Traffic flow prediction fused adaptive control method and system for traffic lights

Abstract

The application discloses a traffic flow prediction fused traffic light adaptive control method and system, and relates to the technical field of intelligent traffic, comprising: obtaining real-time traffic flow data of a target intersection; obtaining a current traffic light control strategy of the target intersection; obtaining a traffic flow prediction result; performing change degree evaluation according to the real-time traffic flow data and the traffic flow prediction result, obtaining a traffic flow change coefficient, and whether it is greater than or equal to a traffic flow change threshold; if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, generating a traffic light adjustment instruction, adjusting the current traffic light control strategy according to the traffic flow prediction result, generating a future traffic light control strategy, and performing traffic light adaptive control of the target intersection. The application solves the technical problems of inaccurate traffic flow prediction and inflexible traffic light control strategy in the prior art, and achieves the technical effects of improving traffic efficiency, reducing traffic congestion and optimizing traffic light control.

Description

Technical Field

[0001] This invention relates to the field of intelligent transportation technology, specifically to a traffic light adaptive control method and system that integrates traffic flow prediction. Background Technology

[0002] With the accelerating pace of urbanization and the rapid increase in the number of motor vehicles, traffic flow is increasing daily, leading to increasingly severe traffic congestion. Especially during peak hours and at key intersections, traditional traffic light control methods are no longer effective in handling complex and ever-changing traffic demands, thus affecting traffic efficiency and safety. Although some traffic light control strategies based on time cycles or flow perception have emerged, these traditional methods generally suffer from low control precision, slow response speed, and poor adaptability, making it difficult to effectively alleviate traffic congestion. Summary of the Invention

[0003] This application provides a traffic light adaptive control method and system that integrates traffic flow prediction, in order to address the technical problems of inaccurate traffic flow prediction and inflexible traffic light control strategies in existing technologies.

[0004] In view of the above problems, this application provides a traffic light adaptive control method and system that integrates traffic flow prediction.

[0005] The first aspect of this application provides a traffic light adaptive control method that integrates traffic flow prediction, the method comprising:

[0006] Based on the intersection traffic perception module, real-time traffic flow data of the target intersection is obtained; the current traffic light control strategy of the target intersection is obtained; based on the real-time traffic flow data, traffic flow of the target intersection is predicted according to the future time zone, and a traffic flow prediction result is obtained; the degree of change is evaluated based on the real-time traffic flow data and the traffic flow prediction result, and a traffic flow change coefficient is obtained; it is determined whether the traffic flow change coefficient is greater than or equal to the traffic flow change threshold; if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, a traffic light adjustment command is generated; based on the traffic light adjustment command, the current traffic light control strategy is adjusted according to the traffic flow prediction result to generate a future traffic light control strategy, and adaptive traffic light control of the target intersection is executed according to the future traffic light control strategy.

[0007] A second aspect of this application provides a traffic light adaptive control system that integrates traffic flow prediction, the system comprising:

[0008] The system includes the following modules: a real-time traffic flow data acquisition module for obtaining real-time traffic flow data of a target intersection based on the intersection traffic perception module; a traffic light control strategy acquisition module for obtaining the current traffic light control strategy of the target intersection; a traffic flow prediction module for predicting traffic flow at the target intersection based on the real-time traffic flow data and the future time zone, and obtaining a traffic flow prediction result; a variability evaluation module for evaluating variability based on the real-time traffic flow data and the traffic flow prediction result, and obtaining a traffic flow variability coefficient; a judgment module for determining whether the traffic flow variability coefficient is greater than or equal to a traffic flow variability threshold; an adjustment instruction generation module for generating a traffic light adjustment instruction if the traffic flow variability coefficient is greater than or equal to the traffic flow variability threshold; and an adaptive control module for adjusting the current traffic light control strategy based on the traffic light adjustment instruction and the traffic flow prediction result, generating a future traffic light control strategy, and executing adaptive traffic light control at the target intersection according to the future traffic light control strategy.

