Traffic signal lamp intelligent control method and system, electronic device, and storage medium

Abstract

The application provides a traffic signal intelligent control method and system, an electronic device and a storage medium, and belongs to the technical field of traffic signal control. The method collects intersection traffic state data through different types of detectors to obtain monitoring data of each control parameter, extracts traffic characteristic values after standardizing the monitoring data, performs correlation analysis on the traffic states to obtain the correlation degree between the parameters, evaluates the data quality by using a pre-trained signal control evaluation model, determines the control reliability in combination with the correlation degree, and generates a signal control scheme by optimizing the data according to the correlation degree when the reliability is lower than a threshold. The control parameters include vehicle flow, queue length and the like. The system comprises acquisition, processing and execution modules. The electronic device and the storage medium implement the execution of the method. The scheme improves the data reliability, optimizes the control scheme, improves the intersection passing efficiency and is suitable for the intelligent traffic signal control scene through multi-dimensional data acquisition, correlation analysis and quality evaluation.

Description

Technical Field

[0001] This application belongs to the field of traffic signal control technology, and more specifically, relates to intelligent control methods and systems for traffic lights, electronic devices, and storage media. Background Technology

[0002] In urban traffic management, the effectiveness of traffic light control directly impacts intersection efficiency and traffic flow. Traditional traffic light control methods mostly employ fixed timing schemes. These schemes pre-set signal durations based on historical traffic data, making it difficult to adapt to dynamic factors such as real-time changes in traffic flow, vehicle queue lengths, and pedestrian waiting times. When sudden traffic situations or flow fluctuations occur at intersections, fixed timing schemes can easily lead to situations where vehicles in one direction wait for extended periods while resources are wasted in the other direction, resulting in traffic congestion or low efficiency.

[0003] With the development of intelligent transportation technology, some control methods have begun to introduce detectors to collect real-time traffic data in order to achieve dynamic adjustment of traffic lights. However, there are still many shortcomings in existing technologies: most schemes rely on only a single type or a small number of detectors to collect data. The limitation of data sources makes it difficult for the acquired traffic monitoring data to comprehensively and accurately reflect the actual traffic conditions at intersections. For example, only obtaining traffic flow data through geomagnetic detectors ignores other important parameters such as pedestrian flow and vehicle queue length, resulting in an incomplete basis for control decisions. When processing traffic data, existing technologies often do not fully consider the inherent correlation between various control parameters, treating parameters such as traffic flow, queue length, and green light utilization rate as independent variables for analysis. They fail to uncover their mutual influence relationships in the traffic structure, making it impossible for control schemes to optimize for the linkage effects between parameters, thus affecting the scientificity and effectiveness of control strategies.

[0004] Existing technologies lack adequate quality assessment for traffic monitoring data, and lack effective mechanisms to evaluate data reliability and accuracy. When detectors malfunction or errors occur during data transmission, poor-quality data directly impacts the generation of signal control schemes, leading to erroneous control decisions and further exacerbating traffic chaos. Furthermore, facing complex traffic environments and diverse traffic scenarios, existing control methods struggle to dynamically adjust control strategies based on data reliability and cannot perform reasonable optimization when data quality is poor, limiting the system's adaptability and robustness. Therefore, a smart traffic light control technology is needed that can comprehensively collect multi-dimensional traffic data, deeply analyze parameter correlations, accurately assess data quality, and optimize control schemes accordingly to address the aforementioned problems in existing technologies. Summary of the Invention

[0005] The purpose of this application is to provide a method and system for intelligent control of traffic lights, electronic devices, and storage media to solve the problems mentioned in the background art.

[0006] To achieve the above objectives, the present invention provides an intelligent control method for traffic lights, the method comprising:

[0007] Traffic status data at intersections is collected by different types of detectors, and traffic monitoring data corresponding to various control parameters are obtained.

[0008] Traffic characteristic values ​​of corresponding control parameters are extracted from various traffic monitoring data, and then traffic state correlation analysis is performed on all traffic characteristic values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure.

[0009] Based on the pre-trained signal control evaluation model, the quality mapping evaluation of traffic monitoring data corresponding to each control parameter is performed to obtain the quality score of traffic monitoring data corresponding to each control parameter. Then, the control reliability of traffic monitoring data corresponding to each control parameter is determined by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure.

[0010] When the control confidence of the traffic monitoring data corresponding to the control parameter is less than the preset confidence threshold, the traffic monitoring data corresponding to the control parameter is optimized according to the state correlation of the traffic structure between the control parameter and other control parameters, and the signal control scheme of the intersection is generated based on the optimized traffic monitoring data.

[0011] In one embodiment of this application, extracting traffic feature values ​​corresponding to control parameters from various traffic monitoring data specifically includes:

[0012] The traffic monitoring data are standardized to obtain the standardized traffic monitoring data.

[0013] Traffic characteristic values ​​corresponding to control parameters are extracted from various traffic monitoring data after standardization.

[0014] In one embodiment of this application, traffic state correlation analysis is performed on all traffic feature values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure. Specifically, this includes:

[0015] Select one control parameter as the selected control parameter;

[0016] A direct correlation analysis is performed on the traffic characteristic values ​​between the selected control parameters and other control parameters to obtain the direct correlation degree of traffic.

[0017] An indirect correlation analysis of traffic structure is performed on the traffic characteristic values ​​between the selected control parameters and other control parameters to obtain the degree of indirect correlation of traffic.

[0018] The state correlation between the selected control parameter and other control parameters in the traffic structure is determined by the direct correlation and the indirect correlation, and then the state correlation between the remaining control parameters and other control parameters in the traffic structure is determined.

[0019] In one embodiment of this application, the quality mapping evaluation of traffic monitoring data corresponding to each control parameter based on a pre-trained signal control evaluation model, to obtain a quality score for the traffic monitoring data corresponding to each control parameter, specifically includes:

[0020] For each control parameter, extract the statistical and distribution characteristics of the traffic monitoring data corresponding to the control parameter;

[0021] The statistical characteristics and the distribution characteristics are used as input parameters for the signal control evaluation model;

[0022] The quality score of traffic monitoring data corresponding to the control parameters is obtained by mapping the signal control evaluation model, and then the quality score of traffic monitoring data corresponding to each control parameter is obtained.

