Smart city traffic resource optimal configuration method and platform and storage medium

Abstract

The invention discloses a smart city traffic resource optimal configuration method, a platform and a storage medium, and relates to the technical field of traffic optimization. When a road accident is detected, accident information is collected, and a traffic accident position and traffic accident characteristics are obtained; time consumption prediction is carried out according to the traffic accident characteristics, and traffic control duration is obtained; calling real-time traffic data from an intelligent traffic system to obtain associated road data; executing multi-thread traffic flow back-stepping analysis to obtain associated traffic flow prediction information; carrying out shunting performance evaluation; and dynamically updating the signal lamp configuration of the plurality of shunting candidate roads by taking the associated traffic flow prediction information as a constraint. According to the invention, the technical problems of aggravation of traffic jam and low emergency management efficiency caused by lack of scientificity and dynamic adaptability of shunting planning and lagging of signal lamp regulation in a traffic accident scene in the prior art are solved, and the technical effects of optimizing traffic resource configuration, relieving traffic jam and improving traffic emergency management efficiency are achieved.

Description

Technical Field

[0001] The present invention relates to the field of traffic optimization technology, and in particular to a method, platform and storage medium for optimizing the configuration of traffic resources in a smart city. Background Art

[0002] With the acceleration of urbanization, the construction of smart cities has become an important direction for the development of modern cities. As the key link of urban operation, the optimal allocation of resources of the transportation system is particularly important. In the field of urban transportation, traditional traffic management methods have exposed many limitations when facing road traffic accidents. On the one hand, traffic flow prediction technology is relatively single and data integration is insufficient. Some existing methods often rely only on limited data of local sections for simple traffic flow estimation, and cannot fully consider the complex topological structure of the entire associated road network and the comprehensive impact of multi-source traffic data. For example, after an accident, it is difficult to accurately predict the traffic flow change trend and redistribution of vehicle flow directions in different sections of the surrounding area, resulting in the lack of pertinence and effectiveness of subsequent traffic diversion plans. On the other hand, the formulation of diversion strategies lacks systematicity and dynamic adaptability. In most cases, the selection of diversion roads is based only on empirical judgment or simple road capacity indicators, and the comprehensive impact of multiple factors such as the real-time road conditions, traffic signal layout, and road geometric characteristics of the diversion roads on the diversion effect is not deeply analyzed. Moreover, once the diversion plan is determined, it cannot be adjusted in time according to real-time changes in traffic conditions, which can easily lead to an unbalanced situation in which congestion on some diversion roads is exacerbated while other road resources are not fully utilized.

[0003] The existing technology has technical problems such as the lack of scientificity and dynamic adaptability of diversion planning in traffic accident scenarios, and the lag in traffic light control, which in turn leads to increased traffic congestion and inefficient emergency management. Summary of the invention

[0004] This application provides a method, platform and storage medium for optimizing the allocation of traffic resources in a smart city, which is used to solve the technical problems in the prior art that diversion planning in traffic accident scenarios lacks scientificity and dynamic adaptability, and traffic light control lags, which leads to aggravated traffic congestion and inefficient emergency management.

[0005] In view of the above problems, the present application provides a method, platform and storage medium for optimizing the configuration of traffic resources in a smart city.

[0006] The first aspect of the present application provides a method for optimizing the configuration of traffic resources in a smart city, the method comprising:

[0007] The road condition is monitored by an intelligent transportation system. When a road accident is detected, an accident detection task is initiated to collect accident information and obtain the location and characteristics of the traffic accident. A time consumption prediction is performed based on the characteristics of the traffic accident to obtain the traffic control duration. After the interactive GIS system obtains the associated road network of the traffic accident location, the associated road network is used as a constraint to call real-time traffic data from the intelligent transportation system to obtain associated road data. With the traffic accident location and traffic control duration as constraints, the associated road data is used to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information. The associated road network is evaluated for diversion performance based on the associated traffic flow prediction information, and multiple diversion candidate roads are located based on the evaluation results. 1 / K of the traffic control duration is used as a diversion update window, and the signal light configurations of the multiple diversion candidate roads are dynamically updated with the associated traffic flow prediction information as a constraint.

[0008] The second aspect of the present application provides a smart city traffic resource optimization configuration platform, the platform comprising:

[0009] A traffic accident feature acquisition module, which is used to monitor road conditions through an intelligent transportation system. When a road accident is detected, the accident detection task is started to collect accident information to obtain the location and characteristics of the traffic accident; a traffic control duration acquisition module, which is used to predict the time consumption according to the traffic accident characteristics and obtain the traffic control duration; an associated road data acquisition module, which is used to obtain the associated road network of the traffic accident location in the interactive GIS system, and then use the associated road network as a constraint to call real-time traffic data from the intelligent transportation system to obtain associated road data; A module for acquiring associated traffic flow prediction information, wherein the module is used to use the associated road data to perform multi-threaded traffic flow reverse analysis based on the traffic accident location and traffic control duration as constraints to obtain associated traffic flow prediction information; a diversion performance evaluation module, wherein the diversion performance evaluation module is used to evaluate the diversion performance of the associated road network according to the associated traffic flow prediction information, and locate multiple diversion candidate roads based on the evaluation results; and a dynamic update module, wherein the dynamic update module is used to use 1 / K of the traffic control duration as a diversion update window, and dynamically update the signal light configurations of the multiple diversion candidate roads based on the associated traffic flow prediction information as a constraint.

[0010] The third aspect of the present application provides a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the smart city traffic resource optimization configuration method provided in the present application.

[0011] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0012] Through the intelligent transportation system, the road condition is monitored. When a road accident is detected, the accident detection task is started to collect accident information, obtain the location and characteristics of the traffic accident; the time consumption is predicted according to the characteristics of the traffic accident to obtain the traffic control duration; after obtaining the associated road network of the traffic accident location, real-time traffic data is called to obtain the associated road data; with the traffic accident location and traffic control duration as constraints, multi-threaded traffic flow reverse analysis is performed to obtain associated traffic flow prediction information; the diversion performance of the associated road network is evaluated according to the associated traffic flow prediction information, and multiple diversion candidate roads are located based on the evaluation results; 1 / K of the traffic control duration is used as the diversion update window to dynamically update the signal light configuration of the multiple diversion candidate roads. The technical effect of optimizing traffic resource allocation, alleviating traffic congestion and improving the efficiency of traffic emergency management is achieved. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0014] Figure 1 A schematic diagram of a flow chart of a method for optimizing the configuration of smart city traffic resources provided in an embodiment of the present application;

[0015] Figure 2 A schematic diagram of the structure of the smart city traffic resource optimization configuration platform provided in the embodiment of the present application.

[0016] Explanation of the reference numerals: traffic accident feature acquisition module 10 , traffic control duration acquisition module 20 , associated road data acquisition module 30 , associated traffic flow prediction information acquisition module 40 , diversion performance evaluation module 50 , dynamic update module 60 . DETAILED DESCRIPTION

[0017] This application provides a method, platform and storage medium for optimizing the allocation of traffic resources in a smart city, which is used to solve the technical problems in the prior art that diversion planning in traffic accident scenarios lacks scientificity and dynamic adaptability, and traffic light control lags, which leads to increased traffic congestion and inefficient emergency management.

[0018] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0019] Embodiment 1, as Figure 1 As shown, the present application provides a method for optimizing the configuration of traffic resources in a smart city, the method comprising:

[0020] Step S100: Monitor the road condition through the intelligent transportation system. When a road accident is detected, start the accident detection task to collect accident information and obtain the location and characteristics of the traffic accident.

[0021] Specifically, the intelligent transportation system relies on various monitoring devices widely distributed in the urban road network, such as high-definition cameras, geomagnetic sensors, microwave radars, etc., to carry out all-round and real-time monitoring of traffic conditions. These devices continuously collect massive amounts of traffic data, including vehicle speed, traffic volume, road occupancy rate and other information. When the system accurately detects a road accident based on the in-depth analysis of these data and intelligent algorithm judgment, it will immediately trigger the start of the accident detection task. The accident detection task integrates multi-source data, such as using the image recognition technology of high-definition cameras to identify the shape and location information of the accident vehicle, combining the data of geomagnetic sensors and microwave radars to determine the abnormal fluctuation characteristics of traffic flow at the moment of the accident, and then comprehensively analyzes the accident scene, so as to accurately determine the location of the traffic accident, and sort out in detail the characteristics of the traffic accident, such as whether the accident type is a collision or a breakdown, the model and number of vehicles involved, and the severity of the accident, whether it is a minor scratch or serious damage, etc., to provide timely and accurate basic data support for subsequent traffic management decisions.