[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0010] This application obtains real-time traffic flow data of a target intersection based on an intersection traffic perception module; obtains the current traffic light control strategy for the target intersection; predicts traffic flow for the target intersection based on the real-time traffic flow data and future time zones, obtaining a traffic flow prediction result; evaluates the degree of change based on the real-time traffic flow data and the traffic flow prediction result, obtaining a traffic flow change coefficient; determines whether the traffic flow change coefficient is greater than or equal to a traffic flow change threshold; if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, generates a traffic light adjustment command; adjusts the current traffic light control strategy based on the traffic light adjustment command and the traffic flow prediction result, generating a future traffic light control strategy, and executes adaptive traffic light control for the target intersection according to the future traffic light control strategy. This invention solves the technical problems of inaccurate traffic flow prediction and inflexible traffic light control strategies in existing technologies. By combining real-time traffic flow data and traffic flow prediction results, and adjusting the traffic light control strategy based on the prediction results, it achieves the technical effects of improving traffic efficiency, reducing traffic congestion, and optimizing traffic light control. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1 A schematic flowchart of the traffic light adaptive control method integrating traffic flow prediction provided in an embodiment of this application;

[0013] Figure 2 A schematic diagram of the structure of the traffic light adaptive control system that integrates traffic flow prediction provided in the embodiments of this application.

[0014] Figure labeling: 11 Real-time traffic flow data acquisition module, 12 Traffic light control strategy acquisition module, 13 Traffic flow prediction module, 14 Change degree evaluation module, 15 Judgment module, 16 Adjustment instruction generation module, 17 Adaptive control module. Detailed Implementation

[0015] This application provides a traffic light adaptive control method and system that integrates traffic flow prediction. It addresses the technical problems of inaccurate traffic flow prediction and inflexible traffic light control strategies in existing technologies. By combining real-time traffic flow data with traffic flow prediction results and adjusting the traffic light control strategy based on the prediction results, it achieves the technical effects of improving traffic efficiency, reducing traffic congestion, and optimizing traffic light control.

[0016] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this application, and not all of them. All other embodiments obtained by those skilled in the art based on the embodiments of this application without creative effort are within the scope of protection of this application.

[0017] It should be noted that any variation of the terms "comprising" and "having" is intended to cover non-exclusive inclusion, for example, a process, method, system, product, or server that includes a series of steps or units is not necessarily limited to those steps or units that are explicitly listed, but may include other steps or modules that are not explicitly listed or that are inherent to such processes, methods, products, or devices.

[0018] Example 1, as Figure 1 As shown, this application provides a traffic light adaptive control method that integrates traffic flow prediction, the method comprising:

[0019] Step S100: Obtain real-time traffic flow data of the target intersection based on the intersection traffic perception module.

[0020] In this embodiment, traffic data is collected by a sensing module to form a real-time traffic sensing dataset. Then, the dataset is filtered and denoised to remove environmental interference and generate a denoised traffic sensing dataset. Next, outlier detection is performed to identify abnormal data and obtain the outlier detection results of the sensing data. Finally, the data is corrected based on the detection results to improve the accuracy of the data and ultimately obtain reliable real-time traffic flow data.

[0021] Furthermore, in the method provided in the application embodiments, obtaining real-time traffic flow data of the target intersection based on the intersection traffic perception module further includes:

[0022] Based on the intersection traffic perception module, the real-time traffic perception dataset of the target intersection is obtained; the real-time traffic perception dataset is filtered and denoised to generate a traffic perception denoised dataset; outlier detection is performed on the traffic perception denoised dataset to obtain perception data outlier detection results; the traffic perception denoised dataset is corrected based on the perception data outlier detection results to obtain the real-time traffic flow data.

[0023] In this embodiment, during the data acquisition phase, the intersection traffic perception module is responsible for real-time monitoring of traffic conditions. This module consists of video surveillance cameras, millimeter-wave radar, geomagnetic induction devices, infrared sensors, microwave detectors, etc. Different types of sensing devices can respectively sense information such as vehicle passage, vehicle speed, lane occupancy, and pedestrian detection. The data collected by these devices are then integrated to form a real-time traffic perception dataset.

