However, during weekend daytime, the weights of these indicators change: the weight of the total travel time loss indicator is 0.2, the effective redundant path ratio is 0.25, the speed maintenance rate is 0.2, the average accident recovery time is 0.1, and the planned time index is 0.25. The indicator weights obtained based on different scenario attribute information can accurately reflect the degree of impact of each resilience assessment indicator on the resilience of the transportation network under different conditions. After collecting traffic monitoring information, several indicator weights for several resilience assessment indicators are obtained based on the scenario attribute information of the target city within a preset time zone.

[0030] Furthermore, using the analytic hierarchy process (AHP), a judgment matrix is ​​first constructed. By comparing the relative importance of each indicator, the element values ​​of the judgment matrix are determined. Then, the largest eigenvalue and the corresponding eigenvector of the judgment matrix are calculated, and after normalization, the weights of each indicator are obtained. The normalized weights ensure that the sum of the weights of each indicator is 1, thereby guaranteeing the scientific and reasonable nature of the assessment. Based on the indicator weights, the impact of each indicator under different scenarios can be more accurately considered when predicting the resilience of the transportation network.

[0031] S30: Based on the meteorological fluctuation characteristics of the target city within the preset time zone and several indicator weights, perform data sensitivity evaluation on multiple monitoring data types respectively, and set the data type set to be analyzed according to multiple data sensitivities; In this embodiment, firstly, based on the meteorological fluctuation characteristics of the target city within a preset time zone and several indicator weights, the data sensitivity of multiple monitoring data types is evaluated. Meteorological fluctuation characteristics can have varying degrees of impact on transportation networks. For example, heavy rain can cause slippery roads, affecting vehicle speed and consequently traffic flow and congestion; strong winds may threaten the driving safety of large vehicles, increasing the probability of traffic accidents. Different monitoring data types have different sensitivities to meteorological fluctuations. Combining this with the previously obtained indicator weights allows for a more accurate evaluation of the sensitivity of each monitoring data type.

[0032] For example, for basic traffic flow data sequences, traffic volume, average vehicle speed, and other data may change significantly under severe weather conditions, making this data sequence quite sensitive to weather fluctuations. However, for vehicle characteristic data sequences, such as vehicle type classification and license plate number recognition, the impact of weather fluctuations is relatively small.

[0033] When evaluating data sensitivity, considering indicator weights can highlight the importance of data types that are highly correlated with traffic network resilience. For example, if the indicator weight for speed maintenance rate is high in a certain scenario, then the sensitivity evaluation of monitoring data types related to speed will be relatively high.

[0034] Secondly, the data types to be analyzed are defined based on multiple data sensitivities. Data types with high sensitivity are included in the data types to be analyzed; while data types with low sensitivity, if they have little impact on the resilience of the transportation network, are not included in the data types to be analyzed, in order to reduce the workload of data processing and improve computational efficiency.

[0035] For example, during heavy rain, data types related to road conditions and vehicle speed are included in the set of data types to be analyzed, while some vehicle feature data types that are not sensitive to weather fluctuations are excluded.

[0036] By evaluating data sensitivity and defining the data set of the types to be analyzed, data that is of great significance to the prediction of traffic network resilience can be selected more effectively, providing more effective data support for subsequent prediction work, further improving the accuracy and efficiency of traffic network resilience prediction, and better meeting the real-time requirements of urban traffic.

[0037] Specifically, step S30 in this application includes: Obtain predicted meteorological information for a target city within a preset time zone, wherein the predicted meteorological information includes at least precipitation information, visibility information, wind information, and temperature information; The volatility of precipitation, visibility, wind, and temperature information was assessed and weighted to obtain the meteorological volatility. Based on the predicted meteorological information, the data acquisition complexity, data analysis complexity, and data reliability of the multiple monitoring data types are evaluated respectively, and multiple data processing complexities are obtained by weighted calculation. The data processing complexity is positively correlated with the data acquisition complexity and data analysis complexity, and negatively correlated with the data reliability. Based on historical traffic monitoring data of the target city, the correlation sets of multiple types between the several resilience assessment indicators and the multiple monitoring data types are analyzed and obtained. The correlation sets of multiple types are then weighted and fused according to the weights of the several indicators to output the comprehensive correlation of multiple types. Based on the meteorological fluctuation, multiple data sensitivities of the multiple monitoring data types are obtained by weighted evaluation according to the multiple data processing complexities and multiple types of comprehensive correlation. Among them, data sensitivity is negatively correlated with data processing complexity and positively correlated with type comprehensive correlation.

