Traffic light calculation method and system based on Internet of Things

Abstract

The present application provides a method and system for calculating traffic lights based on the Internet of Things. Among them, the real-time data stream from the intelligent sensor network of the Internet of Things is received, and the spatiotemporal consistency is processed through high-precision spatiotemporal synchronization technology to generate a dynamic panoramic view of the traffic situation; the causal relationship is inferred by the causal inference algorithm, and the direct and indirect impact paths are calculated. The spatiotemporal sequence modeling technology is used to capture the spatiotemporal characteristics, simulate the time-varying laws, and generate a future traffic state prediction model; the evolutionary strategy optimization algorithm is used to simulate the natural evolution process, search for the optimal traffic light switching strategy, use multi-objective optimization technology to consider mutually exclusive goals, introduce constraints for comprehensive evaluation, and generate traffic light switching strategy evaluation results; introduce a feedback loop mechanism, collect model operation results in real time, automatically form an adaptive improvement report, and generate a traffic light passing calculation plan. The technical solution provided by the present application significantly improves the efficiency of traffic light passing calculations.

Description

Technical Field

[0001] The embodiments of the present application relate to the technical field of Internet of Things, and in particular to a traffic light passing calculation method and system based on Internet of Things. Background Art

[0002] With the development of smart cities, effective management and optimization of traffic flow has become one of the key requirements. This requires the ability to receive and process data from distributed IoT smart sensor networks in real time, integrate vehicle dynamic information with pedestrian flow statistics, and generate a dynamic panoramic view of traffic conditions through high-precision spatiotemporal synchronization technology. In addition, causal inference algorithms are required to identify the causal relationship between traffic events and their impact paths, and spatiotemporal sequence modeling technology is used to simulate the time-varying laws of traffic flow, thereby providing an accurate future traffic state prediction model. The ultimate goal is to generate the optimal traffic light switching strategy based on this model, while considering multiple goals such as minimizing waiting time and maximizing traffic flow.

[0003] Currently, many cities have begun to use smart sensors and basic traffic management systems to monitor and adjust traffic flow. These systems usually include basic data collection and analysis functions, which can monitor traffic flow to a certain extent and adjust the timing of traffic lights based on simple rules or historical data. Some advanced systems can even make predictive adjustments to traffic flow based on factors such as weather and special events.

[0004] However, existing solutions generally have several major problems. First, traffic lights have low computational efficiency and it is difficult to respond to rapidly changing traffic conditions in real time and efficiently. Second, due to the lack of understanding of complex causal relationships, existing systems are often unable to accurately identify the root causes of traffic congestion, limiting their ability to prevent and alleviate. Finally, existing optimization algorithms tend to find local optimal solutions rather than global optimal solutions, which may result in that although the traffic flow of certain sections of the road has improved, the performance of the overall traffic system has not been maximized. In addition, the lack of an effective feedback loop mechanism makes it difficult for the system to self-learn and adapt to new traffic patterns or environmental changes, further reducing the long-term effectiveness and reliability of the system. These problems jointly restrict the pace of urban traffic management to develop towards a higher level of intelligence. Summary of the invention

[0005] The embodiments of the present application provide a traffic light passing calculation method and system based on the Internet of Things, which are used to solve the problem of low efficiency of traffic light passing calculation in the prior art.

[0006] In a first aspect, an embodiment of the present application provides a traffic light passing calculation method based on the Internet of Things, comprising:

[0007] Receive real-time data streams from a distributed IoT smart sensor network, integrate vehicle dynamic information and pedestrian flow statistics in the real-time data streams, perform spatiotemporal consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision spatiotemporal synchronization technology, and generate a dynamic panoramic view of traffic conditions;

[0008] Using a causal inference algorithm, inferring the causal relationship of historical data in the dynamic panoramic map of traffic conditions, identifying key influencing factors of the causal relationship, and constructing a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions, using spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulating the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics, and generating a future traffic state prediction model;

[0009] Generate a solution space based on the future traffic state prediction model, use an evolutionary strategy optimization algorithm to simulate the natural evolution process, eliminate local optimal traps in the solution space by iterating the natural evolution process, and search for the optimal traffic light switching strategy in the solution space, adopt a multi-objective optimization technology to consider mutually exclusive goals, introduce minimizing waiting time and maximizing traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the solution space, and generate a traffic light switching strategy evaluation result;

[0010] A feedback loop mechanism is introduced to collect the operation results of the future traffic state prediction model in real time, and an adaptive improvement report is automatically generated according to the potential improvement points in the traffic light switching strategy evaluation results. By comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted to optimize the traffic light passing signals in the dynamic panoramic view of the traffic conditions, and generate a traffic light passing calculation plan.

[0011] Optionally, the causal inference algorithm is used to infer the causal relationship of historical data in the dynamic panoramic view of traffic conditions, identify the key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and the spatiotemporal sequence modeling technology is used to capture the spatiotemporal characteristics of the direct and indirect impact paths, and the time-varying law of traffic flow in the dynamic panoramic view of traffic conditions is accurately simulated according to the spatiotemporal characteristics to generate a future traffic state prediction model, including:

[0012] Applying data smoothing technology to perform outlier detection and noise removal processing on vehicle dynamic information and pedestrian flow statistical data in the dynamic panoramic view of traffic conditions, applying feature selection algorithm to extract traffic flow state features from the outlier detection and noise removal processing results, and generating pre-processed traffic feature data;

[0013] Using a causal inference algorithm, inferring the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic characteristic data, identifying key influencing factors of the causal relationship, and constructing a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and generating a causal relationship graph of traffic events;

[0014] Adopting the spatiotemporal sequence modeling technology, taking the traffic event causal relationship map as the input variable to perform spatiotemporal sequence modeling, capturing the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulating the time-varying law of the traffic flow in the dynamic panoramic map of the traffic situation according to the spatiotemporal characteristics, and generating a spatiotemporal characteristic model;

[0015] The long-term structural impact in the causal relationship map of the traffic event and the real-time dynamic change details in the spatiotemporal characteristic model are extracted, and the long-term structural impact and the real-time dynamic change details are segmented using the sliding window technology. The segmented processing results are integrated into a hybrid prediction model to generate a future traffic status prediction model.

[0016] Optionally, the use of a causal inference algorithm to infer the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic feature data, identify key influencing factors of the causal relationship, and construct a causal graph to calculate direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and generate a causal relationship graph of traffic events, includes:

[0017] Introducing a cross-correlation function to perform correlation analysis on the pre-processed traffic characteristic data, so as to mine the potential influencing patterns and periodic changes of traffic flow stability in the pre-processed traffic characteristic data, and generate correlation analysis results;

[0018] Using a causal inference algorithm, inferring the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic characteristic data, identifying the key influencing factors of the causal relationship based on the correlation analysis results, and constructing a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and generating a traffic condition causal chain;

[0019] Applying the Shapley value decomposition method to quantify the influence of each causal path in the traffic condition causal chain on the traffic events in the traffic condition dynamic panoramic map, separating the long-term trend and seasonal trend of the traffic condition causal chain, and generating the causal chain time trend;

[0020] A graph layout algorithm is applied to define a basic graph structure according to the causal chain time trend, and the long-term trend and seasonal trend are mapped to the basic graph structure to generate a traffic event causal relationship graph.

[0021] Optionally, the spatiotemporal sequence modeling technology is used to perform spatiotemporal sequence modeling with the traffic event causal relationship map as an input variable, the spatiotemporal characteristics of the direct and indirect impact paths are captured, and the time-varying law of the traffic flow in the dynamic panoramic map of the traffic condition is accurately simulated according to the spatiotemporal characteristics to generate a spatiotemporal characteristic model, including:

[0022] The data normalization technology is applied to standardize the nodes and edges in the traffic incident causal relationship map, and the spatial attributes of the traffic incident causal relationship map are refined by introducing a geographic information system to generate standardized causal relationship vector data;

[0023] Adopting the spatiotemporal sequence modeling technology, taking the standardized causal relationship vector data and the traffic event causal relationship map as input variables to perform spatiotemporal sequence modeling, capturing the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulating the time-varying laws of the traffic flow in the dynamic panoramic map of the traffic situation according to the spatiotemporal characteristics, and generating dynamic pattern recognition results;

[0024] Applying pattern clustering technology to deeply analyze the dynamic pattern recognition results, identifying and classifying typical behavior patterns in the dynamic pattern recognition results, and using adaptive filtering algorithms to optimize the spatiotemporal sequence modeling process to generate optimized dynamic pattern recognition results;

[0025] Multi-scale analysis technology is introduced to capture the dynamic changes of traffic flow at different time scales in the optimized dynamic pattern recognition results, while considering the detailed spatial distribution information in the geographic information system, and an integrated learning framework is applied to integrate the dynamic changes of traffic flow and the detailed spatial distribution information to generate a spatiotemporal characteristic model.

[0026] Optionally, the solution space is generated based on the future traffic state prediction model, and an evolutionary strategy optimization algorithm is used to simulate the natural evolution process. By iterating the natural evolution process, local optimal traps in the solution space are eliminated, and the optimal traffic light switching strategy is searched in the solution space. Multi-objective optimization technology is used to consider mutually exclusive goals, and minimizing waiting time and maximizing traffic smoothness are introduced as constraints. The optimal traffic light switching strategy in the solution space is comprehensively evaluated to generate a traffic light switching strategy evaluation result, including:

[0027] Applying data mapping technology, converting the output result of the future traffic state prediction model into a multi-dimensional vector representation, introducing an initial population generation algorithm, randomly distributing a group of candidate solutions as the starting point of the solution space based on the multi-dimensional vector representation, and generating an initialized solution space;

[0028] An evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and individuals with high fitness are selected by iterating the natural evolution process. A mutation operator is used to introduce random disturbances into the evolutionary strategy optimization algorithm to eliminate local optimal traps in the initialization solution space, and an optimal traffic light switching strategy is searched in the initialization solution space to generate an evolutionary optimization solution set.

[0029] Multi-objective optimization technology is used to consider mutually exclusive objectives, the evolutionary optimization solution set is introduced as the initial test condition, and minimizing waiting time and maximizing traffic flow are additionally introduced as constraint conditions, and the optimal traffic light switching strategy in the initial solution space is comprehensively evaluated to generate a multi-objective optimization solution;

[0030] Taking the evolutionary optimization solution set and the multi-objective optimization solution as decision criteria, a decision matrix analysis method is introduced to construct a scoring system, further verify the effectiveness of the optimal traffic light switching strategy, and generate a traffic light switching strategy evaluation result.