[0024] Next, in the data preprocessing stage, the data is filtered and denoised to reduce sensor errors and environmental noise. For example, data collected by video cameras may be affected by changes in lighting, while geomagnetic sensors may exhibit short-term abnormal readings due to external electromagnetic interference. To remove this noise, methods such as Kalman filtering, median filtering, or wavelet transform are used to smooth the real-time traffic perception dataset, thereby generating a denoised traffic perception dataset.

[0025] Then, in the anomaly detection phase, outlier detection is performed on the denoised data to identify unreasonable abnormal data. For example, if the vehicle speed detected by the radar exceeds the normal range for a certain period of time, or if the geomagnetic sensor fluctuates frequently in a short period of time, these may be abnormal data caused by sensor malfunction or environmental interference. By using statistical methods (such as Z-Score), outliers are identified and marked, generating outlier detection results for the perception data.

[0026] Finally, in the data correction phase, unreasonable data is corrected based on the outlier detection results of the perception data. For example, if the traffic flow suddenly drops to zero during a certain period, but the cameras and other sensors still detect the presence of vehicles, time series interpolation (such as linear interpolation and spline interpolation) is used to correct the data, thereby obtaining real-time traffic flow data that better reflects the actual situation.

[0027] Step S200: Obtain the current traffic light control strategy for the target intersection.

[0028] In this embodiment of the application, the current traffic light control strategy of the target intersection is obtained by acquiring signal control data from the traffic signal controller or intelligent traffic management platform, including information such as signal cycle, phase duration, phase sequence arrangement and control mode, and then parsing and standardizing the data to form the current traffic light control strategy.

[0029] Step S300: Based on the real-time traffic flow data, predict the traffic flow of the target intersection according to the future time zone to obtain the traffic flow prediction result.

[0030] In this embodiment, traffic flow prediction is performed based on real-time traffic flow data. First, a set of traffic flow records for the target intersection is acquired, and the data is processed in a time-series manner to form a traffic flow analysis set. Then, this dataset is used to perform supervised learning on a Long Short-Term Memory (LSTM) neural network to train and optimize the traffic flow prediction model. Finally, the real-time traffic flow data is input into the trained model, and the traffic flow prediction result is output.

[0031] Furthermore, in the method provided in the application embodiments, based on the real-time traffic flow data, traffic flow prediction is performed on the target intersection according to the future time zone to obtain a traffic flow prediction result, which further includes:

[0032] Obtain a traffic flow record set for the target intersection; perform time-series processing on the traffic flow record set to obtain a traffic flow summary set; perform supervised learning on a long short-term memory neural network based on the traffic flow summary set to obtain a traffic flow prediction model; input the real-time traffic flow data into the traffic flow prediction model and output the traffic flow prediction result.

[0033] In this embodiment of the application, the traffic flow record set of the target intersection is first obtained from the historical traffic database. This dataset includes information such as traffic flow, average vehicle speed, lane occupancy, and traffic light status of the target intersection over a historical period.

[0034] Next, the traffic flow record set is processed for time series analysis. Since traffic flow data has time-series characteristics, the original data may have issues such as uneven timestamps and inconsistent data granularity. Therefore, by performing time alignment and using the sliding window method, the data is converted into a format with fixed time intervals to obtain the traffic flow summary set.

[0035] Subsequently, supervised learning of the Long Short-Term Memory (LSTM) neural network was performed using a traffic flow analysis dataset. During training, the traffic flow analysis dataset was divided into training, testing, and validation sets. The Adam optimizer was used for parameter updates, and the prediction performance was evaluated based on mean squared error (MSE) or mean absolute error (MAE). Simultaneously, hyperparameter tuning was performed, optimizing the number of layers, neurons, and learning rate of the LSTM to improve the model's generalization ability. Through this process, a traffic flow prediction model was obtained.

[0036] Finally, the real-time traffic flow data is input into the trained traffic flow prediction model to generate traffic flow prediction results.