[0023] In one embodiment of this application, the statistical features specifically include the maximum value, the minimum value, and the median.

[0024] In one embodiment of this application, determining the control reliability of the traffic monitoring data corresponding to each control parameter by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure specifically includes:

[0025] Select one control parameter as the selected control parameter;

[0026] The reliability adjustment coefficient of the traffic monitoring data corresponding to the selected control parameter is determined based on the state correlation between the selected control parameter and other control parameters in the traffic structure.

[0027] The quality score of the traffic monitoring data corresponding to the selected control parameter is adjusted by the confidence adjustment coefficient to obtain the control confidence of the traffic monitoring data corresponding to the selected control parameter.

[0028] Continue to determine the control reliability of the traffic monitoring data corresponding to the remaining control parameters.

[0029] In one embodiment of this application, the control parameters specifically include traffic flow, queue length, pedestrian waiting time, and green light utilization rate.

[0030] In one embodiment of this application, the present invention also includes an intelligent traffic signal control system, the system comprising:

[0031] The data acquisition module is used to collect traffic status data at intersections through different types of detectors, and then obtain traffic monitoring data corresponding to various control parameters.

[0032] The processing module is used to extract the traffic feature values ​​of the corresponding control parameters from various traffic monitoring data, and then perform traffic state correlation analysis on all traffic feature values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure.

[0033] The processing module is also used to perform quality mapping evaluation on the traffic monitoring data corresponding to each control parameter based on the pre-trained signal control evaluation model, to obtain the quality score of the traffic monitoring data corresponding to each control parameter, and then determine the control reliability of the traffic monitoring data corresponding to each control parameter by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure.

[0034] The execution module is used to optimize the traffic monitoring data corresponding to the control parameters based on the state correlation of the traffic structure between the control parameters and other control parameters when the control confidence of the traffic monitoring data corresponding to the control parameters is less than the preset confidence threshold, and generate the signal control scheme of the intersection based on the optimized traffic monitoring data.

[0035] In one embodiment of this application, the present invention also includes an electronic device comprising a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the traffic light intelligent control method as described above.

[0036] In one embodiment of this application, the present invention also includes a computer-readable storage medium storing a computer program that, when executed by a processor, implements the above-described intelligent traffic light control method.

[0037] The beneficial effects of the intelligent traffic light control method and system, electronic device, and storage medium provided by this invention are as follows:

[0038] This intelligent traffic light control method and system collects traffic state data at intersections using multiple types of detectors. It comprehensively acquires monitoring data for various control parameters, including traffic flow, queue length, pedestrian waiting time, and green light utilization rate, providing a rich data foundation for subsequent analysis. By extracting feature values ​​from various traffic monitoring data and performing traffic state correlation analysis, the system can uncover the state correlation between each control parameter and other parameters within the traffic structure. This allows the system to accurately grasp the intrinsic relationships between parameters, thereby gaining a more precise understanding of the operational characteristics of traffic flow.

[0039] A pre-trained signal control evaluation model is used to perform quality mapping assessment on traffic monitoring data. This effectively identifies noise and anomalies in the data, obtains a quality score for each data point, and determines control reliability by combining the state correlation between parameters. This further improves the reliability and usability of the data. When the control reliability is lower than a preset threshold, the data is optimized based on the correlation between parameters. This corrects the impact of poor-quality data, ensuring that the generated traffic monitoring data is more accurate and providing a reliable basis for the formulation of signal control schemes.

[0040] Through the aforementioned series of processes, an intersection signal control scheme is generated based on optimized traffic monitoring data. This scheme better adapts to changes in real-time traffic conditions, rationally allocates travel time in each direction, effectively reduces waiting time for vehicles and pedestrians, improves green light utilization and overall intersection efficiency, and alleviates traffic congestion. The application of this system enables intelligent and precise traffic signal control, enhances the automation level and service quality of urban traffic management, and provides strong technical support for improving the urban traffic environment. The configuration of electronic equipment and storage media ensures the stable operation and widespread application of this method, demonstrating high practicality and promotional value. Attached Figure Description

[0041] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0042] Figure 1 This is a schematic diagram illustrating the working principle of the intelligent traffic light control method described in this invention.

[0043] Figure 2 Design diagram for quality mapping evaluation;

[0044] Figure 3 The flowchart is used to determine the reliability of the control.

[0045] Figure 4 This is a flowchart of a method executed by an electronic device. Detailed Implementation

[0046] In the following description, specific details such as particular system architectures and techniques are set forth for illustrative purposes and not for limitation, in order to provide a thorough understanding of the embodiments of this application. However, those skilled in the art will understand that this application may also be implemented in other embodiments without these specific details. In other instances, detailed descriptions of well-known systems, apparatuses, circuits, and methods have been omitted so as not to obscure the description of this application with unnecessary detail.

[0047] To make the objectives, technical solutions, and advantages of this application clearer, the following description will be provided in conjunction with the accompanying drawings and specific embodiments.

[0048] Please see Figures 1-4 This invention provides an intelligent control method for traffic lights, the specific implementation steps of which are as follows:

[0049] Traffic status data at intersections is collected using different types of detectors, which in turn yield traffic monitoring data corresponding to various control parameters. These detectors include, but are not limited to, video detectors, radar detectors, and geomagnetic detectors. Specific control parameters include traffic flow, queue length, pedestrian waiting time, and green light utilization rate.

[0050] Traffic characteristic values ​​of corresponding control parameters are extracted from various traffic monitoring data, and then traffic state correlation analysis is performed on all traffic characteristic values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure.

[0051] The pre-trained signal control evaluation model performs quality mapping evaluation on the traffic monitoring data corresponding to each control parameter, and obtains the quality score of the traffic monitoring data corresponding to each control parameter. Then, the control reliability of the traffic monitoring data corresponding to each control parameter is determined by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure.