[0031] Optionally, the use of an evolutionary strategy optimization algorithm to simulate a natural evolution process, screening individuals with high fitness by iterating the natural evolution process, introducing random disturbances in the evolutionary strategy optimization algorithm using a mutation operator to eliminate local optimal traps in the initialization solution space, and searching for an optimal traffic light switching strategy in the initialization solution space to generate an evolutionary optimization solution set includes:

[0032] Introducing a fitness evaluation function, calculating the fitness value of each candidate initialization solution in the initialization solution space under the current traffic conditions, increasing the genetic diversity of the initialization solution space through diversity enhancement technology, and generating an optimized initialization solution space;

[0033] An evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and high-fitness individuals are selected by iterating the natural evolution process. The genetic information between the high-fitness individuals is exchanged by a crossover operator, and a random disturbance is introduced into the evolutionary strategy optimization algorithm by a mutation operator to eliminate the local optimal trap in the initialization solution space, and the optimal traffic light switching strategy is searched in the optimized initialization solution space to generate a natural evolution intermediate solution set;

[0034] Applying dynamic fitness adjustment technology, adjusting the fitness evaluation standard for high fitness individuals in real time during the natural evolution process, and using a local search algorithm to perform local fine adjustment on the natural evolution intermediate solution set to generate an optimized natural evolution solution set;

[0035] A simulated annealing algorithm is introduced to set the initial temperature and cooling rate to simulate the thermal motion process in the physical system. During the initial temperature reduction process, the global search and local optimization processes of the local search algorithm are balanced to generate an evolutionary optimization solution set.

[0036] Optionally, the feedback loop mechanism is introduced to collect the operation results of the future traffic state prediction model in real time, automatically form an adaptive improvement report according to the potential improvement points in the traffic light switching strategy evaluation results, and cyclically adjust the spatiotemporal sequence parameters in the future traffic state prediction model by comparing the operation results with the adaptive improvement report, optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation scheme, including:

[0037] Introducing a feedback loop mechanism to collect the operation results of the future traffic state prediction model in real time, and applying data mining technology to integrate heterogeneous data in the operation results to generate a heterogeneous data set;

[0038] According to the potential improvement points in the traffic light switching strategy evaluation results, automatically learning the potential improvement points using a machine learning algorithm, obtaining difference patterns, and generating an adaptive improvement report;

[0039] By comparing the operation result with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are adjusted cyclically, the traffic light passing signal in the dynamic panoramic view of the traffic condition is optimized, and a passing signal optimization plan is generated;

[0040] A path planning algorithm is introduced to perform signal timing on key intersections in the through signal optimization scheme, and the traffic light switching time in the through signal optimization scheme is adjusted according to the signal timing result to generate a traffic light through calculation scheme.

[0041] In a second aspect, an embodiment of the present application provides a traffic light calculation system based on the Internet of Things, including:

[0042] A receiving module is used to receive a real-time data stream from a distributed IoT smart sensor network, integrate vehicle dynamic information and pedestrian flow statistics in the real-time data stream, perform spatiotemporal consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision spatiotemporal synchronization technology, and generate a dynamic panoramic view of traffic conditions;

[0043] A construction module is used to use a causal inference algorithm to infer the causal relationship of historical data in the dynamic panoramic map of traffic conditions, identify key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions, use spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics to generate a future traffic state prediction model;

[0044] A search module is used to generate a solution space based on the future traffic state prediction model, use an evolutionary strategy optimization algorithm to simulate a natural evolution process, eliminate local optimal traps in the solution space by iterating the natural evolution process, and search for an optimal traffic light switching strategy in the solution space, adopt a multi-objective optimization technique to consider mutually exclusive goals, introduce minimizing waiting time and maximizing traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the solution space, and generate a traffic light switching strategy evaluation result;

[0045] The adjustment module is used to introduce a feedback loop mechanism, collect the operation results of the future traffic state prediction model in real time, automatically form an adaptive improvement report based on the potential improvement points in the traffic light switching strategy evaluation results, and cyclically adjust the spatiotemporal sequence parameters in the future traffic state prediction model by comparing the operation results with the adaptive improvement report, optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation plan.

[0046] In a third aspect, an embodiment of the present application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a traffic light passing calculation method based on the Internet of Things as described in the first aspect.

[0047] In a fourth aspect, an embodiment of the present application provides a computer storage medium storing a computer program. When the computer program is executed by a computer, it implements a traffic light passing calculation method based on the Internet of Things as described in the first aspect.

[0048] In an embodiment of the present application, a real-time data stream is received from a distributed IoT intelligent sensor network, vehicle dynamic information and pedestrian flow statistics in the real-time data stream are integrated, and the vehicle dynamic information and pedestrian flow statistics are processed for spatiotemporal consistency through high-precision spatiotemporal synchronization technology to generate a dynamic panoramic view of traffic conditions; a causal inference algorithm is used to infer the causal relationship of historical data in the dynamic panoramic view of traffic conditions, key influencing factors of the causal relationship are identified, and a causal graph is constructed to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions; spatiotemporal sequence modeling technology is used to capture the spatiotemporal characteristics of the direct and indirect impact paths, and the time-varying laws of traffic flow in the dynamic panoramic view of traffic conditions are accurately simulated according to the spatiotemporal characteristics to generate a future traffic status prediction model; a solution space is generated based on the future traffic status prediction model, and evolution is used to generate a solution space based on the future traffic status prediction model. The strategy optimization algorithm simulates the natural evolution process, removes the local optimal traps in the solution space by iterating the natural evolution process, and searches for the optimal traffic light switching strategy in the solution space. The multi-objective optimization technology is used to consider mutually exclusive goals, and the constraints of minimizing waiting time and maximizing traffic fluency are introduced to comprehensively evaluate the optimal traffic light switching strategy in the solution space and generate the evaluation results of the traffic light switching strategy. The feedback loop mechanism is introduced to collect the operation results of the future traffic state prediction model in real time, and automatically form an adaptive improvement report based on the potential improvement points in the evaluation results of the traffic light switching strategy. By comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted to optimize the traffic light passing signal in the dynamic panoramic view of the traffic situation and generate a traffic light passing calculation scheme. By integrating vehicle dynamic information and pedestrian flow statistics in real time, an accurate dynamic panoramic view of the traffic situation can be generated, which can more effectively manage and optimize traffic flow and reduce congestion and accidents. The causal relationship between traffic events and their impact paths can be accurately identified by using causal inference algorithms and spatiotemporal sequence modeling technology, providing a reliable prediction model for future traffic conditions. This will help formulate more scientific and reasonable traffic light switching strategies, further improving traffic flow and pedestrian safety.

[0049] Furthermore, data smoothing technology and feature selection algorithms are applied to detect outliers, remove noise, and extract features of traffic flow status, which improves the quality and accuracy of input data and enhances the reliability of subsequent analysis. By constructing a causal relationship map of traffic events and combining it with spatiotemporal series modeling technology, it can not only capture the spatiotemporal characteristics of direct and indirect impact paths, but also effectively simulate the time-varying laws of traffic flow, making the prediction of future traffic status more accurate. The sliding window technology is used to segment the long-term structural impacts and the details of immediate dynamic changes, and integrate them into the hybrid prediction model, which enhances the adaptability and flexibility of the system under different environmental conditions.

[0050] Furthermore, the evolutionary strategy optimization algorithm is used to eliminate local optimal traps and search for the optimal traffic light switching strategy in a wide solution space to ensure that the solution found is globally optimal or close to the global optimal solution. Minimizing waiting time and maximizing traffic smoothness are introduced as constraints, and multi-objective optimization technology is used to comprehensively evaluate candidate solutions, achieving a reasonable trade-off between multiple mutually exclusive objectives and improving the performance of the overall traffic system. A scoring system is constructed through the decision matrix analysis method, which not only verifies the effectiveness of the optimal traffic light switching strategy, but also increases the transparency and credibility of the decision-making process, facilitating subsequent monitoring and adjustment.

[0051] These and other aspects of the present application will become more clearly understood in the description of the following embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0052] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following is a brief introduction to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0053] Figure 1 A flow chart of a traffic light passing calculation method based on the Internet of Things provided in an embodiment of the present application;

[0054] Figure 2 A schematic diagram of a traffic light calculation system based on the Internet of Things provided in an embodiment of the present application;

[0055] Figure 3 A schematic diagram of the structure of a computing device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the drawings in the embodiments of the present application.

[0057] In some of the processes described in the specification and claims of this application and the above-mentioned figures, multiple operations that appear in a specific order are included, but it should be clearly understood that these operations may not be executed in the order in which they appear in this article or executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish between different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions of "first", "second", etc. in this article are used to distinguish different messages, devices, modules, etc., do not represent the order of precedence, and do not limit the "first" and "second" to be different types.

[0058] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions 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 those skilled in the art without creative work are within the scope of protection of this application.

[0059] Figure 1 A flow chart of a traffic light calculation method based on the Internet of Things is provided for an embodiment of the present application, such as Figure 1 As shown, the method includes:

[0060] 101. Receive real-time data streams from a distributed IoT smart sensor network, integrate vehicle dynamic information and pedestrian flow statistics in the real-time data streams, perform spatiotemporal consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision spatiotemporal synchronization technology, and generate a dynamic panoramic view of traffic conditions;

[0061] In this step, the distributed IoT smart sensor network consists of a series of smart devices that can monitor and transmit vehicle dynamic information and pedestrian flow statistics in real time.

[0062] Vehicle dynamic information and pedestrian flow statistics contain all data about the movement status of vehicles and the changes in the number of pedestrians at different times and locations. These data are used to evaluate current traffic conditions and provide a basis for predicting future trends.

[0063] High-precision spatiotemporal synchronization technology is a technology that ensures that data from different sensors are consistent in time and space dimensions. It enables data from different sources to be seamlessly integrated to generate a panoramic view that accurately reflects actual traffic conditions.

[0064] The dynamic panoramic map of traffic conditions is a map generated based on integrated vehicle dynamic information and pedestrian flow statistics. It shows the real-time status of traffic flow in the city and its changing trends.

[0065] In the embodiments of the present application, first, real-time traffic data is collected through a distributed IoT intelligent sensor network; second, the collected data is preprocessed, including removing noise and outliers; third, high-precision spatiotemporal synchronization technology is used to process data from different sources for spatiotemporal consistency; finally, the integrated data is used to generate a dynamic panoramic view of the traffic conditions.

[0066] Suppose in a smart city project, first, the dynamic information of vehicles and pedestrians is obtained in real time through the intelligent cameras and sensors deployed at various intersections; secondly, the collected data is cleaned and preprocessed to remove noise and outliers; thirdly, high-precision spatiotemporal synchronization technology is applied to ensure the temporal and spatial consistency of data from different sensors; finally, a dynamic panoramic view of the city’s traffic conditions is successfully generated to help traffic management departments monitor and manage traffic flows more effectively.