[0037] Furthermore, in the method provided in the application embodiments, the method of performing supervised learning on a long short-term memory neural network based on the traffic flow analysis set to obtain a traffic flow prediction model further includes:

[0038] The traffic flow analysis set is divided into a traffic flow training set, a traffic flow test set, and a traffic flow validation set. A long short-term memory neural network is trained under supervision using the traffic flow training set to obtain a traffic flow prediction network. The traffic flow prediction network is tested using the traffic flow test set to obtain the traffic flow prediction accuracy. Based on the traffic flow prediction accuracy, the hyperparameters of the traffic flow prediction network are tuned using the traffic flow validation set to generate the traffic flow prediction model.

[0039] In this embodiment, the traffic flow data set is first divided to obtain a training set, a test set, and a validation set. During the division, the traffic flow data set is divided in a 7:2:1 ratio, where the training set accounts for 7% (for model learning), the test set accounts for 2% (for evaluating training effectiveness), and the validation set accounts for 1% (for hyperparameter optimization). The division method can employ a time-sliding window, i.e., inputting data in batches according to chronological order, or random splitting and allocation. This process yields the traffic flow training set, the traffic flow test set, and the traffic flow validation set.

[0040] Next, the LSTM neural network is trained under supervised conditions using a traffic flow training set to construct a traffic flow prediction network. The training process employs the backpropagation algorithm, which calculates the prediction error and uses gradient descent to continuously adjust the network's weight parameters, gradually converging the model's loss function. Specifically, the error between the predicted and actual values ​​is first calculated, and then gradient descent is used to update the connection weights of the LSTM network to minimize the error, thereby improving the model's learning ability. After training, the final traffic flow prediction network is obtained.

[0041] The trained LSTM model was then tested using a traffic flow test set to evaluate its prediction accuracy. During testing, mean squared error (MSE) was used as the evaluation metric. This metric calculates the mean of the squared errors between the predicted and actual values, measuring the overall error level of the model. A smaller MSE value indicates lower prediction error and better prediction performance. Through this testing process, the traffic flow prediction accuracy was ultimately obtained.

[0042] Finally, based on the traffic flow prediction accuracy, grid search is used to tune the hyperparameters of the LSTM model to obtain the optimal traffic flow prediction model. Grid search is a method that systematically traverses combinations of hyperparameters. By pre-setting multiple possible hyperparameter combinations, the model is trained one by one, and the MSE on the validation set is calculated. Finally, the hyperparameter combination that minimizes the MSE is selected. For example, grid search can be used to adjust parameters such as the number of LSTM layers, the number of hidden units, the learning rate, and the time window length to optimize model performance and improve prediction accuracy. After hyperparameter tuning, a traffic flow prediction model is generated.

[0043] After the above steps, the optimized traffic flow prediction model is finally obtained, which is used to predict the traffic flow trend in future time zones and provide data support for the optimization of traffic light adaptive control strategies, thereby improving road traffic efficiency and the level of intelligence in urban traffic management.

[0044] Step S400: Evaluate the degree of change based on the real-time traffic flow data and the traffic flow prediction results to obtain the traffic flow change coefficient.

[0045] In this embodiment, by comparing and analyzing real-time traffic flow data and traffic flow prediction results, the changing trend of traffic flow is identified, and its upward, downward, or stable pattern is determined. Subsequently, the degree of fluctuation in traffic flow is assessed based on the changing trend, and the traffic flow change coefficient is calculated.

[0046] Furthermore, in the method provided in the application embodiments, the method further includes evaluating the degree of change based on the real-time traffic flow data and the traffic flow prediction results to obtain a traffic flow change coefficient, and also includes:

[0047] Based on the real-time traffic flow data, the traffic flow prediction results are analyzed to identify the trend of traffic flow changes, thereby obtaining the traffic flow change trend; based on the traffic flow change trend, the degree of change is assessed to generate the traffic flow change coefficient.

[0048] In this embodiment, the trend of traffic flow change is first identified by calculating the difference between real-time traffic flow data and traffic flow prediction results. Specifically, the difference method is used to determine whether the traffic flow is increasing, decreasing, or remaining stable by calculating the difference in traffic flow at each time point. To more accurately capture traffic flow change patterns, a sliding window method is used. In this method, the data is divided into multiple time periods (e.g., every 10 minutes), and the difference in each window is analyzed. If the difference in traffic flow continues to increase within a certain time period, it indicates that the traffic flow is increasing; if the difference decreases, it may indicate that the traffic flow is decreasing. Finally, the trend of traffic flow change is obtained through the changes in these differences, specifically manifested as the traffic flow change trend.