[0022] Step S200: Make a time consumption prediction based on the traffic accident characteristics to obtain the traffic control duration.

[0023] Specifically, in order to accurately predict the duration of traffic control, the acquired traffic accident characteristics will first be deeply analyzed and quantified. For example, if the traffic accident characteristics show that there are multiple serious collisions involving large trucks and buses, and there are vehicle fires and serious road deformations at the scene of the accident, the difficulty of clearing the accident and the scale of rescue resources required will be evaluated based on this. On the one hand, considering that the towing of large vehicles requires professional equipment and takes a long time, the fire extinguishing and subsequent treatment of the burning vehicle also takes a certain amount of time; on the other hand, for the road damage caused by the accident, it is necessary to estimate the manpower, material resources and time costs required for repair. At the same time, a comprehensive analysis is conducted based on factors such as the traffic flow of the section where the accident occurred, the road grade, and the traffic capacity of the surrounding detour routes. If the accident occurs on a city trunk road during peak traffic hours with a large traffic volume, the traffic control time will inevitably increase accordingly, because more time is needed to divert the backlog of vehicles and arrange reasonable detour plans. By inputting these complex factors into a specially constructed traffic control duration prediction model, which is based on a large amount of historical accident data and traffic flow simulation algorithms, a relatively accurate traffic control duration estimate is ultimately obtained. This result will provide a critical time basis for the subsequent rational allocation and optimization of traffic resources, so that traffic management departments can plan and implement effective traffic diversion strategies in advance, minimizing the overall impact of accidents on the urban traffic network.

[0024] Step S300: After the interactive GIS system obtains the associated road network of the traffic accident location, real-time traffic data is called from the intelligent transportation system with the associated road network as a constraint to obtain associated road data.

[0025] Specifically, when the interactive GIS system determines the associated road network based on the location of the traffic accident, it starts the precise data call process. With the help of powerful geographic information processing capabilities, it can accurately outline the associated road network architecture radiating outward from the accident point, covering different levels and types of roads such as main roads, branch roads, and loop roads. Subsequently, with this associated road network as a clear constraint, a data call instruction is issued to the intelligent transportation system. The intelligent transportation system responds quickly, relying on its huge traffic data storage and processing system to filter out data information closely related to the associated road network from a large amount of real-time traffic data. These data cover multi-dimensional information such as real-time traffic flow, speed distribution, vehicle density, and traffic conditions at intersections on each associated road, comprehensively and meticulously reflecting the real-time operating status of the associated roads, and finally integrating to form associated road data, providing a rich and accurate data foundation for the subsequent in-depth analysis of traffic flow trends and the formulation of scientific and reasonable traffic diversion strategies.

[0026] Step S400: Taking the traffic accident location and traffic control duration as constraints, the associated road data is used to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information.

[0027] Specifically, the location of the traffic accident is used as the spatial constraint reference point, and the traffic control duration is used as the time series constraint condition. The multi-threaded traffic flow reverse analysis is carried out using the associated road data. First, the location of the traffic accident is accurately located using the geographic information system (GIS) technology, and the impact area of ​​the traffic flow disturbance on the surrounding road network is determined based on the principles of traffic engineering. Then, with the traffic control duration as the time axis, combined with the flow conservation law, the following model and the queuing theory in the traffic flow theory, a multi-threaded analysis framework is constructed. In each thread, the parameters such as the traffic flow, speed, and vehicle density in the associated road data are input into the pre-built microscopic traffic simulation model. By simulating the individual behaviors of the vehicles in the model, such as the acceleration, deceleration, lane change, parking and waiting of the vehicles, the propagation process of the traffic flow on each associated road is reversed with the time of the accident as the starting point. In a multi-threaded parallel computing environment, threads interact and synchronize data through shared memory or message passing mechanisms, taking into account the coupling of multiple factors such as road topology, traffic signal control strategy, and driver behavior characteristics, and ultimately obtaining related traffic flow prediction information, including traffic flow prediction values ​​for each section at different time intervals, traffic speed distribution, formation and dissipation time series of congested sections, and traffic diversion and merging node locations, etc., to provide accurate data support for the formulation of subsequent traffic diversion plans.

[0028] Step S500: Evaluate the diversion performance of the associated road network according to the associated traffic flow prediction information, and locate multiple diversion candidate roads based on the evaluation results.

[0029] Specifically, firstly, the key elements such as the peak traffic volume, traffic flow change trend, congestion duration and degree of each road section in the associated traffic flow prediction information are carefully analyzed. Combined with the topological structure of the road network, the capacity analysis model in traffic engineering is used to accurately calculate the actual capacity and saturation of each road at different times. Then, the inherent attributes such as the grade of the road, the number of lanes, the type and spacing of intersections, and the influence of the surrounding land use properties on the generation and attraction of traffic flow are comprehensively considered to construct a multi-dimensional diversion performance evaluation index system. By comparing and quantitatively analyzing the actual traffic flow data of each road with the evaluation index system one by one, mathematical methods such as hierarchical analysis method and fuzzy comprehensive evaluation are used to determine the diversion performance score of each road. Finally, according to the ranking of the scores, multiple roads with excellent diversion performance and able to effectively alleviate the traffic pressure in the accident area are selected as diversion candidate roads. These roads will become the core resources for the subsequent formulation of traffic diversion strategies and optimization of signal light configuration, so as to maximize the overall operation efficiency of the urban traffic network during the accident.

[0030] Step S600: Using 1 / K of the traffic control duration as a diversion update window, and dynamically updating the signal light configurations of the plurality of diversion candidate roads based on the associated traffic flow prediction information as a constraint.

[0031] Specifically, 1 / K of the traffic control duration is clearly set as the diversion update window, and this fixed-duration window becomes the time rhythm controller for the dynamic adjustment of the signal light configuration. With accurate associated traffic flow prediction information as the core basis and constraint conditions, the dynamic update operation of the signal light configuration is carried out on multiple diversion candidate roads. When each diversion update window is opened, the key data such as the traffic flow change rate of each diversion candidate road, the traffic flow saturation of different sections, and the flow direction dynamics of the traffic flow in the associated traffic flow prediction information are deeply analyzed. Subsequently, based on the traffic signal control theory and algorithm and the dynamic timing model of real-time traffic flow data, the duration of each phase of the signal lights of each diversion candidate road is recalculated and allocated. For example, when the traffic flow of a diversion candidate road increases rapidly and tends to be congested during the window period, the vehicle traffic efficiency of the road is improved by shortening the red light duration, extending the green light duration, and optimizing the phase conversion sequence; on the contrary, if the traffic flow is stable or reduced, the signal light timing is fine-tuned accordingly to balance the resource utilization of each road. After the diversion update window ends, this dynamic update process is repeated based on the new associated traffic flow prediction information, so that the signal light configuration can continuously adapt to the dynamic changes in traffic flow, ensure the efficient operation of the diversion candidate road network during traffic control, effectively alleviate the traffic congestion pressure caused by accidents and maintain traffic order stability.

[0032] In a possible implementation, step S100 further includes:

[0033] Step S110: In the intelligent transportation system, real-time road condition monitoring is performed on multiple monitored road blocks through multiple parallel road condition data collection tasks of the road condition data collection container to obtain road condition data of multiple blocks.

[0034] Step S120: the road condition data collection container publishes the plurality of block road condition data across containers to a road condition anomaly analysis container, wherein the road condition anomaly analysis container is deployed in the intelligent transportation system.

[0035] Step S130: the plurality of parallel road condition anomaly analysis tasks in the road condition anomaly analysis container perform road anomaly analysis based on the plurality of block road condition data, and output a road accident warning, wherein the road accident warning has a road block identifier.

[0036] Step S140: the intelligent transportation system activates the accident detection container according to the road accident warning, and assigns the highest priority to the accident detection container.

[0037] Step S150: In the accident detection container, with the road block identifier as a constraint, the accident detection task is run to collect accident information to obtain the traffic accident location and traffic accident characteristics.