[0067] 102. Use a causal inference algorithm to infer the causal relationship of historical data in the dynamic panoramic map of traffic conditions, identify key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions, use spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics to generate a future traffic state prediction model;

[0068] In this step, the causal inference algorithm is a statistical method used to identify causal relationships in data and understand the influence mechanism between variables. The algorithm identifies which factors are the main causes of a phenomenon by analyzing the dependencies between variables.

[0069] Causal graphs show graphical representations of direct and indirect influence paths between variables, which helps to understand and predict the behavior of complex systems. Causal graphs represent variables and their relationships through nodes and edges, where nodes represent variables and edges represent causal relationships.

[0070] Spatiotemporal series modeling technology is a method of time series analysis, which is particularly suitable for considering the changing patterns of spatial factors and can capture the spatiotemporal characteristics of traffic events. This method combines data in time and space dimensions to simulate the changing laws of traffic flow.

[0071] The future traffic status prediction model predicts future traffic conditions based on historical data and current conditions. The model provides prediction results for future traffic conditions by comprehensively considering historical data and current real-time data.

[0072] In an embodiment of the present application, first, a causal inference algorithm is used to analyze historical data in a dynamic panoramic view of traffic conditions to identify key influencing factors; secondly, a causal graph is constructed to depict the direct and indirect impact paths between traffic events; thirdly, spatiotemporal sequence modeling technology is used to capture the spatiotemporal characteristics of these paths; finally, a future traffic status prediction model is generated based on the above analysis results.

[0073] For example, continuing with the above example, first, through the analysis of traffic data over the past year, it was found that the main reason for traffic congestion around schools in a specific time period was that parents were picking up their children from school; secondly, a causal map was further constructed to depict the key factors that lead to congestion and their interrelationships; thirdly, the time and scope of similar events in the future were predicted using spatiotemporal sequence modeling technology; finally, the traffic light strategy in the relevant area was optimized based on these analysis results. First, historical data was analyzed to identify the main influencing factors; secondly, a causal map was constructed to clarify the causal relationship between the various factors; thirdly, the spatiotemporal sequence modeling technology was used to simulate the changing patterns of traffic flow; finally, a future traffic state prediction model was generated to provide a basis for subsequent optimization.

[0074] 103. Generate a solution space based on the future traffic state prediction model, use an evolutionary strategy optimization algorithm to simulate the natural evolution process, eliminate local optimal traps in the solution space by iterating the natural evolution process, and search for the optimal traffic light switching strategy in the solution space, adopt a multi-objective optimization technology to consider mutually exclusive goals, introduce minimizing waiting time and maximizing traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the solution space, and generate a traffic light switching strategy evaluation result;

[0075] In this step, the evolutionary strategy optimization algorithm is an algorithm that simulates the natural selection process. It searches for the optimal solution through iteration and is often used to solve complex multi-objective optimization problems. The algorithm simulates the mechanism of biological evolution, gradually improves the individuals in the population, and finds the optimal solution.

[0076] A local optimal trap refers to a solution reached during the search process that appears to be the best but is not actually the global optimal solution. The local optimal trap can cause the algorithm to converge prematurely and fail to find the global optimal solution.

[0077] Multi-objective optimization technology is the process of finding a balance point by considering multiple conflicting objectives. Multi-objective optimization technology finds a set of Pareto optimal solutions by weighing different objectives.

[0078] Traffic Light Switching Strategy Evaluation Results Evaluate the effects of different traffic light switching strategies and determine the most appropriate implementation plan.

[0079] In the embodiments of the present application, first, a solution space is generated based on a future traffic state prediction model; second, an evolutionary strategy optimization algorithm is used to explore the solution space to avoid falling into a local optimal trap; third, a multi-objective optimization technique is used to evaluate different traffic light switching strategies; finally, the optimal strategy is determined based on the evaluation results.

[0080] For example, continuing the above case, firstly, a solution space is generated based on the future traffic state prediction model, which includes all possible traffic light switching strategies; secondly, the evolutionary strategy optimization algorithm is used to simulate the effects of different traffic light switching strategies to avoid falling into the local optimal trap; thirdly, multi-objective optimization technology is used to consider the two goals of minimizing waiting time and maximizing traffic flow to evaluate the effects of different strategies; finally, the optimal traffic light switching strategy is determined based on the evaluation results, which not only reduces traffic congestion during the students' commuting time, but also improves the overall traffic flow. Firstly, a solution space is generated, which includes all candidate strategies; secondly, the evolutionary strategy optimization algorithm is applied to gradually improve the strategy; thirdly, the multi-objective optimization technology is used to evaluate the comprehensive performance of each strategy; finally, the optimal strategy is determined to ensure the global optimality.

[0081] 104. Introduce a feedback loop mechanism to collect the operation results of the future traffic state prediction model in real time, automatically generate an adaptive improvement report based on the potential improvement points in the traffic light switching strategy evaluation results, and cyclically adjust the spatiotemporal sequence parameters in the future traffic state prediction model by comparing the operation results with the adaptive improvement report, optimize the traffic light passing signals in the dynamic panoramic view of the traffic conditions, and generate a traffic light passing calculation plan.

[0082] The feedback loop mechanism is a continuous improvement process that adjusts and optimizes system performance by continuously collecting the actual results of system operation and comparing them with the expected goals.

[0083] The adaptive improvement report is a document generated based on actual operating results and potential improvement points to guide further optimization of the system. The report provides detailed improvement suggestions to help the system better meet requirements.

[0084] The spatiotemporal sequence parameters are the time- and space-related parameters that affect the prediction accuracy of the dynamic panorama of traffic conditions. These parameters determine the sensitivity of the model to temporal and spatial changes.

[0085] Traffic lights are designed to improve traffic efficiency by calculating the optimal traffic light switching plan. The plan takes various factors into consideration to provide the best traffic light switching strategy.

[0086] In the embodiment of the present application, first, a feedback loop mechanism is established to collect the operating results of the future traffic state prediction model in real time; second, an adaptive improvement report is formed based on the potential improvement points in the traffic light switching strategy evaluation results; third, the operating results are compared with the improvement report, and the spatiotemporal sequence parameters are adjusted; finally, an optimized traffic light calculation plan is generated.

[0087] For example, continuing with the above example, first introduce a feedback loop mechanism to regularly evaluate the actual effect of the new traffic light strategy and collect the operating results of the future traffic status prediction model in real time; secondly, form an adaptive improvement report based on the potential improvement points in the traffic light switching strategy evaluation results, and put forward specific improvement suggestions; thirdly, compare the actual operating results with the improvement report, adjust the spatiotemporal series parameters, improve the model prediction accuracy, and ensure that the system can better respond to changes; finally, generate an optimized traffic light calculation plan, which not only improves the management efficiency of traffic around the school, but also promotes the continuous optimization and development of the entire urban transportation system.

[0088] In summary, steps 101 to 104 cover the complete process from data collection, analysis, optimization to feedback adjustment, aiming to provide a comprehensive and dynamic traffic management system to meet the needs of improving urban traffic efficiency and response speed.

[0089] In order to further improve the accuracy and predictive ability of the dynamic panoramic view of traffic conditions, in some embodiments, the use of a causal inference algorithm in step 102 to infer the causal relationship of historical data includes: applying data smoothing technology to perform outlier detection and noise removal processing on vehicle dynamic information and pedestrian flow statistics in the dynamic panoramic view of traffic conditions, applying a feature selection algorithm to extract traffic flow state features from the results of the outlier detection and noise removal processing to generate pre-processed traffic feature data; using a causal inference algorithm to infer the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic feature data, identifying the key influencing factors of the causal relationship, and constructing a causal graph to calculate the dynamic view of traffic conditions. The direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions are obtained to generate a causal relationship map of traffic events; the spatiotemporal sequence modeling technology is used to perform spatiotemporal sequence modeling with the traffic event causal relationship map as an input variable to capture the spatiotemporal characteristics of the direct and indirect impact paths, and the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions are accurately simulated according to the spatiotemporal characteristics to generate a spatiotemporal characteristic model; the long-term structural impact in the causal relationship map of traffic events and the immediate dynamic change details in the spatiotemporal characteristic model are extracted, the long-term structural impact and the immediate dynamic change details are segmented by using the sliding window technology, the segmented processing results are integrated into a hybrid prediction model to generate a future traffic status prediction model.

[0090] In this embodiment, the data smoothing technology is a data processing method for removing outliers and noise, and the accuracy of subsequent analysis is ensured by smoothing the original data.

[0091] Feature selection algorithm is a technique to extract the most representative features from a large amount of data. By evaluating the importance of each feature, the features with the greatest impact on the target variable are selected to generate preprocessed traffic feature data in order to better understand the potential patterns in the data.

[0092] The preprocessed traffic characteristic data is a high-quality data set processed by data smoothing technology and feature selection algorithm, which contains key features in vehicle dynamic information and pedestrian flow statistics data, aiming to remove noise and outliers and retain the most representative features for subsequent analysis.

[0093] The traffic incident causal relationship map is a graphical representation of the key influencing factors and their interaction paths in historical data identified by causal inference algorithms. It is used to intuitively display the direct and indirect impacts between traffic events. The nodes in the map represent variables and the edges represent causal relationships.

[0094] The spatiotemporal characteristic model uses spatiotemporal sequence modeling technology to capture the spatiotemporal variation characteristics of the direct and indirect influencing paths in the causal relationship map of traffic events, generate a model that can simulate the time-varying laws of traffic flow, and help predict future traffic conditions.

[0095] The future traffic status prediction model combines long-term structural impacts with immediate dynamic change details to accurately predict future traffic conditions through a hybrid prediction model. The model integrates historical data, current conditions and predicted trends to provide a comprehensive prediction of future traffic conditions.

[0096] In the embodiment of the present application, firstly, data smoothing technology is applied to perform outlier detection and noise removal processing on vehicle dynamic information and pedestrian flow statistical data; secondly, a feature selection algorithm is used to extract traffic flow state characteristics from the above processing results to generate preprocessed traffic characteristic data; thirdly, a causal inference algorithm is used to infer the causal relationship in these preprocessed data, identify key influencing factors, and construct a causal relationship map of traffic events; finally, the spatiotemporal sequence modeling technology is used to capture the spatiotemporal characteristics of traffic events based on the causal relationship map, generate a spatiotemporal characteristic model, and generate a future traffic state prediction model based on this.

[0097] Here is a specific example:

[0098] For example, suppose that in a smart city traffic management system, first, the system obtains real-time vehicle dynamic information (such as speed, location) and pedestrian flow statistics (such as number of people, flow direction) from the distributed IoT smart sensor network, and applies data smoothing technology to remove noise and outliers; secondly, the system uses feature selection algorithms to extract key features reflecting the traffic flow status from the purified data and generate pre-processed traffic feature data; thirdly, the system applies causal inference algorithms to identify the causal relationship between these features, construct a causal relationship map of traffic events, and clarify the mutual influence between various factors; then, the system uses spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of traffic events based on the causal relationship map, generate a spatiotemporal characteristic model, and simulate the changing laws of traffic flow; finally, the system integrates long-term structural impacts with immediate dynamic change details into a hybrid prediction model to generate a future traffic status prediction model and provide a comprehensive prediction of future traffic conditions.