[0049] After identifying the trend of traffic flow changes, the next step is to quantify the degree of fluctuation in traffic flow changes using the coefficient of variation (CV). The CV is obtained by calculating the ratio of the standard deviation of traffic flow changes to the mean. Specifically, the standard deviation of traffic flow differences is first calculated to understand the dispersion of traffic flow changes; then the mean of these differences is calculated. Finally, the standard deviation of traffic flow changes is divided by the mean of the traffic flow changes to obtain the traffic flow variation coefficient.

[0050] Step S500: Determine whether the traffic flow change coefficient is greater than or equal to the traffic flow change threshold.

[0051] In this embodiment, the traffic flow change coefficient calculated above is compared with a preset traffic flow change threshold to determine whether the traffic flow change coefficient is greater than or equal to the traffic flow change threshold. The traffic flow change threshold is preset by technical experts according to requirements.

[0052] Step S600: If the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, generate a traffic light adjustment command.

[0053] In this embodiment of the application, if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, it is considered that the current traffic flow change exceeds the normal fluctuation range. Therefore, a traffic light adjustment instruction is generated, which means that the currently operating traffic light control strategy needs to be adjusted.

[0054] Step S700: Based on the traffic light adjustment command, adjust the current traffic light control strategy according to the traffic flow prediction result, generate a future traffic light control strategy, and execute the adaptive traffic light control of the target intersection according to the future traffic light control strategy.

[0055] In this embodiment, based on traffic light adjustment instructions and traffic flow prediction results, the current traffic light control strategy is first adjusted to generate multiple traffic light control adjustment schemes. Then, by predicting the traffic efficiency of these schemes, multiple different predicted traffic efficiencies are obtained. Next, based on these prediction results, the control adjustment scheme that maximizes traffic efficiency is selected to generate a future traffic light control strategy. Finally, adaptive control of the traffic lights at the target intersection is performed according to the future traffic light control strategy to ensure that the traffic lights at the intersection can be adjusted in real time according to predicted traffic flow changes, thereby improving overall traffic efficiency, reducing congestion, and ensuring smooth traffic flow.

[0056] Furthermore, in the method provided in the application embodiments, based on the traffic light adjustment command, the current traffic light control strategy is adjusted according to the traffic flow prediction result to generate a future traffic light control strategy, which further includes:

[0057] Based on the traffic light adjustment instructions, the current traffic light control strategy is adjusted according to the traffic flow prediction results to obtain a set of traffic light control adjustment schemes; traffic efficiency is predicted based on the set of traffic light control adjustment schemes to obtain multiple predicted traffic efficiencies; the set of traffic light control adjustment schemes is then filtered to maximize traffic efficiency based on the multiple predicted traffic efficiencies to obtain the future traffic light control strategy.

[0058] In this embodiment, the current traffic light control strategy is first adjusted based on traffic light adjustment instructions and traffic flow prediction results, thereby obtaining a set of traffic light control adjustment schemes. This process utilizes multi-objective optimization algorithms, such as particle swarm optimization (PSO), to adjust parameters of the traffic light control strategy, such as green light duration and red light duration, by simulating the behavior of a swarm of particles, generating multiple possible traffic light control adjustment schemes. These schemes consider multiple factors, including real-time traffic flow data, predicted traffic flow changes, and historical traffic data. During this process, the PSO algorithm is trained based on historical traffic flow data and real-time control conditions to obtain the optimal solution, thus forming the set of traffic light adjustment schemes.