[0038] In this embodiment of the application, firstly, the predicted meteorological information of the target city within the preset time zone is obtained, including precipitation information, visibility information, wind information, temperature information, and natural disaster information.

[0039] Furthermore, precipitation information reflects whether there is rain or snow. Different amounts of precipitation will have different effects on traffic. For example, light rain may only make the road surface slightly slippery, while heavy rain may lead to road flooding and traffic paralysis. Visibility information assesses the visibility of vehicles. Low visibility increases the probability of traffic accidents. Wind information indicates that strong winds can affect the driving stability of large vehicles. Temperature information can affect road conditions. For example, low temperatures may cause the road surface to freeze. Natural disaster information refers to the damage to the transportation network caused by extreme disasters such as earthquakes and floods.

[0040] Secondly, volatility assessments are conducted on precipitation, visibility, wind, and temperature information, respectively. This involves analyzing the amplitude and frequency of changes in meteorological elements within a preset time zone, obtained by comparing the standard deviation to the mean of the parameters in the parameter series. For example, does the intensity of precipitation suddenly increase or decrease, or does visibility drop sharply in a short period? The assessment results are then weighted to calculate meteorological volatility, and through appropriate weight allocation, the overall volatility of the weather is comprehensively reflected.

[0041] For example, the volatility of outdoor temperature information is assessed by obtaining the ratio of the standard deviation to the mean of the parameters in the parameter series. The total historical maximum temperature of a target area over seven days is 22+18+15+25+20+16+24=140℃, the temperature mean is 140÷7=20℃, the sum of squared deviations is 4+4+25+25+0+16+16=90℃², and the standard deviation is approximately 3.87℃. The volatility is calculated by dividing the standard deviation by the mean, and then converted to a percentage: Volatility = (3.87 / 20)×100%≈19.35%.

[0042] Next, based on predicted meteorological information, the data acquisition complexity, data analysis complexity, and data reliability of multiple monitoring data types were evaluated. Regarding data acquisition complexity, for example, in heavy rain, data acquisition from road surface sensors may become difficult due to water accumulation, resulting in higher complexity; similarly, high-definition cameras experience reduced image clarity in low visibility conditions, increasing the difficulty of data acquisition. In terms of data analysis complexity, when weather conditions are complex, the patterns of change in traffic flow data such as vehicle volume and speed in the basic traffic flow data sequence may be more difficult to analyze. Regarding data reliability, severe weather conditions may cause deviations in some sensor data, reducing data reliability. The weighted calculation of these three evaluation results yielded multiple data processing complexities.

[0043] Furthermore, when extreme weather or natural disasters occur, data collection becomes more difficult, so we should consider increasing the weight of data collection complexity.

[0044] For example, data acquisition complexity is weighted at 0.4, data analysis complexity at 0.3, and data reliability at 0.3. The data processing complexity score ranges from 1 to 10, with higher scores indicating greater data processing difficulty. In the event of extreme weather or natural disasters, data acquisition complexity is weighted at 0.5, data analysis complexity at 0.3, and data reliability at 0.2. Since data processing complexity is positively correlated with data acquisition complexity and data analysis complexity, and negatively correlated with data reliability, when a certain monitoring data type has a data acquisition complexity score of 7, a data analysis complexity score of 6, and a data reliability score of 5, the data processing complexity for that monitoring data type is 7 × 0.4 + 6 × 0.3 + (10 - 5) × 0.3 = 6.1.

[0045] Next, based on historical traffic monitoring data of the target city, several resilience assessment indicators and multiple types of monitoring data are analyzed to obtain a set of correlation degrees for various types. For example, total travel time loss has a high correlation with the basic traffic flow data sequence, while the average accident recovery time has a high correlation with the automatic traffic accident identification alarm information in the traffic status data sequence. Then, the multiple types of correlation degree sets are weighted and fused according to the weights of several indicators to output a comprehensive correlation degree for multiple types.

[0046] For example, the total travel time loss index has a weight of 0.3 and a correlation degree of 0.8 with the basic traffic flow data sequence; the effective redundant path ratio index has a weight of 0.2 and a correlation degree of 0.6 with the basic traffic flow data sequence. Then the comprehensive correlation degree of the basic traffic flow data sequence is 0.3×0.8+0.2×0.6=0.36.