[0052] When the control confidence of the traffic monitoring data corresponding to the control parameter is less than the preset confidence threshold, the traffic monitoring data corresponding to the control parameter is optimized according to the state correlation of the traffic structure between the control parameter and other control parameters, and the signal control scheme of the intersection is generated based on the optimized traffic monitoring data.

[0053] Example 1: In this example, when extracting traffic feature values ​​corresponding to control parameters from various traffic monitoring data, it is necessary to first standardize the traffic monitoring data to obtain standardized traffic monitoring data. Different types of traffic monitoring data often have different dimensions and value ranges. For example, traffic flow data may be measured in vehicles per hour, ranging from 0 to several thousand, while queue length data may be measured in meters, ranging from 0 to several hundred, pedestrian waiting time data is measured in seconds, ranging from 0 to several hundred seconds, and green light utilization rate data is usually presented as a percentage, ranging from 0% to 100%. If these data are not standardized, the different dimensions and ranges will significantly interfere with subsequent feature extraction and analysis, making it impossible to accurately compare and process traffic monitoring data corresponding to different control parameters.

[0054] Standardization processing employs a normalization method, the core of which is to convert data with different dimensions and ranges into data of a uniform scale. Specifically, for a set of traffic monitoring data, assuming the maximum value is Max and the minimum value is Min, for each data point x, its standardized value x' can be calculated using the formula x'=(x-Min) / (Max-Min). In this way, regardless of the dimensions and range of the original data, it will be converted to a value within the interval [0,1]. For example, if the original traffic flow data for a certain time period is 500 vehicles / hour, and the maximum traffic flow during that time period is 1000 vehicles / hour and the minimum is 0 vehicles / hour, then after standardization, the value of this data point is (500-0) / (1000-0)=0.5. Similarly, for data with a queue length of 200 meters, if the maximum queue length during that time period is 500 meters and the minimum is 0 meters, then the standardized value is (200-0) / (500-0)=0.4. This standardization process eliminates the influence of dimensions, bringing traffic monitoring data with different control parameters to the same scale and providing a unified basis for subsequent feature extraction.

[0055] After standardization, traffic feature values ​​for the corresponding control parameters need to be extracted from the standardized traffic monitoring data. Different control parameters have different characteristics and analytical requirements, thus requiring different feature extraction methods. For traffic flow control parameters, which reflect the number of vehicles passing through an intersection per unit time, their performance across different time dimensions needs to be considered when extracting traffic feature values. For example, the average traffic flow per unit time reflects the overall traffic flow level within that time period. By calculating the average traffic flow over a specific time period (e.g., 5 minutes, 10 minutes), the approximate traffic flow at the intersection can be understood. The maximum traffic flow reflects the peak traffic volume that may occur within that time period, which is of significant reference value for determining whether congestion will occur at the intersection. The minimum traffic flow reflects the lowest traffic volume within that time period, helping to understand the fluctuation range of traffic flow at the intersection. Furthermore, the variance of the traffic flow can be extracted to measure the dispersion of the traffic flow data; a larger variance indicates greater fluctuation in traffic flow.

[0056] For queue length control parameters, which represent the length of vehicles waiting to cross at an intersection, it's crucial to analyze their trends and peak values ​​across different time periods when extracting traffic characteristic values. For instance, the trend of queue length changes within a given time period can be determined by analyzing queue length data from adjacent time points—whether it's gradually increasing, decreasing, or remaining stable. The peak queue length represents the maximum queue length within that time period, which is critical for assessing the level of congestion at the intersection. Furthermore, the duration of queue length can also be extracted, such as the duration for which the queue length exceeds a certain threshold (e.g., 100 meters), to understand the severity and duration of congestion.

[0057] For pedestrian waiting time control parameters, which relate to the time pedestrians spend waiting for the green light at intersections, the distribution of pedestrian waiting times needs to be considered when extracting traffic feature values. For example, the average pedestrian waiting time reflects the average waiting time for pedestrians at the intersection, which is important for assessing the convenience of pedestrian crossing. The maximum pedestrian waiting time represents the longest possible waiting time for pedestrians, which relates to their crossing experience and safety. Furthermore, the proportion of pedestrians whose waiting time exceeds a certain reasonable threshold (e.g., 60 seconds) can be statistically analyzed to understand how many pedestrians need to wait an extended period to cross the street.

[0058] The green light utilization rate control parameter represents the efficiency with which vehicles utilize green light time. When extracting traffic characteristic values, it is necessary to analyze its performance in different phases. For example, the average green light utilization rate can reflect the overall utilization efficiency of green light time in that phase. By calculating the average green light utilization rate over multiple cycles, the traffic operation status of that phase can be understood. The variation range of the green light utilization rate reflects the stability of the green light time utilization efficiency in that phase; the smaller the variation range, the more stable the utilization of green light time. Furthermore, the green light utilization rate can be analyzed in conjunction with traffic flow. For example, under conditions of high traffic flow, is the green light utilization rate also correspondingly higher? This helps determine whether the allocation of green light time is reasonable.

[0059] When extracting these traffic feature values, it is necessary to select appropriate feature extraction methods and time windows based on specific control parameters and analysis objectives. For example, for real-time traffic control, a shorter time window (such as 1 minute) may be needed to extract feature values ​​to reflect changes in traffic conditions in a timely manner; while for long-term traffic data analysis, a longer time window (such as 1 hour) can be used to extract feature values ​​to obtain more macroscopic traffic characteristics. In addition, the extracted traffic feature values ​​can be compared and analyzed in conjunction with historical traffic data to discover patterns and trends in traffic condition changes.

[0060] By standardizing various traffic monitoring data and extracting traffic feature values ​​for corresponding control parameters, the raw traffic state data can be transformed into more valuable analytical features. This lays the foundation for subsequent traffic state correlation analysis of all traffic feature values ​​and for determining the state correlation between each control parameter and other control parameters in the traffic structure. These extracted traffic feature values ​​can more accurately reflect the actual situation and changing patterns of each control parameter, contributing to a deeper understanding of the traffic state at intersections and the relationships between various control parameters.