[0099] In order to further improve the accuracy of the dynamic panoramic view of traffic conditions and the precision of causal relationship identification, in some embodiments, the use of a causal inference algorithm in step 102 to infer the causal relationship between historical data and pre-processed traffic feature data includes: introducing a cross-correlation function to perform a correlation analysis on the pre-processed traffic feature data to mine the potential influencing patterns and periodic changes of traffic flow stability in the pre-processed traffic feature data, and generating a correlation analysis result; using a causal inference algorithm to infer the causal relationship between historical data and the pre-processed traffic feature data in the dynamic panoramic view of traffic conditions, identifying the key influencing factors of the causal relationship based on the correlation analysis result, and constructing a causal graph to calculate the direct and indirect influencing paths between traffic events in the dynamic panoramic view of traffic conditions, and generating a traffic condition causal chain; applying the Shapley value decomposition method to quantify the degree of influence of each causal path in the traffic condition causal chain on the traffic events in the dynamic panoramic view of traffic conditions, separating the long-term trend and seasonal trend of the traffic condition causal chain, and generating a causal chain time trend; applying a graphic layout algorithm to define a basic graph structure according to the causal chain time trend, mapping the long-term trend and seasonal trend to the basic graph structure, and generating a traffic event causal relationship graph.

[0100] In this embodiment, the cross-correlation function is a statistical method for analyzing the correlation between time series. By calculating the correlation coefficients at different time lags, the potential patterns and periodic changes in traffic flow stability can be discovered.

[0101] The results of correlation analysis are based on the results of the cross-correlation function, which reveal the hidden influencing factors and their interaction patterns in the traffic characteristic data, and provide a basis for the subsequent causal relationship mining.

[0102] The traffic condition causal chain shows a graphical representation of the direct or indirect impact paths between traffic events, which helps to understand and predict future traffic conditions.

[0103] The Shapley value decomposition method is a method to quantify the contribution of each factor to the overall result. It can separate long-term trends and seasonal trends, thereby better understanding the development of traffic incidents.

[0104] The causal chain time trend is a time trend generated based on the Shapley value decomposition method, which describes the changes in the impact of each causal path on traffic events.

[0105] The graph layout algorithm is used to define the basic structure of the causal relationship graph, map the long-term and short-term trends into the graph according to the time trend of the causal chain, and display the causal relationship of traffic events in a visual way.

[0106] The traffic incident causal relationship map is a visualization tool that helps decision makers understand the causal relationship between traffic incidents and their changing trends over time.

[0107] In the embodiment of the present application, first, the correlation analysis of the pre-processed traffic characteristic data is performed using the cross-correlation function to identify the influencing pattern of traffic flow stability; second, the causal inference algorithm is applied to identify the causal relationship between historical data and the pre-processed traffic characteristic data, and a causal graph is constructed; third, the Shapley value decomposition method is used to quantify the degree of influence of each causal path, and separate the long-term and seasonal trends; finally, the graph layout algorithm is used to define the graph structure according to the time trend of the causal chain, and generate a causal relationship graph of traffic events.

[0108] Here is a specific example:

[0109] In a case study aimed at optimizing an urban public transportation dispatching system, first, the system collects and integrates real-time traffic data from multiple sources, such as bus GPS data, passenger boarding and alighting records, and weather information, to form a preprocessed traffic feature data set. Second, these data sets are analyzed through cross-correlation functions to find out the stability of traffic flow and the periodic change pattern affected by weather. Third, a causal inference algorithm is used to identify the specific impact of weather conditions (such as rainfall) on bus delays, and a causal graph is constructed to show how weather changes lead to worsening traffic conditions. Then, the Shapley value decomposition method is used to evaluate the contribution of bus delays under each weather condition, clearly distinguishing between long-term effects and short-term fluctuations. Finally, a graphical layout algorithm is used to design an intuitive causal relationship graph of traffic events to guide traffic management departments to formulate more scientific and reasonable response strategies, such as increasing capacity in advance or adjusting route planning in severe weather.

[0110] In order to further improve the simulation accuracy of the spatiotemporal characteristics of the dynamic panoramic map of traffic conditions, in some embodiments, the spatiotemporal sequence modeling technology described in step 102 is used to perform spatiotemporal sequence modeling with the traffic event causal relationship map as an input variable to capture the spatiotemporal characteristics of the direct and indirect influencing paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics to generate a spatiotemporal characteristic model, including: applying data normalization technology to standardize the nodes and edges in the traffic event causal relationship map, and introducing a geographic information system to refine the spatial attributes of the traffic event causal relationship map to generate standardized causal relationship vector data; using the spatiotemporal sequence modeling technology with the standardized causal relationship vector data and the traffic event causal relationship map as input variables; The input variables are used to carry out spatiotemporal sequence modeling to capture the spatiotemporal characteristics of the direct and indirect influencing paths, and the time-varying laws of the traffic flow in the dynamic panoramic map of the traffic conditions are accurately simulated according to the spatiotemporal characteristics to generate dynamic pattern recognition results; the pattern clustering technology is used to deeply analyze the dynamic pattern recognition results to identify and classify the typical behavior patterns in the dynamic pattern recognition results, and the adaptive filtering algorithm is used to optimize the spatiotemporal sequence modeling process to generate optimized dynamic pattern recognition results; the multi-scale analysis technology is introduced to capture the dynamic changes of traffic flow at different time scales in the optimized dynamic pattern recognition results, and the detailed spatial distribution information in the geographic information system is considered at the same time, and the integrated learning framework is used to integrate the dynamic changes of traffic flow and the detailed spatial distribution information to generate a spatiotemporal characteristic model.

[0111] In this embodiment, data normalization technology is a method for standardizing data by uniformly processing nodes and edges in a graph to ensure consistency of data from different sources.

[0112] Geographic Information System is a technical system used to manage and analyze geographic spatial data. By refining the spatial attributes of the cause-and-effect relationship map of traffic events, GIS can provide more accurate spatial information and enhance the spatial resolution of the model.

[0113] Standardized causal relationship vector data is a causal relationship graph dataset processed by data normalization technology and GIS. It contains standardized representations of nodes and edges and their detailed spatial attribute information, providing high-quality input for spatiotemporal series modeling.

[0114] Spatiotemporal series modeling technology is a time series analysis method that combines time and space dimensions. It can capture the spatiotemporal variation patterns between traffic events and simulate the time-varying laws of traffic flow.

[0115] The dynamic pattern recognition results are generated based on the spatiotemporal sequence modeling technology, which describes the spatiotemporal characteristics and changing laws of the direct and indirect impact paths between traffic events.

[0116] Pattern clustering technology is a method used to classify and identify typical behavior patterns. It extracts representative patterns and optimizes the spatiotemporal sequence modeling process by deeply analyzing the dynamic pattern recognition results.

[0117] Adaptive filtering algorithm is an algorithm used to optimize the signal processing process. It improves the accuracy and stability of spatiotemporal sequence modeling by dynamically adjusting filter parameters.

[0118] Multiscale analysis technology is a method used to capture the changing characteristics of data at different time scales. It can take into account both short-term fluctuations and long-term trends and provide a more comprehensive analysis of spatiotemporal characteristics.

[0119] The ensemble learning framework is a technical framework that combines multiple machine learning models to generate a more accurate and robust spatiotemporal characteristic model by integrating the advantages of different models.

[0120] The spatiotemporal characteristic model comprehensively considers multi-scale analysis and detailed spatial distribution information in the geographic information system. The generated spatiotemporal characteristic model can accurately simulate the changing laws of traffic flow and provide a scientific basis for traffic management.

[0121] In the embodiment of the present application, firstly, data normalization technology is applied to standardize the nodes and edges in the causal relationship map of traffic events, and the spatial attributes are refined by introducing a geographic information system to generate standardized causal relationship vector data; secondly, spatiotemporal sequence modeling technology is adopted to perform spatiotemporal sequence modeling with standardized causal relationship vector data and the causal relationship map of traffic events as input variables to capture the spatiotemporal characteristics of direct and indirect influencing paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions according to the spatiotemporal characteristics to generate dynamic pattern recognition results; thirdly, pattern clustering technology is applied to deeply analyze the dynamic pattern recognition results, identify and classify typical behavior patterns, and use an adaptive filtering algorithm to optimize the spatiotemporal sequence modeling process to generate optimized dynamic pattern recognition results; finally, multi-scale analysis technology is introduced to capture the dynamic changes of traffic flow at different time scales in the optimized dynamic pattern recognition results, while considering the detailed spatial distribution information in the geographic information system, and applying an integrated learning framework to integrate the dynamic changes of traffic flow and the detailed spatial distribution information to generate a spatiotemporal characteristic model.

[0122] Here is a specific example:

[0123] Suppose that in an intelligent traffic management system in a city's central business district, it is necessary to optimize traffic flow and reduce congestion during peak hours. First, the system obtains data on vehicle speed, pedestrian flow, and surrounding building activities from multiple sensor nodes to generate pre-processed traffic feature data; second, the nodes and edges in the traffic event causal relationship map are standardized through data normalization technology, and their spatial attributes are refined by introducing a geographic information system to generate standardized causal relationship vector data; third, the spatiotemporal sequence modeling technology is used to perform spatiotemporal sequence modeling with standardized causal relationship vector data and traffic event causal relationship maps as input variables to capture the spatiotemporal characteristics of direct and indirect influencing paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics to generate dynamic pattern recognition results; then Then, the pattern clustering technology is used to deeply analyze the dynamic pattern recognition results, identify and classify typical behavior patterns, and use the adaptive filtering algorithm to optimize the spatiotemporal sequence modeling process to generate optimized dynamic pattern recognition results; finally, multi-scale analysis technology is introduced to capture the dynamic changes of traffic flow at different time scales in the optimized dynamic pattern recognition results, while considering the detailed spatial distribution information in the geographic information system, and using the integrated learning framework to integrate the dynamic changes of traffic flow with the detailed spatial distribution information to generate a spatiotemporal characteristic model to help traffic management departments formulate effective traffic diversion strategies, such as adding temporary lanes or adjusting traffic light times during peak hours, thereby improving overall traffic efficiency and citizens' travel experience.