[0047] Finally, based on meteorological fluctuations, multiple data sensitivities for various monitoring data types are obtained through weighted evaluation based on multiple data processing complexities and the comprehensive correlation of multiple types. Since data sensitivity is negatively correlated with data processing complexity and positively correlated with the comprehensive correlation of types, the calculation formula is set as "Data Sensitivity = Comprehensive Correlation of Types × Meteorological Fluctuations / Data Processing Complexity".

[0048] Among them, based on the meteorological fluctuation, multiple data sensitivities of the multiple monitored data types are obtained through weighted evaluation according to the multiple data processing complexities and the comprehensive correlation of multiple types, including: The ratio of the meteorological fluctuation to the average historical meteorological fluctuation of the target city within the historical time period is set as the complexity weight adjustment coefficient, which is multiplied by the initial complexity weight to obtain the adaptive complexity weight. The initial complexity weight is 0.4, and the adaptive complexity weight is greater than or equal to 0.2 and less than or equal to 0.6. The adaptation complexity weight is obtained by subtracting the adaptation relevance weight from 1; Based on the adaptation complexity weight and adaptation relevance weight, multiple data sensitivities are obtained by weighted evaluation according to the multiple data processing complexities and the comprehensive relevance of multiple types.

[0049] In this embodiment, firstly, the complexity weight adjustment coefficient is calculated by comparing the meteorological fluctuation of the target city within a preset time zone with the historical average meteorological fluctuation of the target city over a historical time period. This comparison is set as the complexity weight adjustment coefficient. The complexity weight adjustment coefficient reflects the current meteorological fluctuation relative to historical meteorological fluctuations. If the meteorological fluctuation is relatively large compared to the historical average, the complexity weight adjustment coefficient is greater than 1; if the meteorological fluctuation is relatively small compared to the historical average, the complexity weight adjustment coefficient will be less than 1. The greater the meteorological fluctuation, the higher the timeliness of the analysis, thus requiring faster prediction results, in which case the complexity weight accounts for a larger proportion.

[0050] After obtaining the complexity weight adjustment coefficient, multiply the complexity weight adjustment coefficient by the initial complexity weight of 0.4 to obtain the adaptive complexity weight. The adaptive complexity weight will vary within a range of 0.2 to 0.6 depending on the weather fluctuations. For example, when the weather fluctuation is much higher than the historical average, the complexity weight adjustment coefficient is set to 1.5, then the adaptive complexity weight is 0.4 × 1.5 = 0.6; when the weather fluctuation is much lower than the historical average, the complexity weight adjustment coefficient is set to 0.5, then the adaptive complexity weight is 0.4 × 0.5 = 0.2.

[0051] Secondly, subtract the adaptation complexity weight from 1 to obtain the adaptation relevance weight. For example, when the adaptation complexity weight is 0.6, the adaptation relevance weight is 1-0.6=0.4; when the adaptation complexity weight is 0.2, the adaptation relevance weight is 1-0.2=0.8.

[0052] Finally, based on the adaptation complexity weight and adaptation relevance weight, multiple data processing complexities and multiple types of comprehensive relevance are weighted and evaluated to obtain multiple data sensitivities. The calculation formula can be set as "Data Sensitivity = Adaptation Relevance Weight × Type Comprehensive Relevance + Adaptation Complexity Weight × (1 / Data Processing Complexity)". The higher the complexity, the longer the processing time and the lower the efficiency, which affects the overall analysis efficiency.

[0053] For example, if the overall correlation of a certain monitoring data type is 0.5, the data processing complexity is 5, the adaptation complexity weight is 0.5, and the adaptation correlation weight is 0.5, then the data sensitivity of this monitoring data type is 0.5×0.5+0.5×(1 / 5)=0.25+0.1=0.35.

[0054] Furthermore, the data types to be analyzed are defined based on multiple data sensitivities, including: Based on the multiple data sensitivities, the multiple monitoring data types are arranged in descending order of data sensitivity to obtain a monitoring data type sequence; The number of adaptive monitoring types is obtained by multiplying the complexity weight adjustment coefficient by the initial number of monitoring types and rounding down, wherein the initial number of monitoring types is 5 and the number of adaptive monitoring types is greater than or equal to 2. Select the monitoring data types that match the aforementioned number of monitoring types from the monitoring data type sequence as the data types to be analyzed, and obtain the set of data types to be analyzed.