[0061] Example 2: In this example, traffic state correlation analysis is performed on all traffic characteristic values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure. The specific implementation process is as follows. First, a control parameter needs to be selected as the selected control parameter, for example, traffic flow is selected as the first analysis object. In the actual operation of traffic intersections, changes in traffic flow do not exist in isolation, but have complex mutual influence relationships with other control parameters such as queue length, pedestrian waiting time, and green light utilization rate. Therefore, it is necessary to carry out the analysis from two levels: direct correlation and indirect correlation.

[0062] When performing direct correlation analysis between selected control parameters and other control parameters' traffic characteristic values, the correlation coefficient method is used to calculate the degree of direct correlation between them. The core of the correlation coefficient method is to reflect the tightness of the direct correlation by quantifying the degree of linear correlation between data. Taking traffic flow and queue length as an example, standardized traffic flow data and standardized queue length data are collected at multiple time points over a period of time, for example, every 5 minutes, to obtain a dataset containing traffic flow x and queue length y. The correlation coefficient of this dataset is calculated, with a value ranging from -1 to 1. When the correlation coefficient is close to 1, it indicates a strong positive correlation between traffic flow and queue length, meaning that as traffic flow increases, queue length tends to increase accordingly; when the correlation coefficient is close to -1, it indicates a strong negative correlation; when the correlation coefficient is close to 0, it indicates a weak linear correlation between the two. In this way, the direct correlation degree between traffic flow and queue length can be obtained. Similarly, the correlation coefficients between traffic flow and pedestrian waiting time, and between traffic flow and green light utilization rate, are calculated to obtain the direct correlation degree between traffic flow and other control parameters.

[0063] After completing the direct correlation analysis, an indirect correlation analysis based on traffic structure is needed to examine the traffic characteristic values ​​between the selected control parameters and other control parameters to obtain the degree of indirect correlation. Indirect correlation analysis requires considering the propagation patterns and influence paths of traffic flow at intersections. This correlation is not a simple linear relationship but involves dynamic changes in traffic flow and the interaction of multiple parameters. For example, an increase in traffic volume initially leads to a decrease in vehicle speed, which in turn causes queue length to gradually accumulate. In this process, changes in traffic volume indirectly affect queue length by influencing the intermediate variable of vehicle speed. To analyze this degree of indirect correlation, a traffic flow propagation model needs to be established. This model is based on traffic flow theory and considers factors such as vehicle following behavior, the number of lanes, and intersection geometry. By inputting traffic volume change data, the model can simulate changes in vehicle speed and thus predict the evolution of queue length. Then, by comparing the actual collected queue length data with the model's prediction results, the deviation between the two is calculated to quantify the degree of indirect correlation between traffic volume and queue length. The smaller the deviation, the higher the indirect correlation, meaning the more significant the impact of traffic flow on queue length through intermediate variables; the larger the deviation, the lower the indirect correlation. Similarly, for the indirect correlation analysis between traffic flow and pedestrian waiting time, the impact of traffic flow changes on vehicle travel time can be considered, which in turn affects the allocation of pedestrian green light time, thus indirectly affecting pedestrian waiting time. By establishing a corresponding time allocation model, the indirect correlation path and degree between changes in traffic flow and pedestrian waiting time can be analyzed.

[0064] After obtaining the direct and indirect correlations, it is necessary to determine the state correlation between the selected control parameter and other control parameters in the traffic structure using these two dimensions. Specifically, a weighted combination method is used to integrate the direct and indirect correlations. For example, the weight of the direct correlation is set to 'a', and the weight of the indirect correlation is set to 'b', where a + b = 1. The weights can be determined based on the analysis of historical traffic data and the characteristics of actual traffic scenarios. Suppose that, after analysis, the direct correlation has a more significant impact on the state correlation in a certain traffic scenario, so a = 0.6 and b = 0.4 are set. Then, the state correlation between the selected control parameter and another control parameter = direct correlation × 0.6 + indirect correlation × 0.4. Through this weighted calculation, the influence of both direct and indirect correlations on the state correlation can be comprehensively considered, resulting in a more comprehensive and accurate state correlation value.

[0065] After calculating the state correlation between the selected control parameter (such as traffic flow) and all other control parameters, it is necessary to continue determining the state correlation between the remaining control parameters and other control parameters in terms of traffic structure using the same method. For example, next, queue length is selected as a new selected control parameter, and the above process of direct correlation analysis, indirect correlation analysis, and weighted combination is repeated to calculate the direct and indirect correlation between queue length and traffic flow, pedestrian waiting time, and green light utilization rate, thereby obtaining the state correlation between queue length and each other control parameter. Then, pedestrian waiting time and green light utilization rate are selected as selected control parameters in turn, and the same analysis is performed to finally complete the calculation of the state correlation between all control parameters.

[0066] Throughout the analysis, the timeliness and accuracy of the data must be carefully considered. Traffic conditions at intersections are dynamic, and traffic flow characteristics can vary significantly across different time periods. Therefore, when conducting correlation analysis, it is necessary to calculate the state correlation degree separately for different time periods (e.g., morning peak, off-peak, evening peak) to reflect the correlation characteristics between control parameters at different times. Simultaneously, outlier detection and processing are required for the collected traffic characteristic data to prevent abnormal data from interfering with the correlation degree calculation results. For example, if traffic flow data at a certain point in time shows significant anomalies (e.g., significantly higher than historical data for the same period), it is necessary to check whether the data is caused by detector malfunction or other abnormalities, and correct or remove the abnormal data.

[0067] By performing traffic state correlation analysis on the traffic characteristic values ​​of all control parameters, the state correlation degree between each control parameter and other control parameters in the traffic structure is obtained, enabling a deeper understanding of the interaction relationships between various traffic parameters at the intersection. This state correlation degree reflects the inherent connection between control parameters in the traffic structure, providing an important basis for subsequent optimization of traffic monitoring data based on control reliability and the generation of reasonable signal control schemes. It allows the system to consider the impact of each control parameter from the perspective of the overall traffic structure, rather than viewing each parameter in isolation, thus laying the foundation for more precise traffic signal control.