[0124] In order to further improve the global optimality and multi-objective optimization effect of the traffic light switching strategy, in some embodiments, the solution space is generated based on the future traffic state prediction model in step 103, and the evolutionary strategy optimization algorithm is used to simulate the natural evolution process. By iterating the natural evolution process, the local optimal traps in the solution space are eliminated, and the optimal traffic light switching strategy is searched in the solution space. The multi-objective optimization technology is used to consider mutually exclusive goals, and the minimization of waiting time and maximization of traffic smoothness are introduced as constraints. The optimal traffic light switching strategy in the solution space is comprehensively evaluated to generate a traffic light switching strategy evaluation result, including: applying data mapping technology to convert the output result of the future traffic state prediction model into a multi-dimensional vector representation, introducing an initial population generation algorithm, and randomly distributing a group of candidate solutions as the starting point of the solution space based on the multi-dimensional vector representation to generate an initialized solution space ; Use the evolutionary strategy optimization algorithm to simulate the natural evolution process, select high fitness individuals by iterating the natural evolution process, use the mutation operator to introduce random disturbances in the evolutionary strategy optimization algorithm to eliminate local optimal traps in the initialization solution space, and search for the optimal traffic light switching strategy in the initialization solution space to generate an evolutionary optimization solution set; adopt multi-objective optimization technology to consider mutually exclusive goals, introduce the evolutionary optimization solution set as the initial test condition, and additionally introduce minimizing waiting time and maximizing traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the initialization solution space, and generate a multi-objective optimization solution; use the evolutionary optimization solution set and the multi-objective optimization solution as decision criteria, introduce the decision matrix analysis method to construct a scoring system, further verify the effectiveness of the optimal traffic light switching strategy, and generate a traffic light switching strategy evaluation result.

[0125] In this embodiment, data mapping technology is a method of converting a complex data structure into a multi-dimensional vector representation that is easy to process. Through this technology, the output results of the future traffic state prediction model can be converted into a form suitable for processing by the optimization algorithm.

[0126] The initial population generation algorithm is a method for generating an initial set of candidate solutions. Based on multidimensional vector representation, a set of candidate solutions is randomly distributed as the starting point of the solution space to ensure the diversity of the solution space.

[0127] The mutation operator is a random perturbation mechanism introduced in the process of evolutionary strategy optimization. It explores new areas of the solution space by changing the eigenvalues ​​of certain individuals, thereby increasing the possibility of finding the global optimal solution.

[0128] Decision matrix analysis is a method used to construct a scoring system that determines the effectiveness of the optimal solution by comprehensively evaluating multiple indicators.

[0129] The traffic light switching strategy evaluation results are the final evaluation reports generated based on the evolutionary optimization solution set and the multi-objective optimization solution, providing detailed performance indicators and improvement suggestions to help traffic management departments develop more effective strategies.

[0130] In the embodiment of the present application, firstly, the output result of the future traffic state prediction model is converted into a multi-dimensional vector representation by applying data mapping technology; secondly, an initial population generation algorithm is introduced, and a group of candidate solutions are randomly distributed based on the multi-dimensional vector representation as the starting point of the solution space to generate an initialization solution space; thirdly, an evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and high fitness individuals are iteratively screened, random disturbances are introduced using mutation operators, local optimal traps in the initialization solution space are eliminated, and the optimal traffic light switching strategy is searched in the solution space to generate an evolutionary optimization solution set; then, a multi-objective optimization technology is used to consider mutually exclusive goals, and an evolutionary optimization solution set is introduced as a preliminary test condition, and minimizing waiting time and maximizing traffic smoothness are additionally introduced as constraints, and the optimal traffic light switching strategy in the initialization solution space is comprehensively evaluated to generate a multi-objective optimization solution; finally, the evolutionary optimization solution set and the multi-objective optimization solution are used as decision criteria, and a decision matrix analysis method is introduced to construct a scoring system to further verify the effectiveness of the optimal traffic light switching strategy and generate a traffic light switching strategy evaluation result.

[0131] Here is a specific example:

[0132] Suppose that in a traffic management system of a large smart city, it is necessary to optimize the traffic light switching strategy during peak hours to reduce congestion and improve traffic flow. First, the system obtains real-time vehicle speed, pedestrian flow and surrounding building activity data from multiple sensor nodes, generates the output results of the future traffic state prediction model, and converts it into a multi-dimensional vector representation using data mapping technology; secondly, an initial population generation algorithm is introduced to randomly distribute a set of candidate solutions as the starting point of the solution space based on the multi-dimensional vector representation to generate an initialization solution space; thirdly, an evolutionary strategy optimization algorithm is used to simulate the natural evolution process, iteratively screens high-fitness individuals, uses mutation operators to introduce random disturbances, eliminates local optimal traps in the initialization solution space, and searches for the optimal traffic light switching strategy in the solution space to generate an evolutionary optimization solution set; then, Multi-objective optimization technology is used to consider mutually exclusive objectives, and an evolutionary optimization solution set is introduced as the initial test condition. Minimizing waiting time and maximizing traffic smoothness are additionally introduced as constraints. The optimal traffic light switching strategy in the initial solution space is comprehensively evaluated to generate a multi-objective optimization solution. Finally, the evolutionary optimization solution set and the multi-objective optimization solution are used as decision criteria, and the decision matrix analysis method is introduced to construct a scoring system to further verify the effectiveness of the optimal traffic light switching strategy, generate traffic light switching strategy evaluation results, and help traffic management departments adjust signal light time during peak hours, add temporary lanes or adjust route planning, thereby improving overall traffic efficiency and citizens' travel experience.

[0133] In order to further improve the global search capability of the traffic light switching strategy and avoid local optimal traps, in some embodiments, the evolutionary strategy optimization algorithm described in step 103 is used to simulate the natural evolution process, and high fitness individuals are screened by iterating the natural evolution process. A mutation operator is used to introduce random disturbances in the evolutionary strategy optimization algorithm to eliminate local optimal traps in the initialization solution space, and the optimal traffic light switching strategy is searched in the initialization solution space to generate an evolutionary optimization solution set, including: introducing a fitness evaluation function, calculating the fitness value of each candidate initialization solution in the initialization solution space under the current traffic conditions, increasing the genetic diversity of the initialization solution space through diversity enhancement technology, and generating an optimized initialization solution space; using the evolutionary strategy optimization algorithm to simulate the natural evolution process, and screening individuals with high fitness by iterating the natural evolution process. High fitness individuals, exchange genetic information between the high fitness individuals through the crossover operator, use the mutation operator to introduce random disturbances in the evolution strategy optimization algorithm to eliminate the local optimal traps in the initialization solution space, and search for the optimal traffic light switching strategy in the optimized initialization solution space to generate a natural evolution intermediate solution set; apply dynamic fitness adjustment technology to adjust the fitness evaluation standard for high fitness individuals in real time during the natural evolution process, use the local search algorithm to perform local fine adjustment on the natural evolution intermediate solution set, and generate an optimized natural evolution solution set; introduce a simulated annealing algorithm to set the initial temperature and cooling rate to simulate the thermal motion process in the physical system, balance the global search and local optimization process of the local search algorithm during the initial temperature reduction process, and generate an evolutionary optimization solution set.

[0134] In this embodiment, the fitness evaluation function is a function used to evaluate the performance of candidate solutions under current traffic conditions. By calculating the fitness value of each candidate solution, individuals with high fitness can be screened out for subsequent optimization.

[0135] Diversity enhancement technology is a method to increase the genetic diversity of a population by introducing new individuals or mutating existing individuals to prevent the population from converging to a local optimal solution too early, thereby improving the global search capability.

[0136] The optimized initialization solution space is the initialization solution space processed by the diversity enhancement technology, which contains more potential solutions and helps to find a better traffic light switching strategy.

[0137] The crossover operator is a method of exchanging genetic information between individuals with high fitness during the evolution process, generating new potential solutions by combining the advantages of different individuals.

[0138] The mutation operator is a method of introducing random perturbations to explore new areas of the solution space by changing the eigenvalues ​​of certain individuals, increasing the possibility of finding the global optimal solution.

[0139] The natural evolution intermediate solution set is an intermediate solution set generated after multiple iterations and selections, which contains multiple high-fitness individuals and their combinations, providing a basis for further optimization.

[0140] Dynamic fitness adjustment technology is a technology that adjusts the fitness evaluation criteria in real time according to the evolutionary process, ensuring that high-fitness individuals can remain competitive in a changing environment.

[0141] The local search algorithm is a method of performing a refined search on a specific area, finding a better local solution by deeply exploring the local solution space.

[0142] The optimized natural evolution solution set is a solution set that has been finely adjusted by the local search algorithm and contains individuals with higher fitness, providing high-quality candidate solutions for the final optimization.

[0143] The simulated annealing algorithm is an optimization algorithm that simulates the thermal motion process of a physical system. By setting the initial temperature and cooling rate, it strikes a balance between global search and local optimization to avoid falling into the local optimal solution.

[0144] The evolutionary optimization solution set is the final solution set after being processed by the simulated annealing algorithm, which contains the global optimal or near-global optimal traffic light switching strategy.

[0145] In the embodiment of the present application, first, a fitness evaluation function is introduced to calculate the fitness value of each candidate initialization solution in the initialization solution space under the current traffic conditions, and the genetic diversity of the initialization solution space is increased by diversity enhancement technology to generate an optimized initialization solution space; secondly, an evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and high-fitness individuals are screened by iterating the natural evolution process, and the genetic information between high-fitness individuals is exchanged by the crossover operator. The mutation operator is used to introduce random disturbances, and the local optimal traps in the initialization solution space are eliminated. The optimal traffic light switching strategy is searched in the optimized initialization solution space to generate a natural evolution intermediate solution set; thirdly, the dynamic fitness adjustment technology is applied to adjust the fitness evaluation criteria for high-fitness individuals in real time during the natural evolution process, and the local search algorithm is used to perform local fine adjustment on the natural evolution intermediate solution set to generate an optimized natural evolution solution set; finally, a simulated annealing algorithm is introduced to set the initial temperature and cooling rate to simulate the thermal motion process in the physical system, and the global search and local optimization processes of the local search algorithm are balanced during the initial temperature reduction process to generate an evolutionary optimization solution set.

[0146] Here is a specific example:

[0147] Suppose that in a smart city traffic hub management project, it is necessary to optimize the traffic light switching strategy during peak hours to reduce congestion and improve traffic flow. First, the system obtains real-time vehicle speed, pedestrian flow and surrounding construction activity data from multiple sensor nodes to generate an initial solution space, and introduces a fitness evaluation function to calculate the fitness value of each candidate solution. The genetic diversity of the initial solution space is increased through diversity enhancement technology to generate an optimized initial solution space; secondly, the evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and high-fitness individuals are selected through iterative natural evolution processes. The genetic information between high-fitness individuals is exchanged through the crossover operator, and the mutation operator is used to introduce random disturbances to eliminate local optimal traps in the initial solution space. The optimal traffic light switching strategy is searched in the optimized initial solution space to generate an optimal traffic light switching strategy. The method generates an optimized natural evolution solution set by using a simulated annealing algorithm to set the initial temperature and cooling rate, and balance the global search and local optimization processes of the local search algorithm during the initial temperature reduction process to generate an evolutionary optimized solution set, which helps traffic management departments adjust the traffic light time during peak hours, add temporary lanes or adjust route planning, thereby improving overall traffic efficiency and citizens’ travel experience.