[0055] In this embodiment, firstly, multiple monitoring data types are arranged based on multiple data sensitivity levels. Data sensitivity reflects the importance of the monitoring data type to the prediction of traffic network resilience. Arranging the data sensitivity from highest to lowest facilitates filtering. For example, if there are monitoring data types such as basic traffic flow data sequences, traffic state data sequences, and vehicle characteristic data sequences, after data sensitivity evaluation, they are arranged in descending order of sensitivity to form a monitoring data type sequence.

[0056] Next, the number of suitable monitoring types is calculated. The complexity weight adjustment factor is multiplied by the initial number of monitoring types and rounded to the nearest integer to obtain the number of suitable monitoring types. If the meteorological fluctuations are small, the complexity weight adjustment factor is less than 1, and the number of suitable monitoring types may decrease, but the number of suitable monitoring types must be greater than or equal to 2. For example, when the complexity weight adjustment factor is 1.2, the number of suitable monitoring types is 1.2 × 5 rounded to 6; when the complexity weight adjustment factor is 0.4, the number of suitable monitoring types is 0.4 × 5 rounded to 2.

[0057] Finally, the monitoring data types that match the top number of monitoring types in the monitoring data type sequence are selected as the data types to be analyzed, thus obtaining the set of data types to be analyzed. It is crucial to ensure that the set of data types to be analyzed includes the data types most important for predicting traffic network resilience, while avoiding the inclusion of too many irrelevant or less influential data types, thereby reducing the workload of data processing and improving prediction efficiency. For example, if the number of matching monitoring types is 4, then the top 4 monitoring data types are selected from the monitoring data type sequence as the data types to be analyzed, forming the set of data types to be analyzed.

[0058] By taking the above steps, we can more accurately assess the data sensitivity of each monitored data type based on meteorological fluctuations, combined with data processing complexity and comprehensive correlation of types. This provides a more scientific and reasonable basis for setting the data type set to be analyzed, and further improves the accuracy and reliability of real-time forecasting of urban integrated transportation network resilience.

[0059] S40: According to the data set to be analyzed, valuable data screening and extraction are performed on the traffic monitoring information in several edge processing units to obtain the monitoring information set to be analyzed; In this embodiment of the application, according to the data set to be analyzed, valuable data is filtered and extracted from the traffic monitoring information in several edge processing units to obtain the monitoring information set to be analyzed.

[0060] Edge processing units are distributed across key nodes in the urban transportation network, such as intersections, bridges, and tunnels. First, based on the data set to be analyzed, the edge processing units perform preliminary filtering of traffic monitoring information. For example, for traffic flow data, obviously erroneous or abnormal data points are removed, such as abnormally high or low vehicle speeds, or sudden drops in traffic volume to zero, which do not reflect reality.

[0061] During the screening process, the edge processing unit classifies the data based on its characteristics and format. For image data, it is categorized according to information such as the time, location, and device used for capture. For sensor data, it is divided according to the data type, such as traffic flow information that can assess road congestion, including the number, speed, and intervals of vehicles, and road surface condition monitoring information, including road surface temperature, humidity, cracks, and settlement.

[0062] Next, valuable data is screened and extracted to obtain the monitoring information set to be analyzed.

[0063] Specifically, step S40 in this application includes: A first data type to be analyzed is randomly selected from the set of data types to be analyzed, and a first edge processing unit is randomly selected from the plurality of edge processing units; In the first edge processing unit, a first monitoring data sequence of the first data type to be analyzed is obtained, and valuable data is filtered and extracted from the first monitoring data sequence to obtain the first monitoring data to be analyzed, which is then added to the monitoring information set to be analyzed. The process involves valuable data screening and extraction from the first monitoring data sequence to obtain the first monitoring data to be analyzed, including: The first monitoring data in the first monitoring data sequence is taken as the first value data, and the adjacent data of the first value data are set as the second monitoring data. The first value data and the second monitoring data are compared for similarity. If the similarity between the two is greater than or equal to a preset similarity threshold, the second monitoring data is removed, and the comparison is continued starting from the first value data. If the similarity between the two is less than the preset similarity threshold, the second monitoring data is set as the second value data, and the comparison continues from the second value data until the data is traversed. All the value data are then combined to obtain the first monitoring data to be analyzed.