[0038] Specifically, multiple parallel road data collection tasks in the road data collection container are based on distributed architecture and multi-threaded processing technology to carry out synchronous monitoring operations on multiple preset monitoring road blocks. For the collection of vehicle speed, a high-precision geomagnetic sensor and a laser speed meter are used. The geomagnetic sensor detects the magnetic field changes caused by the passing of the vehicle, combines its internal signal processing algorithm to calculate the time interval of the vehicle passing, and then uses the speed calculation formula to obtain the vehicle speed data based on the known sensor installation spacing; the laser speed meter uses its high-frequency laser pulse emission and reception mechanism to accurately measure the displacement changes of the vehicle in a very short time, thereby obtaining a more accurate instantaneous speed value, and fusion calibration with the geomagnetic sensor data. For the monitoring of traffic density, on the one hand, the geomagnetic sensor counts the number of vehicles passing, and on the other hand, the video image processing technology is used to shoot the road picture through the high-definition camera installed at the monitoring point, and the image recognition algorithm is used to detect and track the vehicles in the picture, and the number of vehicles in a specific area per unit time is counted, and the traffic density is calculated in combination with the road area of ​​the area. These collection tasks run independently in their own threads without interfering with each other. The collected vehicle speed data and traffic density data are packaged and integrated according to specific data formats and coding specifications, and finally multiple detailed and accurate block road condition data are formed.

[0039] When the road condition data collection container successfully obtains road condition data from multiple blocks, the cross-container publishing process is immediately started. Under the complex data interaction framework built within the intelligent transportation system, the road condition data collection container organizes and encapsulates road condition data from multiple blocks according to the pre-set communication protocol and data transmission mechanism to ensure the integrity and accuracy of the data. Subsequently, with the help of the inter-container communication interface at the bottom of the system, the encapsulated data is sent to the container dedicated to road condition anomaly analysis. During this process, the data will be transmitted through middleware such as high-speed data buses or distributed message queues to ensure the efficiency and reliability of data transmission. Once the data arrives at the road condition anomaly analysis container, the container will parse and store the received data based on its own configuration and functional modules, and make sufficient data preparation for the subsequent road condition anomaly analysis tasks based on these data, so that the entire intelligent transportation system can respond to changes in road conditions quickly and accurately and handle them effectively.

[0040] Multiple parallel road condition anomaly analysis tasks inside the road condition anomaly analysis container are activated simultaneously, and they carry out in-depth analysis with the received road condition data of multiple blocks as the core input source. These tasks use a variety of road condition anomaly detection models pre-built and trained by the intelligent transportation system, such as traffic flow anomaly recognition models based on machine learning algorithms, image anomaly detection models based on deep learning (if assisted by camera data), etc. By extracting features and pattern recognition of key parameters such as vehicle speed sequence and flow density change curve in the block road condition data, and comparing and analyzing with standard model parameters under normal traffic conditions, once a significant deviation from the normal traffic pattern is detected, such as a sharp drop in vehicle speed and an instantaneous increase in flow density in a certain block, combined with cross-validation and comprehensive judgment of multi-source data, it is determined that a road accident may occur in the block. At this time, road accident warning information will be generated and assigned a corresponding road block identifier, which can accurately locate the specific area where the accident occurred, so that the intelligent transportation system can quickly start targeted accident handling processes in the future, such as activating the accident detection container, allocating surrounding traffic resources, etc., to minimize the impact of the accident on the entire traffic network.

[0041] When the intelligent transportation system receives a road accident warning issued by the road condition anomaly analysis container, it immediately starts the emergency response mechanism. A key step is to activate the accident detection container that is specifically used for the collection and preliminary processing of accident details. In the multi-container architecture operating environment of the entire intelligent transportation system, in order to ensure the efficiency and timeliness of accident handling, the priority of the accident detection container is set to the highest level according to the pre-set priority allocation strategy. The implementation of this strategy is completed through the system's resource scheduling and task management module, which will comprehensively evaluate and dynamically allocate and adjust the priorities of each container running in the system, including the traffic flow reverse analysis container, the diversion path planning container, the signal light timing control container, etc., according to the current task requirements and the urgency of the task. For example, during the normal traffic flow monitoring and analysis phase, the traffic flow reverse analysis container may be at a higher priority to continuously optimize the traffic flow model; when a road accident warning is generated, the demand for the accident detection container becomes the most urgent, and its priority is raised to the highest. Some non-critical container tasks will be suspended or their resource allocation ratio will be reduced, and more computing resources, data access rights, network bandwidth and other resources will be allocated to the accident detection container first to ensure that it can quickly and smoothly run and start the accident detection task, and obtain the precise location of the traffic accident and detailed traffic accident feature information in time, laying a solid foundation for a series of subsequent emergency response tasks, such as diversion path planning and traffic light timing adjustment, thereby effectively reducing the negative impact of accidents on the traffic network and ensuring the overall stability and safety of the traffic system.

[0042] After the accident detection container is activated, its core task flow is immediately started. The received road block identification is used as the basis for accurate positioning guidance and data screening, and the accident detection task is fully started. First, with the help of the high-definition camera network widely distributed in the intelligent transportation system, multi-view, high-resolution image acquisition and analysis are carried out on the road block, and advanced image recognition technology is used to identify key visual information such as the vehicle outline, position relationship and collision form at the accident scene, so as to preliminarily determine the approximate range of the traffic accident location. At the same time, combined with various sensors installed along the road, such as geomagnetic sensors, pressure sensors, etc., through in-depth mining and analysis of sensor data, the determination of the accident location is further refined. For example, based on the sensor points corresponding to the vehicle's stop position and the abnormal distribution of the vehicle's pressure on the road surface, the specific coordinate point where the accident occurred is accurately locked. In terms of traffic accident feature collection, the type, quantity, color, and degree of damage of the accident vehicles are analyzed in depth based on visual image data. At the same time, the dynamic features such as the vehicle speed change curve and acceleration abnormality at the moment of the accident are analyzed using sensor data. These multi-dimensional data information are combined to comprehensively sort out the detailed characteristics of the traffic accident, including the severity of the accident, possible causes of the accident, and the initial impact range on the surrounding traffic flow. Through this systematic and accurate accident information collection process, detailed and reliable basic data support is provided for the formulation of subsequent traffic control decisions, the planning of diversion plans, and the deployment of rescue resources, ensuring that the entire intelligent transportation system can carry out follow-up work in an orderly manner when dealing with road accidents, minimize the adverse effects of accidents, and quickly restore traffic order.

[0043] In a possible implementation, step S400 further includes:

[0044] Step S410: constructing a traffic simulation model based on the associated road network, and using the associated road data to perform local update of the traffic simulation model.

[0045] Step S420: using the traffic accident location and traffic control duration to predict the accident impact range in the traffic simulation model to obtain an associated diffusion area.

[0046] Step S430: using the associated diffusion area to select a plurality of associated road sections in the associated road network.

[0047] Step S440: taking the associated diffusion area and the multiple associated road sections as constraints, extracting multiple traffic flow data slices from the associated road data.

[0048] Step S450: using the plurality of traffic flow data slices to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information.

[0049] Specifically, when constructing the traffic simulation model, the graph theory method is used to abstractly express the associated road network. Each intersection in the road network is regarded as a node, and the road section is regarded as an edge. The attribute information such as length, number of lanes, and road grade is assigned to the edge as a weight to build a basic network framework. Then, according to the principle of traffic flow dynamics, the vehicle's driving rules are set, and an improved following model is adopted to comprehensively consider the relative speed and spacing of the front and rear vehicles and the driver's reaction characteristics. The mathematical description of the vehicle following behavior is constructed to accurately simulate the sequential movement process of the vehicle on the road section. For lane changing behavior, a lane changing model based on a combination of rules and probabilities is used. In addition to considering conventional factors such as the traffic volume, speed difference, and safety spacing of adjacent lanes, special factors such as road type and traffic control information are also included. By setting the probability of lane changing under different conditions, the vehicle lane changing decision is more in line with the actual traffic scene. In terms of traffic signal control, for the signal light intersections in the associated road network, the phase setting, cycle duration, green signal ratio and other parameters of the signal light are analyzed in detail, and a signal light control logic model is constructed to clarify the vehicle's access rights and waiting rules under each phase. Finally, the acquired associated road data, including real-time traffic flow, vehicle speed, vehicle density and other information of each road section, are input into the constructed traffic simulation model to initialize and dynamically update the state variables such as the initial vehicle distribution and speed distribution in the model. This enables the traffic simulation model to accurately reflect the current actual traffic operation conditions of the associated road network, providing a reliable and accurate simulation platform for subsequent analysis work such as accident impact range prediction.