[0148] In order to further improve the continuous optimization capability and adaptability of the traffic light switching strategy, in some embodiments, the feedback loop mechanism is introduced in step 104 to collect the operation results of the future traffic state prediction model in real time, and automatically form an adaptability improvement report based on the potential improvement points in the traffic light switching strategy evaluation results. By comparing the operation results with the adaptability improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted to optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation scheme, including: introducing a feedback loop mechanism to collect the operation results of the future traffic state prediction model in real time, and applying data mining technology to the heterogeneous data in the operation results. The traffic lights are integrated to generate heterogeneous data sets; according to the potential improvement points in the traffic light switching strategy evaluation results, the potential improvement points are automatically learned by using a machine learning algorithm, the difference patterns are obtained, and an adaptive improvement report is generated; by comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted to optimize the traffic light passing signals in the dynamic panoramic view of the traffic conditions, and a passing signal optimization plan is generated; a path planning algorithm is introduced to perform signal timing on the key intersections in the passing signal optimization plan, and the traffic light switching time in the passing signal optimization plan is adjusted according to the signal timing results to generate a traffic light passing calculation plan.

[0149] In this embodiment, the feedback loop mechanism is a continuous improvement process that adjusts and optimizes system performance by continuously collecting actual results of system operation and comparing them with expected goals.

[0150] Data mining technology is a method to extract useful information from large amounts of heterogeneous data. By integrating the heterogeneous data in the running results, high-quality data sets can be generated, providing a basis for subsequent analysis.

[0151] Heterogeneous datasets are datasets that contain multiple types and sources, including sensor data and historical traffic data.

[0152] Machine learning algorithm is a technology used to automatically learn and identify patterns. By analyzing potential improvement points in the traffic light switching strategy evaluation results, the machine learning algorithm can automatically discover difference patterns and generate adaptive improvement reports.

[0153] The adaptive improvement report is a document generated based on the machine learning algorithm, which contains specific improvement suggestions and optimization directions.

[0154] The spatiotemporal series parameters are the spatiotemporal related parameters that affect the prediction accuracy of the future traffic state prediction model. These parameters determine the sensitivity of the model to temporal and spatial changes.

[0155] The traffic light signal optimization scheme is a traffic light signal optimization scheme generated based on the optimized spatiotemporal sequence parameters, aiming to improve traffic efficiency and smoothness.

[0156] The path planning algorithm is an algorithm used to optimize traffic flow. It ensures smooth and efficient traffic flow by adjusting the signal timing at key intersections.

[0157] The traffic light calculation plan is the final optimization plan, which contains detailed adjustment suggestions for the traffic light switching time to improve the overall traffic management efficiency.

[0158] In the embodiment of the present application, first, a feedback loop mechanism is introduced to collect the operation results of the future traffic state prediction model in real time, and the data mining technology is used to integrate the heterogeneous data in the operation results to generate a heterogeneous data set; secondly, according to the potential improvement points in the traffic light switching strategy evaluation results, the machine learning algorithm is used to automatically learn the potential improvement points, obtain the difference pattern, and generate an adaptive improvement report; thirdly, by comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted, the traffic light passing signal in the dynamic panoramic view of the traffic condition is optimized, and a passing signal optimization plan is generated; finally, a path planning algorithm is introduced to perform signal timing on the key intersections in the passing signal optimization plan, and the traffic light switching time in the passing signal optimization plan is adjusted according to the signal timing results to generate a traffic light passing calculation plan.

[0159] Here is a specific example:

[0160] Assume that in a smart city traffic hub optimization project, it is necessary to dynamically adjust the traffic light switching strategy to cope with traffic flow changes at different times. First, the system obtains real-time vehicle speed, pedestrian flow and surrounding building activity data from multiple sensor nodes, and collects the operation results of the future traffic state prediction model in real time. The data mining technology is used to integrate the heterogeneous data in the operation results to generate a heterogeneous data set; secondly, according to the potential improvement points in the traffic light switching strategy evaluation results, the machine learning algorithm is used to automatically learn the potential improvement points, obtain the difference pattern, and generate an adaptive improvement report; thirdly, by comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted, the traffic light passing signal in the dynamic panoramic view of the traffic situation is optimized, and the passing signal optimization plan is generated; finally, the path planning algorithm is introduced to perform signal timing on the key intersections in the passing signal optimization plan, and the traffic light switching time in the passing signal optimization plan is adjusted according to the signal timing results to generate a traffic light passing calculation plan. In this way, the traffic management department can add temporary lanes or adjust the signal light time during peak hours according to the optimized plan, thereby improving the overall traffic efficiency and citizen travel experience.

[0161] This application considers that in order to further improve the accuracy and reliability of causal relationship recognition of dynamic panoramic images of traffic conditions, the existing technology has problems such as inaccurate causal relationship recognition, difficulty in quantifying indirect impact paths, and poor adaptability to change rules in different time periods. Therefore, the embodiment of the invention proposes this optional solution, which solves the above technical problems by applying Granger causality test, partial correlation analysis, path analysis technology and Bayesian network model. Therefore, a new optional solution is proposed, which includes:

[0162] Using a causal inference algorithm, the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic characteristic data is inferred, the key influencing factors of the causal relationship are identified based on the correlation analysis results, and a causal graph is constructed to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and a traffic condition causal chain is generated, including:

[0163] Granger causality test is applied to verify the validity of the correlation analysis results, significant correlation data pairs in the pre-processed traffic characteristic data are screened out, partial correlation analysis is used to exclude potential interference factors in the significant correlation data pairs, and small constants are introduced for numerical stabilization processing to generate direct impact path coefficients;

[0164] The direct impact path coefficient is calculated using the following formula:

[0165] ;

[0166] in, It is Feature values ​​in preprocessed traffic feature data; It is The corresponding historical observation values ​​of the preprocessed traffic characteristic data; and They are and The average value of To prevent small constants from having zero values ​​in the logarithmic function; is the index of the preprocessed traffic feature data, from 1 to ; is the amount of preprocessed traffic characteristic data;

[0167] Applying path analysis technology to identify the indirect impact path coefficient corresponding to the pre-processed traffic characteristic data, building a Bayesian network model in combination with the direct impact path coefficient, performing probabilistic inference on the nodes in the causal graph, introducing adjustment factors to correct the indirect impact path coefficient, and introducing weight factors to balance the relative importance of the indirect impact path coefficient, so as to generate causal chain strength;

[0168] The causal chain strength is calculated using the following formula:

[0169] ;

[0170] in, is the strength of the causal chain; is the direct impact path coefficient; It is The indirect impact path coefficient corresponding to the pre-processed traffic characteristic data; Parameters for adjusting the degree and weight of nonlinearity, which are used to adjust the relative contribution of indirect effects in the overall causal chain; Parameters for adjusting the weights of indirect effect path coefficients; Parameters for adjusting the weight of the sinusoidal function for the interaction between direct and indirect effects; Parameters for adjusting the cosine transformation weights of the indirect effect path coefficients; The parameter for adjusting the weight of the tangent function for the interaction between direct and indirect effects; Parameters for adjusting the logarithmic transformation weights of the indirect effect path coefficients;

[0171] The dynamic time warping technology is used to dynamically adjust the strength of the causal chain to adapt to the changing rules in different time periods. The strength of the causal chain is simulated and verified through an offline simulation platform to generate a traffic condition causal chain.

[0172] This method aims to improve the accuracy and reliability of causal relationship identification of traffic conditions. It combines Granger causality test, partial correlation analysis, path analysis technology and Bayesian network model to capture multi-level features from local details to global structure, and enhance sensitivity to dynamic changes through spatiotemporal fusion, ultimately achieving accurate causal relationship identification and classification of traffic events.

[0173] Exponential transformation :This part is used to nonlinearly adjust the importance of the indirect effect path coefficient. Exponential transformation can amplify or reduce the impact of the indirect effect path coefficient, making the causal chain strength more reasonable; it helps to highlight the role of important paths and ensure that the results are always positive, avoiding calculation problems caused by negative values; sine interaction terms :This part is used to capture the interaction between direct and indirect impact paths. The sine function can effectively describe the periodicity and nonlinear characteristics of this interaction, help identify complex dynamic relationships, and avoid the influence of extreme values ​​through the cosine adjustment term to ensure the stability of the results; the tangent interaction term :This part is used to further enhance the interaction between direct and indirect impact paths. The tangent function can effectively describe the drastic change characteristics of this interaction, help identify and quantify complex causal relationships, and avoid the influence of extreme values ​​through logarithmic adjustment items to ensure the stability of the results;

[0174] in, Calculated according to the direct impact path coefficient formula; Calculated using path analysis techniques; Set according to the actual application scenario requirements;

[0175] Suppose that we need to evaluate the impact of a traffic accident on a main road during a city's peak traffic period on the congestion of surrounding roads; suppose that the eigenvalue in the i-th preprocessed traffic feature data is for , the corresponding historical observations for , and They are 130 and 0.9 respectively. is 0.01;

[0176] ;

[0177] Hypothesized Moderating Factors , ;

[0178] ;

[0179] Assuming that the threshold is set to 1.0, since the calculated result 1.23 is greater than the set threshold, it shows that the traffic condition causal relationship analysis scheme has high accuracy and reliability, and can effectively identify the impact of traffic accidents on surrounding road congestion. This is because the higher detection results reflect that the model can accurately identify the causal relationship between traffic events under the current conditions without affecting the efficiency and accuracy of the analysis. Through the above steps, the accuracy and dynamic adaptability of the causal relationship of traffic conditions are ensured, the scientificity and reliability of traffic management are improved, and the response speed and optimization capabilities of the entire traffic system are enhanced.