[0064] In this embodiment, firstly, a first data type to be analyzed is randomly selected from the set of data types to be analyzed, such as a basic traffic flow data sequence. Simultaneously, a first edge processing unit is randomly selected from several edge processing units; for example, an edge processing unit located at a major urban intersection is selected.

[0065] Secondly, the first monitoring data sequence of the basic traffic flow data sequence is acquired by the first edge processing unit. This sequence contains data such as traffic volume and speed passing through the intersection over a period of time. Next, valuable data is filtered and extracted from the first monitoring data sequence to obtain the first monitoring data to be analyzed, which is then added to the monitoring information set to be analyzed.

[0066] Further, after obtaining the first monitoring data to be analyzed, the first monitoring data in the first monitoring data sequence is taken as the first value data. The adjacent data of the first value data are set as the second monitoring data. A similarity comparison is performed between the first value data and the second monitoring data. If the similarity between the first value data and the second monitoring data is greater than or equal to a preset similarity threshold, the second monitoring data is discarded, and the comparison continues starting from the first value data. If the similarity between the first value data and the second monitoring data is greater than or equal to the preset similarity threshold, it indicates that the two are highly similar, and only the first value data needs to be compared.

[0067] If the similarity between the first value data and the second monitoring data is less than the preset similarity threshold, then the second monitoring data is set as the second value data, and the comparison continues from the second value data until all data is traversed, and all value data are combined to obtain the first monitoring data to be analyzed.

[0068] For example, the first monitoring data in the first monitoring data sequence is taken as the first value data. Assuming the first monitoring data is a traffic flow of 30 vehicles / minute at a certain moment, its adjacent data is designated as the second monitoring data, for example, a traffic flow of 31 vehicles / minute at an adjacent moment. A preset similarity threshold of 90% is set, and the similarity of the data is calculated. If, through a specific similarity calculation method, the similarity between the two data is greater than or equal to 90%, the second monitoring data is discarded. That is, the 31 vehicles / minute data is not included in subsequent analysis; instead, the first value data of 30 vehicles / minute is used as the starting point to continue comparison with the next adjacent data.

[0069] If the calculated similarity between the two is less than 90%, for example, if the traffic flow at adjacent times is 20 vehicles / minute, then the second monitoring data of 20 vehicles / minute is set as the second value data, and comparison with the next adjacent data point continues from 20 vehicles / minute as the starting point. This process is repeated until all data has been traversed.

[0070] Finally, all the selected valuable data are combined to obtain the first monitoring data to be analyzed, and this data is added to the monitoring information set to be analyzed.

[0071] S50: Input the monitoring information set to be analyzed into the traffic network resilience prediction channel, and output the predicted traffic network resilience index of the target city in the preset time zone.

[0072] In this embodiment of the application, firstly, a traffic network resilience prediction channel is constructed based on a graph neural network. The set of monitoring information to be analyzed, obtained by combining all valuable data, is input into the traffic network resilience prediction channel. After passing through a machine learning algorithm, the predicted traffic network resilience index of the target city within the preset time zone is output.

[0073] The traffic network resilience prediction channel leverages the powerful learning capabilities of graph neural networks to capture the complex relationships between nodes and edges in a traffic network. Graph neural networks can model traffic networks, treating intersections, road segments, etc., as nodes, and the connections between nodes as edges. By learning the features of nodes and edges, as well as their interactions, the resilience of the traffic network can be predicted.

[0074] Specifically, step S50 of this application includes: Based on historical traffic monitoring data of the target city, and constrained by the data type set to be analyzed, multiple sample monitoring information sets are collected, and historical resilience datasets under different sample monitoring information are set as sample resilience datasets, thus obtaining multiple sample resilience datasets. Using the multiple sample monitoring information sets as input and the multiple sample resilience datasets as supervision, the spatiotemporal graph neural network is trained until convergence to obtain the traffic network resilience prediction channel. Input the monitoring information set to be analyzed into the traffic network resilience prediction channel, and output the predicted resilience dataset. Based on the weights of the aforementioned indicators, the predicted transportation network resilience index of the target city within the preset time zone is calculated according to the predicted resilience dataset.

[0075] In this embodiment, firstly, based on historical traffic monitoring data of the target city, and constrained by the data type set to be analyzed, multiple sample monitoring information sets are collected, and historical resilience datasets under different sample monitoring information are obtained as sample resilience datasets, resulting in multiple sample resilience datasets. These sample resilience datasets include traffic facility and equipment stability, traffic network connectivity, travel time loss, effective redundant path ratio, speed maintenance rate, average accident recovery time, and planned time index.