[0050] The traffic accident location and traffic control duration are introduced as key input parameters into the constructed and updated traffic simulation model. First, the node or road section corresponding to the traffic accident location is accurately located in the model, and the initial disturbance of the traffic flow is simulated with this point as the core. The simulation time range of the model is set according to the traffic control duration. Within this time range, the reduction or even blockage of road capacity caused by the accident is considered, such as simulating the number of lanes occupied by the accident vehicle and the speed limit of surrounding vehicles. Based on the synergy of the vehicle following model, lane changing model and traffic signal control model, the propagation and changes of traffic flow around the accident point are observed. As the simulation progresses, the scope of the area affected by the accident where the speed is significantly reduced, the traffic flow is backlogged or the vehicle is forced to detour is tracked, and these areas gradually form associated diffusion areas. For example, on the main road where the accident occurred, vehicles gradually formed queues, and the adjacent branches saw increased traffic volume and congestion due to vehicle diversion. The model accurately determined the boundaries and scope of the associated diffusion area through comprehensive analysis and regional definition of these traffic flow changes, providing a clear spatial scope basis for further analysis of related road sections and related traffic data, so as to formulate more targeted traffic diversion strategies and resource allocation plans.

[0051] Based on the identified associated diffusion area, accurate screening is performed in the associated road network. Using the boundary of the area as a limiting condition, all road sections that intersect or are included in it are identified, and these sections are multiple associated sections. For example, if the associated diffusion area covers a city's main road and several branch roads around it, then the corresponding parts of these main roads and branch roads are identified as associated sections, which constitute a subset of roads that are more directly and critically affected by the accident.

[0052] With the associated diffusion area and multiple associated road sections as dual constraints, we conduct data extraction in-depth on the associated road data. From the massive amount of associated road data, we screen out the traffic flow data related to these specific associated road sections within the associated diffusion area. For each associated road section, according to its location and length and other attributes in the associated diffusion area, we extract the corresponding information fragments such as traffic volume, speed, and vehicle density to form multiple traffic data slices. For example, for a specific section of an associated trunk road, we extract its traffic volume change data for a period of time after the accident, the speed distribution data at different time periods, and the vehicle density data, etc. Each data slice focuses on the traffic flow characteristics of a specific associated road section within the associated diffusion area.

[0053] A multi-threaded processing mechanism is started for multiple traffic data slices. Each thread is precisely assigned the data processing task of a specific associated road section, focusing on the reverse analysis of traffic flow on that road section. The thread first reads the traffic data slice corresponding to the associated road section, which covers key information such as traffic volume, speed, and vehicle density. Subsequently, the reverse model constructed based on traffic flow theory begins to operate, comprehensively considering the following characteristics of the vehicle, and simulating the driving sequence of the vehicle on the road section by analyzing the relationship between the front and rear vehicle distance and speed; combining the lane change rules, the possibility of vehicle transfer between different lanes and its impact on the direction of traffic flow is determined; the periodic change factors of traffic signals are incorporated to determine the start, stop and turn behavior of vehicles under the control of traffic lights. Based on these analyses, each thread independently predicts the traffic flow that may enter the accident location on the road section at different time intervals. For example, based on the traffic conditions of the current road section, the thread will trace back where the vehicles converged and how they evolved when passing through a series of intersections and sections, so as to estimate the scale and change trend of traffic flow heading to the accident area in the future period. During the multi-threaded parallel computing process, each thread maintains efficient data interaction and synchronization, integrates and optimizes their respective prediction results, and ultimately forms comprehensive and accurate correlated traffic flow prediction information, providing key data support for the formulation of subsequent traffic diversion strategies, so as to more effectively respond to traffic pressure and challenges caused by traffic accidents.

[0054] In a possible implementation, step S450 further includes:

[0055] Step S451: historical traffic data is called for the multiple associated road sections and traffic accident locations, and traffic flow diversion features are extracted based on the call results to obtain multiple diversion ratio coefficients.

[0056] Step S452: Calculate and obtain multiple initial associated vehicle flows based on the multiple diversion ratio coefficients and multiple vehicle flow data slice mappings.

[0057] Step S453: Anonymized driving route data collection is performed on the multiple vehicle flow data slices to obtain multiple driving route data sets.

[0058] Step S454: traversing the plurality of driving route data sets using the traffic accident location to obtain a plurality of real-time diversion coefficients.

[0059] Step S455: After the primary data of the multiple initial associated traffic flows are updated according to the distribution relationship of the multiple associated road sections, the multiple real-time diversion coefficients are used to perform secondary updates on the multiple initial associated traffic flows to obtain multiple sliced ​​traffic flow prediction information.

[0060] Step S456: Fusing the plurality of sliced ​​traffic flow prediction information to obtain the associated traffic flow prediction information.

[0061] Specifically, for multiple related road sections and traffic accident locations, we deeply explore the historical traffic database and accurately call relevant data. Through detailed analysis of these historical data, we extract the diversion characteristics of traffic flow under similar traffic conditions. For example, we analyze the diversion ratio of traffic entering the accident area from different directions and at different times on the nearby accident site and surrounding related road sections in the past, and calculate multiple diversion ratio coefficients. These coefficients can reflect the relative contribution of each related road section to traffic diversion in historical experience.

[0062] Based on the obtained multiple diversion ratio coefficients, multiple traffic data slices are mapped and calculated. The diversion ratio coefficient is used as a weight to perform weighted calculations with data such as traffic volume in the traffic data slice, so as to preliminarily estimate multiple initial associated traffic flows. This step makes a preliminary predictive integration of the current traffic flow conditions based on historical diversion rules, laying the foundation for subsequent more accurate analysis.

[0063] Anonymized driving route data collection is carried out for multiple traffic flow data segments. Using a variety of traffic monitoring technologies and data collection methods, the driving route information of vehicles on each associated road section is collected, but the specific identification of the vehicle is anonymized during the collection process to protect privacy. These driving route data are classified and sorted according to different associated road sections to form multiple driving route data sets, each of which reflects the driving path distribution of vehicles on the corresponding associated road section.

[0064] Taking the traffic accident location as the core reference point, we traverse multiple driving route data sets and calculate multiple real-time diversion coefficients by analyzing the relative relationship between each driving route and the traffic accident location, such as the changes in vehicle steering and diversion when approaching the accident location. These real-time diversion coefficients can better reflect the diversion trend of vehicles under current actual traffic conditions compared to historical diversion ratio coefficients, because they are generated based on current driving route data.

[0065] According to the distribution relationship of multiple associated road sections, the first-level data of multiple initial associated traffic flows is updated, and the traffic flow convergence and evacuation rules between the associated road sections are considered, such as the traffic flow exchange relationship between the main road and the branch road, to adjust and optimize the initial associated traffic flow. After that, multiple real-time diversion coefficients are used for secondary update, and the real-time diversion coefficients and the associated traffic flow after the first-level update are weighted again. Operations such as calculation, etc. are performed to obtain multiple sliced ​​traffic flow prediction information. Each sliced ​​traffic flow prediction information accurately reflects the traffic flow prediction of the corresponding associated road section after considering the historical diversion rules, real-time driving routes, and the distribution relationship of the road sections.

[0066] First, a comprehensive data fusion framework is constructed. Based on the continuity principle and spatial correlation principle of traffic flow, the framework regards each slice of traffic flow prediction information as a local representation of the overall traffic flow. Using the weighted average fusion algorithm, the corresponding weight is assigned to each slice of traffic flow prediction information according to the importance of the associated road section corresponding to each slice in the entire associated road network and the reliability of the data. For example, the slice traffic flow prediction information of the associated road section of the main road will be given a higher weight due to its large traffic volume and significant impact on the overall traffic flow; while the slice traffic flow prediction information of some branch roads will be given a relatively low weight. Then, the key elements such as traffic volume, speed, and vehicle density in each slice of traffic flow prediction information are fused and calculated separately. In terms of traffic flow fusion, the traffic flow prediction values ​​of each slice are multiplied by the corresponding weights and then summed to obtain the comprehensive traffic flow prediction value; for speed fusion, the weighted sum is also used, and corrections are made according to the road type and traffic control conditions to ensure that the fused speed conforms to the actual traffic operation logic; when fusion of vehicle density, the spatial range differences of each slice area are considered, and normalization is performed before weighted fusion. At the same time, Kalman filtering technology is used to optimize the fusion results. Kalman filtering predicts the current state based on the fusion results of the previous moment through two steps of prediction and update, and then corrects the observations in the traffic flow prediction information of each slice at the current moment, thereby improving the accuracy and stability of the associated traffic flow prediction information. After such a rigorous fusion process, the associated traffic flow prediction information that can comprehensively and accurately reflect the future traffic flow change trend of the entire associated road network under the influence of traffic accidents is finally obtained, including the dynamic changes of traffic flow, speed, and vehicle density of each associated road section at different time points, providing solid and reliable data support for subsequent traffic diversion decisions, signal light timing control, etc.