[0180] This application considers that in order to further improve the optimization accuracy and reliability of the traffic light switching strategy, the existing technology is prone to fall into the local optimal solution, it is difficult to find the global optimal solution and the dynamic change adaptability is poor. Therefore, the embodiment of the invention proposes this optional solution, which solves the above technical problems by applying the evolution strategy optimization algorithm, principal component analysis, K-means clustering, fitness function calculation and differential evolution algorithm. Therefore, a new optional solution is proposed, which includes:

[0181] An evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and high fitness individuals are selected by iterating the natural evolution process. The genetic information between the high fitness individuals is exchanged by a crossover operator, and a random disturbance is introduced into the evolutionary strategy optimization algorithm by a mutation operator to eliminate the local optimal trap in the initialization solution space, and the optimal traffic light switching strategy is searched in the optimized initialization solution space to generate a natural evolution intermediate solution set, including:

[0182] Applying principal component analysis to reduce the dimension of the optimized initialization solution space, extracting strong representative feature vectors, using K-means clustering algorithm to classify the high fitness individuals according to the strong representative feature vectors, introducing adjustment parameters to adjust the traffic flow values ​​of the individuals in the optimized initialization solution space to generate fitness function values;

[0183] The fitness function value is calculated using the following formula:

[0184] ;

[0185] in, For the The fitness function value of each individual; Initialize the index of individuals in the solution space for optimization; is the feature index of the individual contained in the optimized initialization solution space, ranging from 1 to ; To optimize the number of features contained in individuals in the initial solution space; It is The individual in Traffic flow value on each feature; It is the adjustment parameter that controls the rate of influence of the exponential decay part; is the tuning parameter that adjusts the strength of the exponential decay; is the adjustment parameter that adjusts the overall weight of the product term, and They are adjustment parameters for controlling the scale factor and intensity of the hyperbolic sine function, respectively;

[0186] The genetic information between the high-fitness individuals is exchanged through the crossover operator in the evolution strategy optimization algorithm, random disturbance is introduced into the evolution strategy optimization algorithm using the mutation operator, and the actual contribution of each path in the evolution strategy optimization algorithm is determined by introducing the substitution influence coefficient and the dynamic environment influence factor to generate the optimized evolution strategy path weight;

[0187] The optimized evolution strategy path weight is calculated using the following formula:

[0188] ;

[0189] in, To optimize the evolution strategy path weights; is the individual index in the evolution strategy; is the total number of individuals in the evolution strategy; For the The fitness function value of each individual; For the The substitution influence coefficient of an individual indicates the degree of influence of the individual on other factors; For the The dynamic environmental impact factor of each individual indicates the performance of the individual under different environmental conditions; is the time-varying influence function; To control the impact rate of fitness deviation; is the target fitness value, which indicates the ideal state of individual fitness; To adjust the weight of the substitution impact coefficient; is the proportional factor, which is used to adjust the proportion of the substitution influence coefficient; is the offset, ensuring that the logarithmic term is always positive; is the weight of the hyperbolic sine function, which determines its importance in the overall calculation; is the proportional factor of the dynamic environmental impact factor; is the offset of the dynamic environmental impact factor; is the overall weight of the integral part; and The starting point and the end point of the integration interval are defined respectively, reflecting the time span of the evaluation;

[0190] The differential evolution algorithm is applied to globally optimize the weights of the optimized evolution strategy paths, and fuzzy comprehensive evaluation analysis is used to calculate the fuzzy evaluation matrix to perform fuzzy comprehensive evaluation on the global optimization results to generate a natural evolution intermediate solution set.

[0191] This method aims to improve the accuracy and reliability of traffic light switching strategy optimization. It combines evolutionary strategy optimization algorithm, principal component analysis, K-means clustering, fitness function calculation and differential evolution algorithm to capture multi-level features from local details to global structure, and enhances sensitivity to dynamic changes through global optimization, ultimately achieving accurate traffic light switching strategy optimization.

[0192] In the fitness function value, the exponential decay term ): used to control the rate of influence of the exponential decay part, helping to avoid the local optimal trap; hyperbolic sine product term ): It is used to adjust the overall weight of the product term, and describe the complex nonlinear relationship through the hyperbolic sine function to help identify and separate the interactions between different features;

[0193] in, Acquired from real-time traffic monitoring system; Set according to the actual application scenario requirements;

[0194] In optimizing the evolution strategy path weight, the fitness deviation term : Used to control the impact rate of fitness deviation to ensure that individual fitness is close to the ideal state; replace the influence coefficient term ): used to adjust the weight of the alternative impact coefficient to ensure its importance in the overall calculation; dynamic environmental impact factor term ( ): used to determine the importance of dynamic environmental factors and reflect the performance of individuals under different environmental conditions; integral part : It is used to evaluate the overall contribution of the time-varying impact function and ensure the rationality of the evaluation time span;

[0195] in, Calculated according to the fitness function value formula; and Obtained through real-time monitoring and historical data analysis; Set according to the actual application scenario requirements; It is a time-varying influence function, determined according to the specific scenario;

[0196] Suppose you need to optimize the traffic light switching strategy on the main road during a city’s traffic rush hour. Suppose the feature index of the kth individual is for , the corresponding traffic flow value for , adjust the parameters ;

[0197] ;

[0198] Hypothesized Moderating Factors

[0199] ;

[0200] ;

[0201] Assuming that the threshold is set to 1.0, since the calculated result 1.15 is greater than the set threshold, it shows that the traffic light switching strategy optimization scheme has high accuracy and reliability, and can effectively optimize the traffic light switching strategy. This is because the higher detection result reflects that the model can accurately identify and optimize the traffic light switching strategy under the current conditions without affecting the analysis efficiency and accuracy. Through the above steps, the accuracy and dynamic adaptability of the traffic light switching strategy are ensured, the scientificity and reliability of traffic management are improved, and the response speed and optimization ability of the entire traffic system are enhanced. Through the above steps, the accuracy and dynamic adaptability of the causal relationship of traffic conditions are ensured, the scientificity and reliability of traffic management are improved, and the response speed and optimization ability of the entire traffic system are enhanced.

[0202] Figure 2 A schematic diagram of a traffic light computing system based on the Internet of Things is provided for an embodiment of the present application, such as Figure 2As shown, the device comprises:

[0203] The receiving module 21 is used to receive the real-time data stream from the distributed IoT smart sensor network, integrate the vehicle dynamic information and pedestrian flow statistics in the real-time data stream, perform time-space consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision time-space synchronization technology, and generate a dynamic panoramic view of the traffic situation;

[0204] A construction module 22 is used to use a causal inference algorithm to infer the causal relationship of historical data in the dynamic panoramic map of traffic conditions, identify key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions, use spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics to generate a future traffic state prediction model;

[0205] A search module 23 is used to generate a solution space based on the future traffic state prediction model, use an evolutionary strategy optimization algorithm to simulate a natural evolution process, eliminate local optimal traps in the solution space by iterating the natural evolution process, and search for an optimal traffic light switching strategy in the solution space, adopt a multi-objective optimization technique to consider mutually exclusive goals, introduce minimization of waiting time and maximization of traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the solution space, and generate a traffic light switching strategy evaluation result;

[0206] The adjustment module 24 is used to introduce a feedback loop mechanism, collect the operation results of the future traffic state prediction model in real time, automatically form an adaptive improvement report based on the potential improvement points in the traffic light switching strategy evaluation results, and cyclically adjust the spatiotemporal sequence parameters in the future traffic state prediction model by comparing the operation results with the adaptive improvement report, optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation plan.

[0207] Figure 2 The traffic light based on the Internet of Things can be executed by the computing system Figure 1 The implementation principle and technical effect of the traffic light calculation method based on the Internet of Things described in the embodiment shown are not repeated here. The specific way in which each module and unit performs operations in the traffic light calculation system based on the Internet of Things in the above embodiment has been described in detail in the embodiment of the method, and will not be elaborated here.

[0208] In one possible design, Figure 2The traffic light based on the Internet of Things in the embodiment shown can be implemented as a computing device through a computing system, such as Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0209] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are called and executed by the processing component 32 .

[0210] The processing component 32 is used to: receive real-time data streams from a distributed IoT smart sensor network, integrate vehicle dynamic information and pedestrian flow statistics in the real-time data stream, and perform spatiotemporal consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision spatiotemporal synchronization technology to generate a dynamic panoramic view of traffic conditions; use a causal inference algorithm to infer the causal relationship of historical data in the dynamic panoramic view of traffic conditions, identify key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, use spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic view of traffic conditions based on the spatiotemporal characteristics to generate a future traffic status prediction model; generate a solution space based on the future traffic status prediction model, and use the progress The strategy optimization algorithm simulates the natural evolution process, eliminates the local optimal traps in the solution space by iterating the natural evolution process, searches for the optimal traffic light switching strategy in the solution space, adopts multi-objective optimization technology to consider mutually exclusive goals, introduces minimizing waiting time and maximizing traffic fluency as constraints, comprehensively evaluates the optimal traffic light switching strategy in the solution space, and generates a traffic light switching strategy evaluation result; introduces a feedback loop mechanism to collect the operation results of the future traffic state prediction model in real time, and automatically forms an adaptive improvement report based on the potential improvement points in the traffic light switching strategy evaluation results. By comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted to optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation plan.

[0211] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above method. Of course, the processing component may also be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors or other electronic components to perform the above method.

[0212] The storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic disk or optical disk.

[0213] Of course, the computing device may also include other components, such as input / output interfaces, display components, communication components, etc.

[0214] The input / output interface provides an interface between the processing component and the peripheral interface module, which may be an output device, an input device, etc.

[0215] The communication component is configured to facilitate, among other things, wired or wireless communications between the computing device and other devices.

[0216] Among them, the computing device can be a physical device or an elastic computing host provided by a cloud computing platform, etc. In this case, the computing device can refer to a cloud server, and the above-mentioned processing components, storage components, etc. can be basic server resources rented or purchased from the cloud computing platform.

[0217] The present application also provides a computer storage medium storing a computer program, wherein the computer program can achieve the above-mentioned Figure 1 The illustrated embodiment is a traffic light calculation method based on the Internet of Things.

[0218] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0219] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Those of ordinary skill in the art may understand and implement it without creative work.

[0220] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0221] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit it. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A traffic light calculation method based on the Internet of Things, characterized in that: include: Receive real-time data streams from a distributed IoT smart sensor network, integrate vehicle dynamic information and pedestrian flow statistics in the real-time data streams, perform spatiotemporal consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision spatiotemporal synchronization technology, and generate a dynamic panoramic view of traffic conditions; Using a causal inference algorithm, inferring the causal relationship of historical data in the dynamic panoramic map of traffic conditions, identifying key influencing factors of the causal relationship, and constructing a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions, using spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulating the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics, and generating a future traffic state prediction model; Generate a solution space based on the future traffic state prediction model, use an evolutionary strategy optimization algorithm to simulate the natural evolution process, eliminate local optimal traps in the solution space by iterating the natural evolution process, and search for the optimal traffic light switching strategy in the solution space, adopt a multi-objective optimization technology to consider mutually exclusive goals, introduce minimizing waiting time and maximizing traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the solution space, and generate a traffic light switching strategy evaluation result; A feedback loop mechanism is introduced to collect the operation results of the future traffic state prediction model in real time, and an adaptive improvement report is automatically generated according to the potential improvement points in the traffic light switching strategy evaluation results. By comparing the operation results with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are cyclically adjusted to optimize the traffic light passing signals in the dynamic panoramic view of the traffic conditions, and generate a traffic light passing calculation plan.