[0068] Example 3: In this example, a pre-trained signal control evaluation model is used to perform quality mapping evaluation on the traffic monitoring data corresponding to each control parameter to obtain a quality score. The specific implementation process is as follows. For each control parameter, the statistical and distribution characteristics of its corresponding traffic monitoring data must first be extracted. These characteristics are important bases for evaluating data quality. The statistical characteristics specifically include the maximum value, minimum value, and median. Taking traffic flow monitoring data as an example, within a certain set time period, such as 1 hour, traffic flow data at each time point within that time period is collected. The largest traffic flow value is taken as the maximum value, and the smallest traffic flow value is taken as the minimum value. The value in the middle position after sorting all the data from smallest to largest is the median. These three statistical characteristics can reflect the central tendency and dispersion of the data from different perspectives. The maximum value reflects the peak traffic flow within that time period, the minimum value reflects the lowest level of traffic flow, and the median is not affected by extreme values ​​and better represents the general level of the data.

[0069] In addition to statistical characteristics, distribution characteristics also need to be extracted. Distribution characteristics are mainly obtained by analyzing the probability distribution of traffic monitoring data, such as determining whether the data conforms to a normal distribution, a uniform distribution, or another type of distribution. Taking queue length monitoring data as an example, queue length data over a certain time period are compiled, the frequency of data occurrences within different queue length intervals is statistically analyzed, and a frequency distribution histogram is plotted. By observing the shape of the histogram and comparing it with the standard distribution curve, the distribution characteristics can be determined. If the data distribution is relatively concentrated, it indicates good data stability; if the distribution is relatively dispersed, it indicates greater data fluctuation.

[0070] After extracting the statistical and distribution features, these features are used as input parameters for the signal control evaluation model. The signal control evaluation model is pre-trained using a large amount of historical traffic data and corresponding signal control performance data. The model structure can employ common machine learning architectures such as neural networks. During training, the model learns how to establish a mapping relationship between the input statistical and distribution features and the quality of traffic monitoring data, thereby acquiring the ability to evaluate the quality of new traffic monitoring data.

[0071] Taking traffic monitoring data corresponding to pedestrian waiting time control parameters as an example, assuming the extracted statistical features are a maximum value of 120 seconds, a minimum value of 20 seconds, and a median of 60 seconds, and the distribution features show that the data approximately follows a normal distribution with a mean of 65 seconds and a standard deviation of 20 seconds, these features are input into a pre-trained signal control evaluation model. The model processes and analyzes these features, and through internal neural network calculations and weight adjustments, ultimately outputs a score representing the quality of the pedestrian waiting time monitoring data. This score can be a value within a specific range, such as between 0 and 100. The higher the value, the higher the data quality, indicating that the monitoring data more accurately reflects the actual pedestrian waiting time situation.

[0072] For traffic monitoring data corresponding to the green light utilization rate control parameter, it is also necessary to extract its statistical and distribution characteristics. Assuming green light utilization rate data is collected over multiple signal cycles, with a maximum value of 90%, a minimum value of 30%, and a median of 65%, the distribution characteristics show a skewed distribution. After inputting these characteristics into the model, the model will evaluate the quality of the green light utilization rate monitoring data based on the rules and patterns obtained during training, and output a corresponding quality score.

[0073] In practical applications, different control parameters may have different feature importance, and the model will automatically learn the degree of influence of each feature on data quality during training. For example, for traffic flow monitoring data, the maximum and minimum values ​​may have a greater impact on data quality because they can directly reflect extreme traffic flow conditions; while for pedestrian waiting time monitoring data, the median and distribution characteristics may better reflect the reliability of the data.

[0074] When extracting features, it is important to pay attention to the selection of the time window. Different time windows will affect the calculation results of the features. For example, features within a short time window (such as 5 minutes) may better reflect real-time changes in traffic conditions, but are easily affected by random factors; features within a long time window (such as 30 minutes) better reflect the overall trend of traffic conditions, but may mask some short-term fluctuations. Therefore, it is necessary to reasonably select the length of the time window based on the actual application scenario and requirements.

[0075] Furthermore, handling outlier data is also crucial. Before feature extraction, traffic monitoring data needs to be preprocessed, including removing noise and filling in missing values. For example, if traffic flow data at a certain point in time is significantly higher than historical data for the same period, and this is confirmed to be due to a detector malfunction, the data needs to be corrected or removed to avoid affecting feature extraction and model evaluation results.

[0076] By using a signal control evaluation model to assess the quality of traffic monitoring data corresponding to each control parameter, the resulting quality scores objectively reflect the reliability and accuracy of the data. These scores serve as crucial criteria for determining control reliability, enabling the system to adjust its dependence on each control parameter based on data quality, thus making the generation of signal control schemes more scientific and rational. Data with high quality scores means it more accurately reflects traffic conditions and will be given higher weight in control decisions; while data with low quality scores needs to be optimized in conjunction with the correlation of other control parameters to improve its reliability.

[0077] Throughout the process, pre-training of the signal control assessment model is a crucial step, as the quality and quantity of training data directly impact the model's assessment accuracy. Therefore, during the model training phase, it is necessary to collect a large amount of historical data covering different traffic scenarios and time periods, and to rigorously label and preprocess the data to ensure the model can learn comprehensive and accurate mapping relationships. Simultaneously, the model also needs to be regularly updated and optimized to adapt to changes in the traffic environment and improvements in detector performance, ensuring the continued effectiveness of its assessment capabilities.

[0078] Example 4: In this example, the control reliability of the traffic monitoring data corresponding to each control parameter is determined by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure. The specific implementation process is as follows. First, a control parameter needs to be selected as the selected control parameter, taking queue length as an example. In the actual operation of traffic intersections, the quality score of queue length monitoring data may be affected by various factors, such as the installation location of the detector and changes in ambient light. The state correlation between queue length monitoring data and other control parameters (such as traffic flow, pedestrian waiting time, and green light utilization rate) reflects the degree of mutual influence of this parameter in the overall traffic structure. Therefore, it is necessary to combine these two aspects of information to determine its control reliability.