[0067] In a possible implementation, step S500 further includes:

[0068] Step S510: locating an associated inflow end and an associated outflow end in the associated road network according to the traffic accident location.

[0069] Step S520: locating H diversion adaptation roads in the associated road network according to the associated inflow end and the associated outflow end.

[0070] Step S530: extracting H real-time traffic flow data of the H diversion adaptation roads from the associated road data.

[0071] Step S540: interactively obtaining H traffic carrying capacities of the H diversion adaptation roads, and calculating and obtaining H real-time carrying capacities based on the H traffic carrying capacities and H real-time traffic flow data.

[0072] Step S550: extracting and obtaining H diversion road features of the H diversion adaptive roads from the associated road network, wherein each of the H diversion road features includes the number of intersections, the number of traffic lights, and the number of road turns.

[0073] Step S560: comprehensively evaluating the H diversion adaptation roads using the H real-time carrying capacities and the H diversion road characteristics to obtain H diversion fitnesses.

[0074] Step S570: Filtering the H diversion adaptive roads according to the H diversion fitnesses to obtain the multiple diversion candidate roads that meet a preset fitness threshold.

[0075] Specifically, with the determined traffic accident location as the key basis, we carry out precise positioning operations in the related road network. We comprehensively analyze the connection status, traffic flow direction, and historical traffic data of each road in the network, trace back along the path of traffic flow to the accident location, and comprehensively consider factors such as the road's capacity and number of lanes to accurately determine the related inflow end, that is, the starting area of ​​the road where traffic flow converges and flows to the accident point; at the same time, based on the direction of traffic flow evacuation at the accident location, combined with the connectivity and traffic control of surrounding roads, we accurately find the related outflow end, that is, the starting point where traffic flow flows out and disperses in the accident area.

[0076] With the located associated inflow and outflow, H diversion-adaptive roads are further selected in the associated road network. By analyzing the connectivity between the road and the inflow and outflow, the rationality of the traffic flow, and the basic conditions of the road itself, the roads with the potential to undertake diversion tasks in the event of an accident are selected. These roads constitute the basic road set for the subsequent diversion strategy formulation.

[0077] The H real-time traffic flow data corresponding to each of the H diversion adaptive roads are extracted from the associated road data resources. These data record in detail the current traffic operation status of each diversion adaptive road, including key information such as the size of traffic flow, vehicle speed, and vehicle distribution density, providing real-time data support for evaluating the diversion feasibility of each road.

[0078] With the help of the interaction function with other modules of the traffic management system or related databases, H traffic carrying capacity data of H diversion adaptive roads are obtained. These data cover the carrying upper limit determined by comprehensive factors such as the maximum traffic flow, number of lanes, road grade, etc. of the road design. Then, based on the obtained traffic carrying capacity and real-time traffic flow data, H real-time carrying capacities are obtained by calculating the proportional relationship between real-time traffic flow and traffic carrying capacity, so as to measure the current remaining carrying space and diversion potential of each diversion adaptive road.

[0079] H diversion road features of H diversion adaptive roads are extracted from the detailed information database of the associated road network. Each road diversion feature contains important information such as the number of intersections, the number of traffic lights, the number of road turns, etc. These features have a significant impact on the smoothness, speed and efficiency of the diversion traffic flow, and are an important basis for the comprehensive evaluation of diversion adaptive roads.

[0080] The following diversion fitness formula is constructed to comprehensively evaluate H diversion fitness roads and obtain H diversion fitnesses: Among them, F i represents the diversion fitness of the i-th diversion adaptation road, C i is the traffic carrying capacity of the i-th diversion adaptation road (such as the maximum traffic volume that can be accommodated per unit time); R i is the traffic volume in the real-time traffic data of the i-th diversion adaptation road; I i is the number of intersections of the i-th diversion-adaptive road; T i is the number of traffic lights on the i-th diversion adaptation road; W i is the number of turns on the i-th diversion adaptive road; α, β, γ, δ are weight coefficients, and α+β+γ+δ=1. These weight coefficients can be adjusted according to the actual traffic management focus and experience. For example, if the current traffic conditions focus more on the remaining carrying capacity of the road, α can be set relatively large; if more attention is paid to the smoothness of the road, the coefficients related to road traffic obstruction in β, γ, and δ can be appropriately increased. Through this formula, the real-time carrying capacity and diversion road characteristics can be comprehensively considered, and a quantitative diversion fitness evaluation result can be given for each diversion adaptive road, so as to subsequently screen out more suitable diversion candidate roads.

[0081] Based on the calculated H diversion fitness, the pre-set fitness threshold is compared and screened. The diversion adaptation roads that meet the pre-set fitness threshold are determined as multiple diversion candidate roads. These candidate roads will become the core objects for subsequent traffic diversion strategy formulation, signal light configuration optimization and other operations, so that when a traffic accident occurs, the traffic flow can be effectively guided for reasonable diversion, the traffic pressure in the accident area can be reduced, and the normal operation of the entire transportation network can be guaranteed.

[0082] In a possible implementation, step S600 further includes:

[0083] Step S610: extracting multiple diversion fitnesses of the multiple diversion candidate roads from the H diversion fitnesses.

[0084] Step S620: extracting multiple real-time carrying capacities of the multiple diversion candidate roads from the H real-time carrying capacities.

[0085] Step S630: Performing diversion weight configuration according to the multiple diversion fitnesses and multiple real-time carrying capabilities to obtain multiple diversion weight coefficients.

[0086] Step S640: Calculate and obtain multiple real-time diversion demands based on the associated traffic flow prediction information and multiple diversion weight coefficients.

[0087] Step S650: performing diversion fitting of the multiple real-time diversion demands according to the road connection relationship between the multiple associated road sections and the multiple diversion candidate roads, and configuring the traffic lights of the multiple diversion candidate roads according to the fitting results.

[0088] Step S660: Similarly, 1 / K of the traffic control duration is used as a diversion update window to update the associated traffic flow prediction information, and the traffic light configurations of the multiple diversion candidate roads are dynamically updated based on the update result.

[0089] Specifically, the H calculated diversion fitness data are systematically sorted out and identified. Taking the previously determined multiple diversion candidate roads as the target objects, through data retrieval and matching technology, the diversion fitness values ​​corresponding to these diversion candidate roads are accurately screened out in the complete H diversion fitness data set. This process is like accurately locating specific elements in a huge information library. With the help of preset road signs or index information, the diversion fitness of each diversion candidate road is separated and extracted from a large number of data, ensuring that the subsequent diversion weight configuration based on these data can be carried out accurately, providing a solid data foundation for the scientific formulation of the entire traffic diversion plan, so that the diversion strategy can be closely built around the characteristics and advantages of these representative diversion candidate roads, thereby effectively improving the diversion efficiency and operation stability of the traffic network under the influence of traffic accidents.

[0090] Faced with the H sets of real-time carrying capacity data that have been acquired, multiple diversion candidate roads are used as the core screening basis, and data screening and matching algorithms are used to go deep into the H real-time carrying capacity data, carefully identify and lock the real-time carrying capacity information corresponding to each diversion candidate road. This requires a clear understanding of the data structure, and the real-time carrying capacity data of the diversion candidate roads are separated from the overall data through key identification information such as road numbers and section names. These extracted real-time carrying capacity data can intuitively reflect the actual space and potential for each diversion candidate road to accommodate additional traffic flow under the current traffic conditions, providing indispensable key data support for the subsequent reasonable configuration of diversion weights and scientific planning of traffic diversion strategies, ensuring that the diversion plan can fully consider the actual carrying conditions of each candidate road to achieve efficient and orderly traffic diversion and guidance goals.