2. The method according to claim 1, characterized in that The causal inference algorithm is used to infer the causal relationship of historical data in the dynamic panoramic view of traffic conditions, identify the key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions. The spatiotemporal sequence modeling technology is used to capture the spatiotemporal characteristics of the direct and indirect impact paths, and the time-varying laws of traffic flow in the dynamic panoramic view of traffic conditions are accurately simulated according to the spatiotemporal characteristics to generate a future traffic state prediction model, including: Applying data smoothing technology to perform outlier detection and noise removal processing on vehicle dynamic information and pedestrian flow statistical data in the dynamic panoramic view of traffic conditions, applying feature selection algorithm to extract traffic flow state features from the outlier detection and noise removal processing results, and generating pre-processed traffic feature data; Using a causal inference algorithm, inferring the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic characteristic data, identifying key influencing factors of the causal relationship, and constructing a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and generating a causal relationship graph of traffic events; Adopting the spatiotemporal sequence modeling technology, taking the traffic event causal relationship map as the input variable to perform spatiotemporal sequence modeling, capturing the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulating the time-varying law of the traffic flow in the dynamic panoramic map of the traffic situation according to the spatiotemporal characteristics, and generating a spatiotemporal characteristic model; The long-term structural impact in the causal relationship map of the traffic event and the real-time dynamic change details in the spatiotemporal characteristic model are extracted, and the long-term structural impact and the real-time dynamic change details are segmented using the sliding window technology. The segmented processing results are integrated into a hybrid prediction model to generate a future traffic status prediction model.

3. The method according to claim 2, characterized in that The causal inference algorithm is used to infer the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic characteristic data, identify the key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and generate a causal relationship graph of traffic events, including: Introducing a cross-correlation function to perform correlation analysis on the pre-processed traffic characteristic data, so as to mine the potential influencing patterns and periodic changes of traffic flow stability in the pre-processed traffic characteristic data, and generate correlation analysis results; Using a causal inference algorithm, inferring the causal relationship between the historical data in the dynamic panoramic view of traffic conditions and the pre-processed traffic characteristic data, identifying the key influencing factors of the causal relationship based on the correlation analysis results, and constructing a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic view of traffic conditions, and generating a traffic condition causal chain; Applying the Shapley value decomposition method to quantify the influence of each causal path in the traffic condition causal chain on the traffic events in the traffic condition dynamic panoramic map, separating the long-term trend and seasonal trend of the traffic condition causal chain, and generating the causal chain time trend; A graph layout algorithm is applied to define a basic graph structure according to the causal chain time trend, and the long-term trend and seasonal trend are mapped to the basic graph structure to generate a traffic event causal relationship graph.

4. The method according to claim 2, characterized in that: The spatiotemporal sequence modeling technology is used to perform spatiotemporal sequence modeling with the traffic event causal relationship map as an input variable, capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic view of the traffic situation according to the spatiotemporal characteristics to generate a spatiotemporal characteristic model, including: The data normalization technology is applied to standardize the nodes and edges in the traffic incident causal relationship map, and the spatial attributes of the traffic incident causal relationship map are refined by introducing a geographic information system to generate standardized causal relationship vector data; Adopting the spatiotemporal sequence modeling technology, taking the standardized causal relationship vector data and the traffic event causal relationship map as input variables to perform spatiotemporal sequence modeling, capturing the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulating the time-varying laws of the traffic flow in the dynamic panoramic map of the traffic situation according to the spatiotemporal characteristics, and generating dynamic pattern recognition results; Applying pattern clustering technology to deeply analyze the dynamic pattern recognition results, identifying and classifying typical behavior patterns in the dynamic pattern recognition results, and using adaptive filtering algorithms to optimize the spatiotemporal sequence modeling process to generate optimized dynamic pattern recognition results; Multi-scale analysis technology is introduced to capture the dynamic changes of traffic flow at different time scales in the optimized dynamic pattern recognition results, while considering the detailed spatial distribution information in the geographic information system, and an integrated learning framework is applied to integrate the dynamic changes of traffic flow and the detailed spatial distribution information to generate a spatiotemporal characteristic model.

5. The method according to claim 1, characterized in that The solution space is generated based on the future traffic state prediction model, and the evolutionary strategy optimization algorithm is used to simulate the natural evolution process. By iterating the natural evolution process, the local optimal trap in the solution space is eliminated, and the optimal traffic light switching strategy is searched in the solution space. The multi-objective optimization technology is used to consider mutually exclusive goals, and the minimization of waiting time and maximization of traffic smoothness are introduced as constraints. The optimal traffic light switching strategy in the solution space is comprehensively evaluated, and the traffic light switching strategy evaluation result is generated, including: Applying data mapping technology, converting the output result of the future traffic state prediction model into a multi-dimensional vector representation, introducing an initial population generation algorithm, randomly distributing a group of candidate solutions as the starting point of the solution space based on the multi-dimensional vector representation, and generating an initialized solution space; An evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and individuals with high fitness are selected by iterating the natural evolution process. A mutation operator is used to introduce random disturbances into the evolutionary strategy optimization algorithm to eliminate local optimal traps in the initialization solution space, and an optimal traffic light switching strategy is searched in the initialization solution space to generate an evolutionary optimization solution set. Multi-objective optimization technology is used to consider mutually exclusive objectives, the evolutionary optimization solution set is introduced as the initial test condition, and minimizing waiting time and maximizing traffic flow are additionally introduced as constraint conditions, and the optimal traffic light switching strategy in the initial solution space is comprehensively evaluated to generate a multi-objective optimization solution; Taking the evolutionary optimization solution set and the multi-objective optimization solution as decision criteria, a decision matrix analysis method is introduced to construct a scoring system, further verify the effectiveness of the optimal traffic light switching strategy, and generate a traffic light switching strategy evaluation result.

6. The method according to claim 5, characterized in that The method uses the evolution strategy optimization algorithm to simulate the natural evolution process, selects individuals with high fitness by iterating the natural evolution process, introduces random disturbances into the evolution strategy optimization algorithm by using mutation operators to eliminate local optimal traps in the initialization solution space, and searches for the optimal traffic light switching strategy in the initialization solution space to generate an evolution optimization solution set, including: Introducing a fitness evaluation function, calculating the fitness value of each candidate initialization solution in the initialization solution space under the current traffic conditions, increasing the genetic diversity of the initialization solution space through diversity enhancement technology, and generating an optimized initialization solution space; An evolutionary strategy optimization algorithm is used to simulate the natural evolution process, and high-fitness individuals are selected by iterating the natural evolution process. The genetic information between the high-fitness individuals is exchanged by a crossover operator, and a random disturbance is introduced into the evolutionary strategy optimization algorithm by a mutation operator to eliminate the local optimal trap in the initialization solution space, and the optimal traffic light switching strategy is searched in the optimized initialization solution space to generate a natural evolution intermediate solution set; Applying dynamic fitness adjustment technology, adjusting the fitness evaluation criteria for high fitness individuals in real time during the natural evolution process, and using a local search algorithm to perform local fine adjustment on the natural evolution intermediate solution set to generate an optimized natural evolution solution set; A simulated annealing algorithm is introduced to set the initial temperature and cooling rate to simulate the thermal motion process in the physical system. During the initial temperature reduction process, the global search and local optimization processes of the local search algorithm are balanced to generate an evolutionary optimization solution set.

7. The method according to claim 1, characterized in that The feedback loop mechanism is introduced to collect the operation results of the future traffic state prediction model in real time, automatically form an adaptive improvement report according to the potential improvement points in the traffic light switching strategy evaluation results, and cyclically adjust the spatiotemporal sequence parameters in the future traffic state prediction model by comparing the operation results with the adaptive improvement report, optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation scheme, including: Introducing a feedback loop mechanism to collect the operation results of the future traffic state prediction model in real time, and applying data mining technology to integrate heterogeneous data in the operation results to generate a heterogeneous data set; According to the potential improvement points in the traffic light switching strategy evaluation results, automatically learning the potential improvement points using a machine learning algorithm, obtaining difference patterns, and generating an adaptive improvement report; By comparing the operation result with the adaptive improvement report, the spatiotemporal sequence parameters in the future traffic state prediction model are adjusted cyclically, the traffic light passing signal in the dynamic panoramic view of the traffic condition is optimized, and a passing signal optimization plan is generated; A path planning algorithm is introduced to perform signal timing on key intersections in the through signal optimization scheme, and the traffic light switching time in the through signal optimization scheme is adjusted according to the signal timing result to generate a traffic light through calculation scheme.

8. A traffic light calculation system based on the Internet of Things, characterized in that: include: A receiving module is used to receive a real-time data stream from a distributed IoT smart sensor network, integrate vehicle dynamic information and pedestrian flow statistics in the real-time data stream, perform spatiotemporal consistency processing on the vehicle dynamic information and pedestrian flow statistics through high-precision spatiotemporal synchronization technology, and generate a dynamic panoramic view of traffic conditions; A construction module is used to use a causal inference algorithm to infer the causal relationship of historical data in the dynamic panoramic map of traffic conditions, identify key influencing factors of the causal relationship, and construct a causal graph to calculate the direct and indirect impact paths between traffic events in the dynamic panoramic map of traffic conditions, use spatiotemporal sequence modeling technology to capture the spatiotemporal characteristics of the direct and indirect impact paths, and accurately simulate the time-varying laws of traffic flow in the dynamic panoramic map of traffic conditions based on the spatiotemporal characteristics to generate a future traffic state prediction model; A search module is used to generate a solution space based on the future traffic state prediction model, use an evolutionary strategy optimization algorithm to simulate a natural evolution process, eliminate local optimal traps in the solution space by iterating the natural evolution process, and search for an optimal traffic light switching strategy in the solution space, adopt a multi-objective optimization technique to consider mutually exclusive goals, introduce minimizing waiting time and maximizing traffic smoothness as constraints, comprehensively evaluate the optimal traffic light switching strategy in the solution space, and generate a traffic light switching strategy evaluation result; The adjustment module is used to introduce a feedback loop mechanism, collect the operation results of the future traffic state prediction model in real time, automatically form an adaptive improvement report based on the potential improvement points in the traffic light switching strategy evaluation results, and cyclically adjust the spatiotemporal sequence parameters in the future traffic state prediction model by comparing the operation results with the adaptive improvement report, optimize the traffic light passing signal in the dynamic panoramic view of the traffic condition, and generate a traffic light passing calculation plan.

9. A computing device, characterized in that It comprises a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement a traffic light passing calculation method based on the Internet of Things as described in any one of claims 1 to 7.

10. A computer storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a computer, a traffic light passing calculation method based on the Internet of Things as described in any one of claims 1 to 7 is implemented.

Images