[0079] After selecting queue length as the chosen control parameter, the reliability adjustment coefficient of the corresponding traffic monitoring data is determined based on the state correlation between this control parameter and other control parameters in the traffic structure. The state correlation is obtained based on the analysis method described in Example 2. For example, the state correlation between queue length and traffic flow is calculated by weighting direct and indirect correlation. Assume that the analysis shows a state correlation of 0.7 between queue length and traffic flow, 0.3 between queue length and pedestrian waiting time, and 0.5 between queue length and green light utilization rate. These values ​​reflect the close correlation between queue length and other parameters. The higher the state correlation, the more relevant the change in queue length is to the change in this parameter, and the more reliable the monitoring data can be verified and adjusted using data from other parameters.

[0080] Determining the credible adjustment coefficient requires considering the state correlation between the selected control parameter and all other control parameters. A common approach is to comprehensively calculate the state correlation between the selected control parameter and the other control parameters. For example, assuming there are three other control parameters—traffic flow, pedestrian waiting time, and green light utilization rate—with queue length having state correlations of 0.7, 0.3, and 0.5 with them respectively, the credible adjustment coefficient can be calculated by summing or weighted summing. If a simple summation method is used, the credible adjustment coefficient is 0.7 + 0.3 + 0.5 = 1.5. If a weighted summation method is used, weights need to be assigned according to the importance of each parameter in the traffic structure. For example, if the weight of traffic flow is 0.5, the weight of pedestrian waiting time is 0.2, and the weight of green light utilization rate is 0.3, then the credible adjustment coefficient is 0.7 × 0.5 + 0.3 × 0.2 + 0.5 × 0.3 = 0.35 + 0.06 + 0.15 = 0.56. The specific calculation method can be determined based on the analysis of actual traffic scenarios and historical data. Its core purpose is to measure the reliability of the monitoring data of the selected control parameters through state correlation.

[0081] After obtaining the confidence adjustment coefficient, the quality score of the traffic monitoring data corresponding to the selected control parameter is adjusted using this coefficient to obtain the control confidence level. The quality score is output by the signal control evaluation model described in Example 3. For example, the quality score of queue length monitoring data is 70 points. If the confidence adjustment coefficient is 1.5, the control confidence level is 70 × 1.5 = 105 points (assuming the score range can exceed the original range); if the confidence adjustment coefficient is 0.56, the control confidence level is 70 × 0.56 = 39.2 points. This adjustment reflects the corrective effect of state correlation on data confidence: when the state correlation between the selected control parameter and other parameters is high, it indicates that its monitoring data has strong consistency with the data of other parameters, and the confidence adjustment coefficient is greater than 1. The adjusted control confidence level will be higher than the quality score, indicating that the data is more reliable in the overall traffic structure; when the state correlation is low, the confidence adjustment coefficient is less than 1, and the adjusted control confidence level will be lower than the quality score, indicating that the data may contradict the data of other parameters and needs further verification or optimization.

[0082] After calculating the control confidence level for the selected control parameter (queue length), the same method is used to determine the control confidence level for the remaining control parameters corresponding to the traffic monitoring data. For example, if traffic flow is selected as a new control parameter, the state correlation between traffic flow and queue length, pedestrian waiting time, and green light utilization rate is first obtained, assuming they are 0.8, 0.4, and 0.6 respectively. Then, the confidence adjustment coefficient is calculated based on these state correlations. If a simple summation method is used, the coefficient is 0.8 + 0.4 + 0.6 = 1.8; if the quality score of the traffic flow monitoring data is 80 points, then the control confidence level is 80 × 1.8 = 144 points. Next, pedestrian waiting time is selected as the selected control parameter. Assuming its state correlation with traffic flow, queue length, and green light utilization rate are 0.3, 0.2, and 0.4 respectively, a weighted summation (with the same weights as above) is used to calculate the confidence adjustment coefficient: 0.3×0.5 + 0.2×0.2 + 0.4×0.3 = 0.15 + 0.04 + 0.12 = 0.31. If the quality score is 60, the control confidence level is 60×0.31 = 18.6. Finally, green light utilization rate is selected as the selected control parameter. Assuming its state correlation with traffic flow, queue length, and pedestrian waiting time are 0.5, 0.6, and 0.3 respectively, a simple summation yields a confidence adjustment coefficient of 1.4. If the quality score is 75, the control confidence level is 75×1.4 = 105.

[0083] In practical applications, it is important to note that the calculation method and weighting of the reliability adjustment coefficient should be adjusted based on the characteristics of different traffic intersections and historical data. For example, intersections in urban centers have higher traffic volume, and the correlation between traffic volume and other parameters may be stronger; therefore, the weight of traffic volume can be appropriately increased when setting weights. Conversely, pedestrian traffic at suburban intersections may be lower, and the correlation between pedestrian waiting time and other parameters may be relatively weaker; therefore, its weight can be reduced accordingly. Furthermore, it is necessary to consider changes in traffic characteristics at different times of day. For instance, the correlation between traffic volume and queue length may be stronger during morning rush hour than during off-peak hours; therefore, the weights and calculation methods can be dynamically adjusted according to the time period.

[0084] Meanwhile, the scoring range for control confidence needs to be set reasonably according to actual needs. If the scoring range is set to 0 to 100 points, when the adjusted control confidence exceeds 100 points, it can be limited to 100 points; when it is below 0 points, it can be limited to 0 points. This limitation can ensure that the control confidence value is within a reasonable range, which is convenient for subsequent comparison with the preset confidence threshold.

[0085] By combining quality scores and state correlation to determine control reliability, the reliability of control parameter monitoring data can be comprehensively evaluated from two dimensions: the quality of the data itself and its correlation with the traffic structure. This allows the system to consider not only the quality of individual data points when processing traffic data, but also the rationality of the data from the perspective of the overall traffic structure, avoiding deviations in signal control schemes due to the anomalies or unreliability of individual data. For example, when a control parameter has a low quality score but a high state correlation with other parameters, its control reliability may not be too low, indicating that although the data itself has some quality issues, it is consistent with the trend of the overall traffic structure and still has some reference value. Conversely, when the quality score is low and the state correlation is also low, the control reliability will decrease significantly, indicating that the data is both unreliable and contradictory to other parameters, requiring optimization.