[0091] Based on the extracted multiple diversion fitness and multiple real-time carrying capacities, a multi-factor weighted decision algorithm is used to configure the diversion weight. First, the diversion fitness and real-time carrying capacity are standardized and their values ​​are mapped to the same dimensional range for subsequent comprehensive evaluation. For the diversion fitness, a basic weight factor is determined based on its importance in the overall evaluation system. This factor can be determined by technical means such as the analytic hierarchy process (AHP) to reflect its potential influence on the diversion effect. For the real-time carrying capacity, the corresponding weight factor is also assigned based on traffic flow theory and actual experience. Then, a diversion weight coefficient calculation model is constructed, for example, using a linear weighted summation formula: diversion weight coefficient = diversion fitness weight × standardized diversion fitness + real-time carrying capacity weight × standardized real-time carrying capacity. In the calculation process, the standardized diversion fitness and standardized real-time carrying capacity of each diversion candidate road are substituted into the formula, and the synergistic influence of the two is comprehensively considered through the adjustment of the weight factor. The weight factors are iteratively optimized using optimization algorithms in machine learning, such as the gradient descent algorithm, to ensure that the calculated diversion weight coefficients can meet the goal of efficient traffic flow diversion to the greatest extent possible. Ultimately, the corresponding diversion weight coefficients are determined for each diversion candidate road. These coefficients will serve as the key basis for the subsequent accurate calculation of real-time diversion needs and the formulation of traffic light configuration strategies, thereby achieving scientific and reasonable diversion and guidance of traffic flow in accident scenarios.

[0092] Based on the existing associated traffic flow prediction information and the calculated multiple diversion weight coefficients, multiple real-time diversion demands are calculated with the help of traffic flow allocation algorithm. First, the associated traffic flow prediction information is parsed to extract the key data such as the expected traffic flow, speed, and vehicle density of each associated road section at different time points, and the traffic flow matrix is ​​constructed to represent the traffic flow state sequence of each associated road section. Then, multiple diversion weight coefficients are integrated into a weight vector, and each element of the vector corresponds to the weight coefficient of a diversion candidate road. Using the allocation model based on the traffic flow conservation principle and the allocation model constructed based on the Wardrop equilibrium criterion, the traffic flow matrix and the weight vector are used as input parameters. Through iterative calculation, the traffic flow gradually reaches the flow distribution equilibrium state on each diversion candidate road under the condition of satisfying the diversion weight coefficient constraint. In the iterative process, according to the weight coefficient of each diversion candidate road, the traffic flow of the associated road section is proportionally allocated to each diversion candidate road, and the real-time diversion demand of each diversion candidate road at each time point is calculated. This process involves complex matrix operations and optimization algorithms, such as the use of Newton-Raphson iteration method to solve the nonlinear equations in the traffic flow distribution model to ensure the accuracy and stability of the calculation results, thereby providing accurate flow data basis for subsequent diversion fitting and traffic light configuration, and realizing reasonable and orderly diversion of traffic flow in traffic accident scenarios.

[0093] Graph theory algorithms are used to deeply analyze the road connections between multiple associated road sections and multiple diversion candidate roads. By constructing a directed graph, road sections are represented as edges, intersections are represented as nodes, and the weights of the edges are marked as the road section's capacity, length and other attribute information, thus clearly presenting the topological structure of the road network. Based on this structure, the minimum cost maximum flow algorithm is used to perform diversion fitting for multiple real-time diversion demands. The algorithm aims to minimize the transportation cost of traffic from associated road sections to diversion candidate roads, comprehensively considering factors such as the congestion level of the road section and turning restrictions, and gradually adjusting the traffic flow distribution on each road section through continuous iteration to find augmented paths until the optimal diversion solution is reached, realizing the reasonable distribution of traffic demand among the candidate roads. Subsequently, based on the results of the diversion fitting, the signal light configuration work was carried out. The hybrid algorithm combining timing control and induction control in the intelligent transportation system was used to set the signal light parameters for each intersection of the candidate diversion road. For time periods and sections with relatively stable traffic, the timing control algorithm was used. According to historical traffic data and the fitted traffic demand, the cycle duration and green-to-signal ratio of the signal light were pre-set to ensure the regular passage of vehicles. For situations with large fluctuations in traffic flow, the induction control algorithm was enabled. By setting up geomagnetic sensors, cameras and other detection equipment at the intersection, the traffic flow, queue length and other information were monitored in real time, and the phase and duration of the signal light were dynamically adjusted. When the traffic flow in a certain direction suddenly increases, the green light time in that direction is promptly extended to reduce the waiting time of vehicles. Thus, through the precise control of the signal light, the traffic flow is smoothly diverted to each candidate road, and efficient diversion around the accident area is achieved, effectively alleviating the impact of traffic accidents on the traffic network and ensuring the smooth operation of traffic.

[0094] The fixed diversion update window period is established with 1 / K of the traffic control duration. When each diversion update window comes, various types of traffic data newly emerged during the period are quickly integrated. These data cover the details of the increase and decrease of traffic volume in each section of the associated road network, the change of vehicle speed, and the change of vehicle driving trajectory. With the help of these rich and real-time data resources, a precise and efficient traffic reverse analysis algorithm is used to comprehensively update the previously obtained associated traffic prediction information (i.e., the traffic reverse analysis results). This algorithm fully considers the complex dynamic characteristics of traffic flow, and re-analyzes and calculates the source, direction, and convergence and dispersion rules of traffic flow based on new data, so as to accurately obtain the updated associated traffic prediction information. Then, based on the updated associated traffic prediction information as the core basis, the dynamic update process of the signal light configuration for multiple diversion candidate roads is immediately started. By comprehensively weighing the changes in traffic load of each candidate road under the updated traffic flow forecast scenario, the real-time traffic capacity of the road itself, the need for coordination with adjacent roads and other key factors, an intelligent and flexible signal timing optimization algorithm is used to recalculate the core parameters such as the cycle duration, green-to-signal ratio and precise order of phase switching of the signal lights of each candidate road. Through this continuous cycle and dynamically adaptive update mechanism, it is ensured that during the long period of traffic control, the configuration of the signal lights can always be closely matched with the actual traffic flow conditions, effectively guiding the traffic flow to achieve efficient, smooth and orderly diversion between multiple candidate roads, minimizing the congestion interference and adverse effects of traffic accidents on the surrounding traffic network, and fully ensuring that the traffic system can still maintain a stable, efficient and orderly operation under special circumstances.

[0095] In a possible implementation, step S200 further includes:

[0096] Step S210: The interactive GIS system obtains the road type characteristics of the traffic accident location.

[0097] Step S220: extracting accident road condition data from the plurality of road condition data blocks according to the road block identifier.

[0098] Step S230: Synchronizing the road type characteristics, accident road condition data and traffic accident characteristics to a pre-built traffic control prediction model to perform a comprehensive prediction of the control time consumption and output the traffic control duration.

[0099] Specifically, it interacts with GIS (Geographic Information System). By sending a request to the GIS system and providing relevant information about the location of the traffic accident, the road type characteristics corresponding to the location are obtained. These characteristics include whether the road belongs to a highway, a city trunk road, a secondary trunk road or a branch road, as well as detailed information such as the number of lanes on the road, whether there is a central dividing strip, and the alignment of the road (straight line, curve, slope, etc.). These data are critical for the subsequent accurate assessment of the scale and scope of impact of traffic control.

[0100] According to the specific road block identification, accurate screening is performed in the existing multiple block traffic data. Each block traffic data covers the traffic condition information in a certain area, such as traffic volume, vehicle speed, road occupancy, etc. Through the road block identification, the road condition data of the block where the accident occurred can be quickly located, that is, the accident road condition data. These data reflect the actual operating status of the surrounding roads at the moment of the accident and are one of the important bases for predicting the duration of traffic control.

[0101] Data cleaning and format conversion are performed on road type characteristics, accident road condition data, and traffic accident characteristics to ensure that they can be accurately identified and processed by the traffic control prediction model. Using the decision tree algorithm, the model uses these data as input nodes to start building a decision process. For road type characteristics, if the road is a highway, its branches may lead to further analysis of factors such as multiple lanes and high design speeds; if it is a city trunk road, sub-nodes such as intersection spacing, surrounding commercial and residential area distribution are considered. In the accident road condition data, situations with large traffic volume, extremely low speed, and high road occupancy will be given higher weights in the corresponding branches of the decision tree, because this usually means that the accident seriously hinders traffic. In the traffic accident characteristics, serious accidents (such as multi-vehicle chain collisions, casualties, etc.) will give priority to branch paths with longer control times in the hierarchical judgment of the decision tree. As the data is screened and judged layer by layer in each branch of the decision tree, the model comprehensively considers the interactive effects between various characteristics. For example, even if it is a main urban road, if the accident is minor and the traffic can be quickly evacuated after the accident (based on the accident road condition data), then the control time may be relatively short. The decision tree branching rules and thresholds obtained by training a large amount of historical data ultimately determine a path from the root node to the leaf node, and the numerical range corresponding to the leaf node is the predicted traffic control duration. This duration combines multiple factors such as the inherent attributes of the road, the real-time status of the accident, and the severity of the accident itself, providing a more scientific and accurate time basis for subsequent traffic diversion planning, helping traffic management departments to rationally arrange resources and efficiently deal with traffic congestion and chaos caused by traffic accidents.