[0076] Secondly, using multiple sample monitoring information sets as input and multiple sample resilience datasets as supervision, a model is trained based on a graph neural network until convergence, resulting in a traffic network resilience prediction channel. The monitoring information set to be analyzed is input into the traffic network resilience prediction channel, and the predicted resilience dataset is output.

[0077] For example, a traffic network resilience prediction channel is constructed and trained based on a graph convolutional neural network.

[0078] First, data preparation: Based on historical traffic monitoring data of the target city, and constrained by the data set to be analyzed, multiple sample monitoring information sets are collected. The input nodes of the traffic network resilience prediction channel are multiple sample monitoring information sets, including historical traffic facility and equipment stability, traffic network connectivity, total travel time loss, effective redundant path ratio, driving speed maintenance rate, average accident recovery time and planned time index.

[0079] Secondly, model construction involves building a graph convolutional neural network model, including convolutional layers, pooling layers, and fully connected layers. Convolutional layers extract image features, pooling layers reduce the size of the feature maps, and fully connected layers convert the feature maps into labels for the predicted resilience dataset. The input layer has the number of nodes equal to the dimension of the input features. For example, if there are 10 features (historical total travel time loss, effective redundant path ratio, speed maintenance rate, average accident recovery time, and planned time exponent), the input layer contains 10 nodes. One to three hidden layers are set, with the number of nodes in each layer adjusted experimentally (e.g., 64, 32, etc.). The activation function used is ReLU. The number of nodes in the output layer equals the number of predicted current total travel time loss, effective redundant path ratio, speed maintenance rate, average accident recovery time, and planned time exponent. For example, predicting only travel time requires one node, while predicting both travel time and energy consumption requires two nodes. The output layer generally does not use an activation function and directly outputs continuous values.

[0080] Next, the model is trained, with the predicted resilience dataset as the output, including information such as the predicted total travel time loss, effective redundant path ratio, speed maintenance rate, average recovery time after an accident, and planned time exponent under the current scenario. Multiple sample resilience datasets are used for supervision. The Adam optimizer and mean squared error loss function are used to construct the training framework, with a batch size of 32 and a total of 50 training epochs. An early stopping mechanism (patience=5) is introduced: when the validation set loss does not decrease for 5 consecutive epochs, the training process is automatically terminated, resulting in a trained model identifying the modeling element type. This effectively avoids overfitting while ensuring the model reaches convergence.

[0081] Finally, the predicted resilience dataset is weighted based on several indicator weights. Different indicators have varying degrees of impact on traffic network resilience; assigning corresponding weights reflects the actual resilience of the traffic network. For example, the weight of the total travel time loss indicator is 0.3, the effective redundant path ratio is 0.2, the speed maintenance rate is 0.25, the average accident recovery time is 0.15, and the planned time index is 0.1. The predicted values ​​for traffic flow recovery capacity, traffic congestion mitigation capacity, and traffic accident response capacity in the predicted resilience dataset are 80, 70, 60, 50, and 45, respectively. Therefore, the predicted traffic network resilience index is 0.3×80 + 0.2×70 + 0.25×60 + 0.15×50 + 0.1×45 = 65.

[0082] In summary, compared to existing technologies, this application, in terms of data screening and extraction, compares the similarity of the first monitoring data sequence and eliminates data with high similarity, ensuring that the screened data is representative and effective. Regarding traffic network resilience prediction, a traffic network resilience prediction channel is constructed based on a graph neural network, fully utilizing the powerful learning capabilities of graph neural networks to more accurately predict the resilience of traffic networks.

[0083] In summary, the embodiments of this application have at least the following technical effects: This application provides a real-time prediction method for the resilience of urban integrated transportation networks by combining IoT sensing. First, IoT sensing devices collect information at key traffic nodes. Then, resilience assessment index weights are obtained by combining scene attribute information. Data sensitivity evaluation and data filtering are performed considering meteorological fluctuation characteristics. Finally, prediction results are output through a trained traffic network resilience prediction channel. This effectively improves the computational efficiency of traffic network resilience prediction, enabling accurate prediction of traffic resilience within a limited time, better meeting the real-time requirements of urban traffic, and providing strong support for the management and optimization of urban integrated transportation networks. Through the above technical solution, IoT sensing devices are deployed at key urban traffic nodes to collect multi-dimensional data such as traffic flow, road conditions, and facility operation status in real time. Edge computing technology is used to preprocess the collected real-time data to filter out core indicators related to network resilience, such as road segment capacity, transfer efficiency, and emergency response speed. Then, a deep learning prediction model integrating spatiotemporal features is constructed and trained to predict traffic network resilience in real time under disturbances such as sudden weather and accidents, providing accurate early warning and decision support for traffic management departments, effectively improving the anti-interference and recovery efficiency of urban transportation systems.