[0086] Throughout the process, the accuracy of state correlation and quality scoring is crucial to ensuring the accuracy of control confidence calculation. Therefore, it is necessary to regularly update and optimize the state correlation analysis model and signal control evaluation model to adapt to changes in the traffic environment and improvements in detector performance, ensuring that the control confidence calculation always accurately reflects the actual situation.

[0087] Example 5: In this example, when the control confidence level of the traffic monitoring data corresponding to the control parameter is less than the preset confidence threshold, the traffic monitoring data corresponding to the control parameter needs to be optimized based on the state correlation of the traffic structure between the control parameter and other control parameters. A signal control scheme for the intersection is then generated based on the optimized traffic monitoring data. The preset confidence threshold is set based on actual traffic control needs and historical data experience; for example, it is set to 60 points. If the control confidence level of a certain control parameter is lower than 60 points, the optimization mechanism is triggered.

[0088] Taking traffic flow control parameters as an example, assuming the control reliability of the traffic flow monitoring data calculated using the method in Example 4 is 50 points, which is less than the preset threshold of 60 points, the data needs to be optimized. First, the state correlation between traffic flow and other control parameters (queue length, pedestrian waiting time, and green light utilization rate) in the traffic structure is obtained. Assuming the analysis method in Example 2 yields a state correlation of 0.8 between traffic flow and queue length, 0.3 with pedestrian waiting time, and 0.6 with green light utilization rate, these correlation values ​​reflect the closeness of the relationship between traffic flow and other parameters. A higher correlation indicates greater reference value of the other parameters for the traffic flow data.

[0089] During the optimization process, the correlation between traffic flow and other control parameters is utilized to obtain relevant information from traffic monitoring data of other parameters to correct the traffic flow data. Considering the control reliability of each correlated parameter and its correlation with traffic flow, the monitoring data of the correlated parameters are weighted to generate optimized traffic flow data. For example, the control reliability of queue length is 70 points, pedestrian waiting time is 65 points, and green light utilization rate is 80 points. Based on the state correlation and control reliability, the contribution weight of each correlated parameter to the traffic flow data is calculated. One possible weight calculation method is to multiply the state correlation by the control reliability of the corresponding parameter and then normalize the result. The weight of queue length is (0.8 × 70) = 56, the weight of pedestrian waiting time is (0.3 × 65) = 19.5, and the weight of green light utilization rate is (0.6 × 80) = 48, for a total weight of 56 + 19.5 + 48 = 123.5. The normalized weights for queue length are 56 / 123.5≈0.453, for pedestrian waiting time are 19.5 / 123.5≈0.158, and for green light utilization are 48 / 123.5≈0.389.

[0090] Assuming the standardized value of queue length monitoring data is 0.6, the standardized value of pedestrian waiting time monitoring data is 0.4, and the standardized value of green light utilization rate monitoring data is 0.7, the optimized standardized value of traffic flow is calculated based on normalized weights as follows: 0.6×0.453+0.4×0.158+0.7×0.389≈0.272+0.063+0.272≈0.607. This value is the optimized standardized value of the traffic flow monitoring data. Compared to the original data (assuming the original standardized value is 0.5), it has been corrected, and the correction process considers its correlation with other parameters and the reliability of other parameter data, making the optimized data more consistent with the overall traffic structure.

[0091] After optimization, a signal control scheme for the intersection is generated based on the optimized traffic monitoring data. Generating the signal control scheme requires comprehensive consideration of the optimized data for various control parameters to rationally allocate green light time for each direction. For example, when generating the scheme, firstly, the optimized traffic flow data is analyzed to determine the traffic flow volume of each approach lane; then, the optimized queue length data is combined to determine the congestion level of each approach lane; next, the optimized pedestrian waiting time data is considered to ensure reasonable pedestrian crossing time; and finally, green light utilization rate data is referenced to evaluate the efficiency of the current green light time usage.

[0092] Assume that after optimization, the standardized values ​​for traffic flow at the east entrance of a certain intersection are: 0.65, queue length: 0.7, pedestrian waiting time: 0.5, and green light utilization rate: 0.6. For the west entrance, the standardized values ​​are: 0.4, queue length: 0.3, pedestrian waiting time: 0.4, and green light utilization rate: 0.7. For the south entrance, the standardized values ​​are: 0.5, queue length: 0.6, pedestrian waiting time: 0.6, and green light utilization rate: 0.5. For the north entrance, the standardized values ​​are: 0.3, queue length: 0.2, pedestrian waiting time: 0.3, and green light utilization rate: 0.8.

[0093] Based on this optimized data, when generating the signal control scheme, for the eastbound approach, due to the relatively large traffic volume and queue length, a longer green light time is needed to alleviate traffic congestion; the southbound approach also has a large queue length, so the green light time needs to be slightly shorter; the westbound and northbound approach have relatively smaller traffic volumes and queue lengths, so the green light time can be appropriately shortened. Simultaneously, considering pedestrian waiting time, reasonable pedestrian crossing times are inserted into the green light time allocation for vehicles in each direction to ensure that pedestrian waiting times do not exceed a reasonable range. Furthermore, referring to green light utilization data, for approach lanes with low green light utilization, the green light time is adjusted to avoid wasting green light time; for approach lanes with high green light utilization, if the traffic volume and queue length are still large, the green light time can be appropriately increased to improve traffic efficiency.

[0094] In the specific process of generating a signal control scheme, traditional signal control algorithms, such as timing control algorithms, can be used to calculate the green light time for each phase based on optimized traffic monitoring data. For example, the Webster algorithm can be used to calculate the optimal cycle length and green light time for each phase based on the traffic flow and saturation flow of each approach lane. Adaptive control algorithms can also be used to dynamically adjust signal timing based on real-time optimized data, making the signal control scheme more adaptable to changes in traffic conditions.

[0095] Throughout the optimization and scheme generation process, it is crucial to consider the interplay between different control parameters. For instance, increasing the green light time in one direction may reduce the green light time in other directions, thereby affecting traffic flow and queue length in those directions. Therefore, a comprehensive balance must be struck when generating the scheme to ensure that traffic demand in all directions is reasonably met, avoiding situations where one direction experiences smooth traffic while another suffers severe congestion.