[0102] Embodiment 2, based on the same inventive concept as the method for optimizing the allocation of traffic resources in a smart city in the above embodiment, Figure 2As shown, the present application provides a smart city traffic resource optimization configuration platform, and the platform and method embodiments in the present application embodiments are based on the same inventive concept. Among them, the platform includes:

[0103] The traffic accident feature acquisition module 10 is used to monitor the road conditions through the intelligent transportation system. When a road accident is detected, the accident detection task is started to collect accident information to obtain the traffic accident location and traffic accident features.

[0104] The traffic control duration acquisition module 20 is used to make a time consumption prediction according to the traffic accident characteristics to obtain the traffic control duration.

[0105] The associated road data acquisition module 30 is used to obtain the associated road network of the traffic accident location in the interactive GIS system, and then call the real-time traffic data from the intelligent transportation system with the associated road network as a constraint to obtain the associated road data.

[0106] The associated traffic flow prediction information acquisition module 40 is used to perform multi-threaded traffic flow reverse analysis using the associated road data with the traffic accident location and traffic control duration as constraints to obtain associated traffic flow prediction information.

[0107] The diversion performance evaluation module 50 is used to evaluate the diversion performance of the associated road network according to the associated traffic flow prediction information, and locate multiple diversion candidate roads based on the evaluation results.

[0108] The dynamic update module 60 is used to dynamically update the signal light configurations of the multiple candidate diversion roads by taking 1 / K of the traffic control duration as a diversion update window and taking the associated traffic flow prediction information as a constraint.

[0109] Furthermore, the traffic accident feature acquisition module 10 also includes:

[0110] The block road condition data acquisition unit is used in the intelligent transportation system to perform real-time road condition monitoring on multiple monitoring road blocks through multiple parallel road condition data acquisition tasks of the road condition data acquisition container to obtain multiple block road condition data.

[0111] A block road condition data cross-container publishing unit, wherein the block road condition data cross-container publishing unit is used for the road condition data collection container to publish the multiple block road condition data across containers to the road condition anomaly analysis container, wherein the road condition anomaly analysis container is deployed in the intelligent transportation system.

[0112] A road accident warning output unit, wherein the road accident warning output unit is used for multiple parallel road condition anomaly analysis tasks in the road condition anomaly analysis container to perform road anomaly analysis based on the multiple block road condition data and output a road accident warning, wherein the road accident warning has a road block identifier.

[0113] An accident detection container activation unit is used by the intelligent transportation system to activate an accident detection container according to the road accident warning and to assign the highest priority to the accident detection container.

[0114] An accident information collection unit is used to collect accident information in the accident detection container, using the road block identifier as a constraint, run the accident detection task, and obtain the traffic accident location and traffic accident characteristics.

[0115] Furthermore, the associated traffic flow prediction information acquisition module 40 further includes:

[0116] A traffic simulation model construction unit is used to construct a traffic simulation model based on the associated road network and to locally update the traffic simulation model using the associated road data.

[0117] The associated diffusion area acquisition unit is used to use the traffic accident location and traffic control duration to predict the accident impact range in the traffic simulation model to obtain the associated diffusion area.

[0118] A plurality of associated road segment acquisition units are used to obtain a plurality of associated road segments by selecting a frame in an associated road network using the associated diffusion area.

[0119] A vehicle flow data fragment acquisition unit is used to extract and obtain multiple vehicle flow data fragments from the associated road data with the associated diffusion area and multiple associated road sections as constraints.

[0120] The associated traffic flow prediction information acquisition unit is used to use the multiple traffic flow data slices to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information.

[0121] Furthermore, the associated traffic flow prediction information acquisition unit further includes:

[0122] The diversion ratio coefficient acquisition unit is used to call historical traffic data for the multiple associated road sections and traffic accident locations, and extract traffic diversion features based on the call results to obtain multiple diversion ratio coefficients.

[0123] A plurality of initial associated vehicle flow acquisition units are used to calculate and obtain a plurality of initial associated vehicle flows according to the plurality of diversion ratio coefficients and a plurality of vehicle flow data slicing mappings.

[0124] A driving route data set acquisition unit is used to collect anonymized driving route data for the multiple vehicle flow data slices to obtain multiple driving route data sets.

[0125] A real-time diversion coefficient acquisition unit is used to traverse the multiple driving route data sets using the traffic accident location to obtain multiple real-time diversion coefficients.

[0126] The pieced traffic flow prediction information acquisition unit is used to perform a primary data update of the multiple initial associated traffic flows according to the distribution relationship of the multiple associated road sections, and then use the multiple real-time diversion coefficients to perform a secondary update of the multiple initial associated traffic flows to obtain multiple pieced traffic flow prediction information.

[0127] The sliced ​​traffic flow prediction information fusion unit is used to fuse the multiple sliced ​​traffic flow prediction information to obtain the associated traffic flow prediction information.

[0128] Furthermore, the diversion performance evaluation module 50 further includes:

[0129] An associated road network positioning unit is used to locate an associated inflow end and an associated outflow end in the associated road network according to the location of the traffic accident.

[0130] A diversion adaptive road positioning unit is used to locate H diversion adaptive roads in the associated road network according to the associated confluence end and the associated outflow end.

[0131] A real-time traffic flow data extraction unit is used to extract H real-time traffic flow data of the H diversion adaptation roads from the associated road data.

[0132] A real-time carrying capacity acquisition unit, wherein the real-time carrying capacity acquisition unit is used to interactively obtain H traffic carrying capacities of the H diversion adaptation roads, and calculate and obtain H real-time carrying capacities based on the H traffic carrying capacities and H real-time traffic flow data.

[0133] A diversion road feature extraction unit is used to extract H diversion road features of the H diversion adaptive roads from the associated road network, wherein each of the H diversion road features includes the number of intersections, the number of traffic lights, and the number of road turns.

[0134] The diversion fitness acquisition unit is used to comprehensively evaluate the H diversion adaptation roads by using the H real-time carrying capacities and the H diversion road characteristics to obtain H diversion fitnesses.

[0135] A diversion candidate road acquisition unit is used to filter and obtain the plurality of diversion candidate roads that meet a preset fitness threshold from the H diversion adaptive roads according to the H diversion fitnesses.

[0136] Furthermore, the dynamic update module 60 also includes:

[0137] A diversion fitness extraction unit is used to extract multiple diversion fitnesses of the multiple diversion candidate roads from the H diversion fitnesses.

[0138] A real-time carrying capacity extraction unit is used to extract multiple real-time carrying capacities of the multiple diversion candidate roads from the H real-time carrying capacities.

[0139] The shunt weight coefficient acquisition unit is used to configure the shunt weight according to the multiple shunt fitnesses and the multiple real-time carrying capacities to obtain multiple shunt weight coefficients.

[0140] A real-time diversion demand acquisition unit is used to calculate and obtain multiple real-time diversion demands based on the associated traffic flow prediction information and multiple diversion weight coefficients.

[0141] A diversion fitting unit is used to perform diversion fitting of the multiple real-time diversion demands according to the road connection relationship between the multiple associated road sections and the multiple diversion candidate roads, and to configure the traffic lights of the multiple diversion candidate roads according to the fitting results.

[0142] An associated traffic flow prediction information updating unit is used to update the associated traffic flow prediction information by analogy, taking 1 / K of the traffic control duration as a diversion update window, and dynamically updating the signal light configurations of the multiple diversion candidate roads based on the update results.

[0143] Furthermore, the traffic control duration acquisition module 20 also includes:

[0144] A road type feature acquisition unit is used to interact with the GIS system to obtain the road type features of the traffic accident location.

[0145] An accident road condition data extraction unit is used to extract the accident road condition data from the plurality of block road condition data according to the road block identifier.

[0146] A traffic control duration output unit is used to synchronize the road type characteristics, accident road condition data and traffic accident characteristics to a pre-built traffic control prediction model to make a comprehensive prediction of the control time and output the traffic control duration.

[0147] Embodiment three, based on the same inventive concept as the method for optimizing the allocation of traffic resources in a smart city in the aforementioned embodiment, the present application also provides a computer-readable storage medium, on which a computer program is stored, and the computer program implements the steps of any one of the methods described in the aforementioned embodiment one when executed.