[0084] Among them, the set of data types to be analyzed is set according to multiple data sensitivities, such as Figure 2 As shown, it includes: Based on the multiple data sensitivities, the multiple monitoring data types are arranged in descending order of data sensitivity to obtain a monitoring data type sequence; The number of adaptive monitoring types is obtained by multiplying the complexity weight adjustment coefficient by the initial number of monitoring types and rounding down, wherein the initial number of monitoring types is 5 and the number of adaptive monitoring types is greater than or equal to 2. Select the monitoring data types that match the aforementioned number of monitoring types from the monitoring data type sequence as the data types to be analyzed, and obtain the set of data types to be analyzed.

[0085] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. Additionally, the processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0086] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0087] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A real-time prediction method for the resilience of urban integrated transportation networks combined with Internet of Things (IoT) sensing, characterized in that the method... include: IoT sensing devices are deployed at several key traffic nodes in the target city to collect traffic information and obtain traffic monitoring information. Each key traffic node includes an edge processing unit. Based on the scenario attribute information of the target city within the preset time zone, several indicator weights of several resilience assessment indicators are obtained. Based on the meteorological fluctuation characteristics of the target city within the preset time zone and several indicator weights, data sensitivity evaluation is performed on multiple monitoring data types, and a set of data types to be analyzed is set according to multiple data sensitivities. According to the data set to be analyzed, valuable data is filtered and extracted from the traffic monitoring information in several edge processing units to obtain the monitoring information set to be analyzed. The set of monitoring information to be analyzed is input into the traffic network resilience prediction channel, and the predicted traffic network resilience index of the target city in the preset time zone is output.

2. The method for real-time prediction of urban integrated transportation network resilience combined with IoT sensing as described in claim 1, characterized in that, Deploy IoT sensing devices at several key traffic nodes in the target city to collect traffic information and obtain traffic monitoring data, including: Deploy IoT sensing devices at several key traffic nodes in the target city to build a traffic monitoring network. The IoT sensing devices include at least high-definition cameras, LiDAR, RFID readers, weather sensors, and road condition sensors. The traffic monitoring network continuously monitors the traffic status of several key traffic nodes and obtains traffic monitoring information, which includes basic traffic flow data sequences, traffic status data sequences, environmental status data sequences, and vehicle characteristic data sequences within a historical time range.

3. The method for real-time prediction of urban integrated transportation network resilience combined with IoT sensing as described in claim 1, characterized in that, Based on the scenario attribute information of the target city within the preset time zone, several resilience assessment indicators and their weights are obtained, including: Obtain scene attribute information of a target city within a preset time zone, wherein the scene attribute information includes time, day type, season, and climate conditions; Several resilience assessment indicators were obtained, including the stability of traffic facilities and equipment, traffic network connectivity, total travel time loss, effective redundant path ratio, speed maintenance rate, average recovery time after an accident and planned time index. Based on historical traffic monitoring data of the target city, the relative importance of the several resilience assessment indicators is evaluated according to the scenario attribute information, and the weights of several indicators are output.

4. The real-time prediction method for urban integrated transportation network resilience combined with IoT sensing according to claim 3, characterized in that, Based on the meteorological fluctuation characteristics of the target city within the preset time zone and several indicator weights, data sensitivity evaluations are performed on multiple monitoring data types, including: Obtain predicted meteorological information for a target city within a preset time zone, wherein the predicted meteorological information includes at least precipitation information, visibility information, wind information, and temperature information; The volatility of precipitation, visibility, wind, and temperature information was assessed and weighted to obtain the meteorological volatility. Based on the predicted meteorological information, the data acquisition complexity, data analysis complexity, and data reliability of the multiple monitoring data types are evaluated respectively, and multiple data processing complexities are obtained by weighted calculation. The data processing complexity is positively correlated with the data acquisition complexity and data analysis complexity, and negatively correlated with the data reliability. Based on historical traffic monitoring data of the target city, the correlation sets of multiple types between the several resilience assessment indicators and the multiple monitoring data types are analyzed and obtained. The correlation sets of multiple types are then weighted and fused according to the weights of the several indicators to output the comprehensive correlation of multiple types. Based on the meteorological fluctuation, multiple data sensitivities of the multiple monitoring data types are obtained by weighted evaluation according to the multiple data processing complexities and multiple types of comprehensive correlation. Among them, data sensitivity is negatively correlated with data processing complexity and positively correlated with type comprehensive correlation.