[0096] Furthermore, the dynamic characteristics of traffic flow must be considered. Traffic conditions change in real time, and the optimized traffic monitoring data reflects the traffic conditions at the current moment or within a short period. Therefore, the generated signal control scheme needs to have a certain degree of real-time performance and adaptability. An update cycle can be set, such as regenerating the signal control scheme every 5 minutes based on the latest optimized data, to adapt to changes in traffic flow.

[0097] In this way, when the reliability of traffic monitoring data corresponding to control parameters is insufficient, optimization is performed by leveraging their correlation with other control parameters. This optimizes the data to more accurately reflect the traffic conditions at the intersection. The resulting signal control scheme better adapts to changes in traffic flow, rationally allocates green light times for each direction, improves intersection efficiency, and ensures pedestrian safety and convenience. The entire process fully utilizes the relationships between control parameters, correcting unreliable data from the perspective of the overall traffic structure. This avoids the adverse effects of unreliable single data points on the signal control scheme, making signal control more scientific and rational.

[0098] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various equivalent modifications or substitutions within the technical scope disclosed in this application, and these modifications or substitutions should all be covered within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A method for intelligent control of traffic lights, characterized in that, Includes the following steps: Traffic status data at intersections is collected by different types of detectors, and traffic monitoring data corresponding to various control parameters are obtained. Traffic characteristic values ​​of corresponding control parameters are extracted from various traffic monitoring data, and then traffic state correlation analysis is performed on all traffic characteristic values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure. Based on the pre-trained signal control evaluation model, the quality mapping evaluation of traffic monitoring data corresponding to each control parameter is performed to obtain the quality score of traffic monitoring data corresponding to each control parameter. Then, the control reliability of traffic monitoring data corresponding to each control parameter is determined by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure. When the control confidence of the traffic monitoring data corresponding to the control parameter is less than the preset confidence threshold, the traffic monitoring data corresponding to the control parameter is optimized according to the state correlation of the traffic structure between the control parameter and other control parameters, and the signal control scheme of the intersection is generated based on the optimized traffic monitoring data. The step of performing traffic state correlation analysis on all traffic characteristic values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure specifically includes: Select one control parameter as the selected control parameter; A direct correlation analysis is performed on the traffic characteristic values ​​between the selected control parameters and other control parameters to obtain the direct correlation degree of traffic. An indirect correlation analysis of traffic structure is performed on the traffic characteristic values ​​between the selected control parameters and other control parameters to obtain the degree of indirect correlation of traffic. The state correlation between the selected control parameter and other control parameters in the traffic structure is determined by the direct correlation degree and the indirect correlation degree, and then the state correlation between the remaining control parameters and other control parameters in the traffic structure is determined. The determination of the control reliability of the traffic monitoring data corresponding to each control parameter, based on the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure, specifically includes: Select one control parameter as the selected control parameter; The reliability adjustment coefficient of the traffic monitoring data corresponding to the selected control parameter is determined based on the state correlation between the selected control parameter and other control parameters in the traffic structure. The quality score of the traffic monitoring data corresponding to the selected control parameter is adjusted by the confidence adjustment coefficient to obtain the control confidence of the traffic monitoring data corresponding to the selected control parameter. Continue to determine the control reliability of the traffic monitoring data corresponding to the remaining control parameters.

2. The intelligent traffic light control method as described in claim 1, characterized in that, The extraction of traffic characteristic values ​​corresponding to control parameters from various traffic monitoring data specifically includes: The traffic monitoring data are standardized to obtain the standardized traffic monitoring data. Traffic characteristic values ​​corresponding to control parameters are extracted from various traffic monitoring data after standardization.

3. The intelligent traffic light control method as described in claim 1, characterized in that, Based on a pre-trained signal control evaluation model, the quality of traffic monitoring data corresponding to each control parameter is mapped and evaluated to obtain a quality score for the traffic monitoring data corresponding to each control parameter. Specifically, this includes: For each control parameter, extract the statistical and distribution characteristics of the traffic monitoring data corresponding to the control parameter; The statistical characteristics and the distribution characteristics are used as input parameters for the signal control evaluation model; The quality score of traffic monitoring data corresponding to the control parameters is obtained by mapping the signal control evaluation model, and then the quality score of traffic monitoring data corresponding to each control parameter is obtained.

4. The intelligent traffic light control method as described in claim 3, characterized in that, The statistical characteristics specifically include the maximum value, minimum value, and median.

5. The intelligent traffic light control method as described in claim 1, characterized in that, The control parameters specifically include traffic flow, queue length, pedestrian waiting time, and green light utilization rate.

6. A traffic signal intelligent control system, used to implement the traffic signal intelligent control method as described in any one of claims 1 to 5, characterized in that, include: The data acquisition module is used to collect traffic status data at intersections through different types of detectors, and then obtain traffic monitoring data corresponding to various control parameters. The processing module is used to extract the traffic feature values ​​of the corresponding control parameters from various traffic monitoring data, and then perform traffic state correlation analysis on all traffic feature values ​​to obtain the state correlation degree between each control parameter and other control parameters in the traffic structure. The processing module is also used to perform quality mapping evaluation on the traffic monitoring data corresponding to each control parameter based on the pre-trained signal control evaluation model, to obtain the quality score of the traffic monitoring data corresponding to each control parameter, and then determine the control reliability of the traffic monitoring data corresponding to each control parameter by the quality score of all traffic monitoring data and the state correlation between each control parameter and other control parameters in the traffic structure. The execution module is used to optimize the traffic monitoring data corresponding to the control parameters based on the state correlation of the traffic structure between the control parameters and other control parameters when the control confidence of the traffic monitoring data corresponding to the control parameters is less than the preset confidence threshold, and generate the signal control scheme of the intersection based on the optimized traffic monitoring data.

7. An electronic device comprising a memory and a processor, the memory storing code, characterized in that, The processor is configured to acquire the code and execute the intelligent traffic light control method as described in any one of claims 1 to 5.

8. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the intelligent traffic light control method as described in any one of claims 1 to 5.

Images