[0148] It should be noted that the above-mentioned sequence of the embodiments of the present application is only for description and does not represent the advantages and disadvantages of the embodiments. And the above-mentioned specific embodiments of this specification are described. In addition, the processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0149] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application should be included in the protection scope of the present application.

[0150] This specification and the drawings are merely exemplary illustrations of the present application and are deemed to cover any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these modifications and variations.

Claims

1. A method for optimizing the allocation of traffic resources in a smart city, characterized in that: The method comprises: The intelligent transportation system is used to monitor road conditions. When a road accident is detected, the accident detection task is initiated to collect accident information and obtain the location and characteristics of the traffic accident. According to the characteristics of the traffic accident, a time consumption prediction is performed to obtain the traffic control duration; After the interactive GIS system obtains the associated road network of the traffic accident location, real-time traffic data is called from the intelligent transportation system with the associated road network as a constraint to obtain associated road data; Taking the traffic accident location and traffic control duration as constraints, the associated road data is used to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information; Performing a diversion performance evaluation on the associated road network according to the associated traffic flow prediction information, and locating a plurality of diversion candidate roads based on the evaluation results; 1 / K of the traffic control duration is used as a diversion update window, and the signal light configurations of the multiple diversion candidate roads are dynamically updated based on the associated traffic flow prediction information.

2. The method for optimizing the allocation of traffic resources in a smart city according to claim 1, characterized in that: The method includes: monitoring the road condition by an intelligent transportation system, and when a road accident is detected, starting an accident detection task to collect accident information, and obtaining the location and characteristics of the traffic accident. In the intelligent transportation system, multiple parallel road condition data collection tasks of the road condition data collection container are used to perform real-time road condition monitoring on multiple monitoring road blocks to obtain road condition data of multiple blocks; The road condition data collection container publishes the multiple block road condition data across containers to a road condition anomaly analysis container, wherein the road condition anomaly analysis container is deployed in the intelligent transportation system; The multiple parallel road condition anomaly analysis tasks in the road condition anomaly analysis container perform road anomaly analysis based on the multiple block road condition data and output a road accident warning, wherein the road accident warning has a road block identifier; The intelligent transportation system activates the accident detection container according to the road accident warning and assigns the highest priority to the accident detection container; In the accident detection container, the road block identifier is used as a constraint to run the accident detection task to collect accident information and obtain the traffic accident location and traffic accident characteristics.

3. The method for optimizing the allocation of traffic resources in a smart city according to claim 2, characterized in that: Taking the traffic accident location and traffic control duration as constraints, the method uses the associated road data to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information, and includes: Building a traffic simulation model based on the associated road network, and using the associated road data to perform local updates on the traffic simulation model; The traffic accident location and traffic control duration are used to predict the accident impact range in the traffic simulation model to obtain the associated diffusion area; Using the associated diffusion area to select a plurality of associated road sections in an associated road network; Taking the associated diffusion area and the multiple associated road sections as constraints, extracting multiple traffic flow data slices from the associated road data; The multiple traffic flow data slices are used to perform multi-threaded traffic flow reverse analysis to obtain associated traffic flow prediction information.

4. The method for optimizing the allocation of traffic resources in a smart city according to claim 3, characterized in that: The method comprises: performing multi-threaded traffic flow reverse analysis using the plurality of traffic flow data slices to obtain associated traffic flow prediction information; Calling historical traffic data for the multiple associated road sections and traffic accident locations, and extracting traffic flow diversion features based on the calling results to obtain multiple diversion ratio coefficients; Calculate and obtain multiple initial associated vehicle flows according to the multiple diversion ratio coefficients and multiple vehicle flow data slice mappings; Anonymizing the driving route data collection of the multiple vehicle flow data slices to obtain multiple driving route data sets; Using the traffic accident location to traverse the multiple driving route data sets to obtain multiple real-time diversion coefficients; After the primary data of the multiple initial associated traffic flows are updated according to the distribution relationship of the multiple associated road sections, the multiple real-time diversion coefficients are used to perform secondary updates on the multiple initial associated traffic flows to obtain multiple sliced ​​traffic flow prediction information; The plurality of sliced ​​traffic flow prediction information are integrated to obtain the associated traffic flow prediction information.

5. The method for optimizing the allocation of traffic resources in a smart city according to claim 4, characterized in that: The method includes: evaluating the diversion performance of the associated road network according to the associated traffic flow prediction information, and locating a plurality of candidate diversion roads based on the evaluation results. Locate an associated inflow end and an associated outflow end in the associated road network according to the location of the traffic accident; Positioning H diversion adaptation roads in the associated road network according to the associated inflow end and the associated outflow end; Extracting H real-time traffic flow data of the H diversion adaptation roads from the associated road data; Interactively obtain H traffic carrying capacities of the H diversion adaptive roads, and calculate and obtain H real-time carrying capacities according to the H traffic carrying capacities and H real-time traffic flow data; Extracting H diversion road features of the H diversion adaptive roads from the associated road network, wherein each of the H diversion road features includes the number of intersections, the number of traffic lights, and the number of road turns; Comprehensively evaluating the H diversion adaptation roads by using the H real-time carrying capacities and the H diversion road characteristics to obtain H diversion fitnesses; The plurality of candidate diversion roads satisfying a preset fitness threshold are obtained by screening the H diversion adaptation roads according to the H diversion fitnesses.

6. The method for optimizing the allocation of traffic resources in a smart city according to claim 5, characterized in that: Using 1 / K of the traffic control duration as a diversion update window and dynamically updating the signal light configurations of the plurality of diversion candidate roads with the associated traffic flow prediction information as a constraint, the method comprises: Extracting a plurality of diversion fitnesses of the plurality of diversion candidate roads from the H diversion fitnesses; Extracting multiple real-time carrying capacities of the multiple diversion candidate roads from the H real-time carrying capacities; Performing diversion weight configuration according to the multiple diversion fitnesses and multiple real-time carrying capacities to obtain multiple diversion weight coefficients; Calculate and obtain multiple real-time diversion demands based on the associated traffic flow prediction information and multiple diversion weight coefficients; Performing diversion fitting of the multiple real-time diversion demands according to the road connection relationship between the multiple associated road sections and the multiple diversion candidate roads, and configuring the traffic lights of the multiple diversion candidate roads according to the fitting results; By analogy, 1 / K of the traffic control duration is used as a diversion update window to update the associated traffic flow prediction information, and the signal light configurations of the multiple diversion candidate roads are dynamically updated based on the update result.

7. The method for optimizing the allocation of traffic resources in a smart city according to claim 2, characterized in that: The time consumption is predicted according to the traffic accident characteristics to obtain the traffic control duration, and the method includes: The interactive GIS system obtains the road type characteristics of the traffic accident location; Extracting accident road condition data from the plurality of block road condition data according to the road block identifier; The road type characteristics, accident road condition data and traffic accident characteristics are synchronized to a pre-built traffic control prediction model to perform a comprehensive prediction of the control time consumption, and the traffic control duration is output.

8. Smart city traffic resource optimization configuration platform, characterized by: The smart city traffic resource optimization configuration platform includes: A traffic accident feature acquisition module, which is used to monitor road conditions through an intelligent transportation system. When a road accident is detected, an accident detection task is initiated to collect accident information and obtain the location and features of the traffic accident. A traffic control duration acquisition module, the traffic control duration acquisition module is used to perform time consumption prediction according to the traffic accident characteristics to obtain the traffic control duration; An associated road data acquisition module, wherein after the interactive GIS system obtains the associated road network of the traffic accident location, the associated road data acquisition module is used to call real-time traffic data from the intelligent transportation system with the associated road network as a constraint to obtain associated road data; An associated traffic flow prediction information acquisition module, the associated traffic flow prediction information acquisition module is used to perform multi-threaded traffic flow reverse analysis using the associated road data based on the traffic accident location and traffic control duration as constraints to obtain associated traffic flow prediction information; a diversion performance evaluation module, the diversion performance evaluation module being used to evaluate the diversion performance of the associated road network according to the associated traffic flow prediction information, and locate a plurality of diversion candidate roads based on the evaluation results; A dynamic update module is used to dynamically update the signal light configurations of the multiple diversion candidate roads by taking 1 / K of the traffic control duration as a diversion update window and taking the associated traffic flow prediction information as a constraint.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, which implements the steps of the method according to any one of claims 1 to 7 when executed.

Images