5. The real-time prediction method for urban integrated transportation network resilience combined with IoT sensing according to claim 4, characterized in that, Based on the meteorological fluctuations, multiple data sensitivities of the multiple monitored data types are obtained through a weighted evaluation based on the complexity of multiple data processing methods and the comprehensive correlation of multiple types, including: The ratio of the meteorological fluctuation to the average historical meteorological fluctuation of the target city within the historical time period is set as the complexity weight adjustment coefficient, which is multiplied by the initial complexity weight to obtain the adaptive complexity weight. The initial complexity weight is 0.4, and the adaptive complexity weight is greater than or equal to 0.2 and less than or equal to 0.

6. The adaptation complexity weight is obtained by subtracting the adaptation relevance weight from 1; Based on the adaptation complexity weight and adaptation relevance weight, multiple data sensitivities are obtained by weighted evaluation according to the multiple data processing complexities and the comprehensive relevance of multiple types.

6. The method for real-time prediction of urban integrated transportation network resilience combined with IoT sensing as described in claim 5, characterized in that, The data types to be analyzed are defined based on multiple data sensitivities, including: Based on the multiple data sensitivities, the multiple monitoring data types are arranged in descending order of data sensitivity to obtain a monitoring data type sequence; The number of adaptive monitoring types is obtained by multiplying the complexity weight adjustment coefficient by the initial number of monitoring types and rounding down, wherein the initial number of monitoring types is 5 and the number of adaptive monitoring types is greater than or equal to 2. Select the monitoring data types that match the aforementioned number of monitoring types from the monitoring data type sequence as the data types to be analyzed, and obtain the set of data types to be analyzed.

7. The method for real-time prediction of urban integrated transportation network resilience combined with IoT sensing as described in claim 1, characterized in that, According to the data set to be analyzed, valuable data is filtered and extracted from the traffic monitoring information in several edge processing units to obtain the monitoring information set to be analyzed, including: A first data type to be analyzed is randomly selected from the set of data types to be analyzed, and a first edge processing unit is randomly selected from the plurality of edge processing units; In the first edge processing unit, a first monitoring data sequence of the first data type to be analyzed is obtained, and valuable data is filtered and extracted from the first monitoring data sequence to obtain the first monitoring data to be analyzed, which is then added to the monitoring information set to be analyzed. The process involves valuable data screening and extraction from the first monitoring data sequence to obtain the first monitoring data to be analyzed, including: The first monitoring data in the first monitoring data sequence is taken as the first value data, and the adjacent data of the first value data are set as the second monitoring data. The first value data and the second monitoring data are compared for similarity. If the similarity between the two is greater than or equal to a preset similarity threshold, the second monitoring data is removed, and the comparison is continued starting from the first value data. If the similarity between the two is less than the preset similarity threshold, the second monitoring data is set as the second value data, and the comparison continues from the second value data until the data is traversed. All the value data are then combined to obtain the first monitoring data to be analyzed.

8. The method for real-time prediction of urban integrated transportation network resilience combined with IoT sensing according to claim 2, characterized in that, The monitoring information set to be analyzed is input into the traffic network resilience prediction channel, and the predicted traffic network resilience index of the target city within the preset time zone is output, including: Based on historical traffic monitoring data of the target city, and constrained by the data type set to be analyzed, multiple sample monitoring information sets are collected, and historical resilience datasets under different sample monitoring information are set as sample resilience datasets, thus obtaining multiple sample resilience datasets. Using the multiple sample monitoring information sets as input and the multiple sample resilience datasets as supervision, the spatiotemporal graph neural network is trained until convergence to obtain the traffic network resilience prediction channel. Input the monitoring information set to be analyzed into the traffic network resilience prediction channel, and output the predicted resilience dataset. Based on the weights of the aforementioned indicators, the predicted transportation network resilience index of the target city within the preset time zone is calculated according to the predicted resilience dataset.

Images