[0059] Subsequently, based on the aforementioned traffic light control scheme set, traffic efficiency is predicted using a traffic simulation model (such as VISSIM). The traffic simulation model models the performance of different schemes under real traffic conditions and calculates key traffic indicators such as traffic density, delay time, and average speed. Traffic density reflects the density of vehicles on the road per unit time, calculated as the number of vehicles per unit road length; delay time measures the waiting time of vehicles under signal control, calculated by tracking vehicle dwell time using simulation software; average speed measures traffic flow, calculated by averaging the speed of each vehicle using simulation software. To quantify the traffic efficiency of each scheme, a normalization method is used to standardize different indicators, ensuring their values ​​are distributed between 0 and 1. Then, a weighted summation is used to calculate the predicted traffic efficiency. During the calculation, traffic density, the difference between 1 and delay time, and average speed are multiplied by corresponding preset weights to obtain the predicted traffic efficiency for the corresponding scheme. Multiple predicted traffic efficiencies are obtained through this process.

[0060] Finally, based on the obtained predicted traffic efficiencies, a ranking and filtering method is used to maximize the selection of alternatives. This filtering process evaluates the traffic efficiency of each alternative and selects the one with the highest efficiency. Ultimately, a future traffic light control strategy that adapts to future traffic flow changes is generated, and traffic light adaptive control is implemented at the target intersection according to this strategy, thereby achieving efficient traffic flow management.

[0061] Furthermore, the method provided in the application embodiments also includes:

[0062] If the traffic flow change coefficient is less than the traffic flow change threshold, the current traffic light control strategy is mapped to the future time zone, and adaptive traffic light control in the future time zone is executed according to the current traffic light control strategy.

[0063] In this embodiment, when the traffic flow change coefficient is less than the traffic flow change threshold, it indicates that the traffic flow change at the target intersection is relatively small, and the traffic flow prediction results for future time zones do not change significantly. Therefore, there is no need to make significant adjustments to the traffic light control strategy. In this case, by mapping the current traffic light control strategy to future time zones, a smooth transition of traffic flow is ensured in future time periods.

[0064] Specifically, the current traffic light control strategy includes traffic flow management information such as the duration and cycle of traffic lights for different directions at the target intersection. Based on the comparison between the traffic flow change coefficient and the traffic flow change threshold, if the change coefficient is less than the threshold, it indicates that the trend of traffic flow change is not significant, and the current control strategy is sufficient to meet future traffic demand.

[0065] Then, by mapping the current traffic light control strategy to future time zones, the existing traffic light timings and strategies are applied to the following time periods without modification. This operation is based on the assumption that traffic flow will not change significantly in the future, thus eliminating the need for complex strategy adjustments. This mapping method avoids over-adjustment.

[0066] Ultimately, based on the mapped traffic light control strategy, adaptive traffic light control for the future time zone is implemented. This means that the traffic light control at the target intersection will follow the current strategy, automatically adapting to future traffic flow demands without additional adjustments. This approach ensures efficient traffic management even under relatively stable traffic conditions.

[0067] In summary, the embodiments of this application have at least the following technical effects:

[0068] This application obtains real-time traffic flow data of a target intersection based on an intersection traffic perception module; obtains the current traffic light control strategy for the target intersection; predicts traffic flow for the target intersection based on the real-time traffic flow data and future time zones, obtaining a traffic flow prediction result; evaluates the degree of change based on the real-time traffic flow data and the traffic flow prediction result, obtaining a traffic flow change coefficient; determines whether the traffic flow change coefficient is greater than or equal to a traffic flow change threshold; if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, generates a traffic light adjustment command; adjusts the current traffic light control strategy based on the traffic light adjustment command and the traffic flow prediction result, generating a future traffic light control strategy, and executes adaptive traffic light control for the target intersection according to the future traffic light control strategy. This invention solves the technical problems of inaccurate traffic flow prediction and inflexible traffic light control strategies in existing technologies. By combining real-time traffic flow data and traffic flow prediction results, and adjusting the traffic light control strategy based on the prediction results, it achieves the technical effects of improving traffic efficiency, reducing traffic congestion, and optimizing traffic light control.

[0069] Example 2, based on the same inventive concept as the traffic light adaptive control method integrating traffic flow prediction in the previous examples, such as... Figure 2 As shown, this application provides a traffic light adaptive control system that integrates traffic flow prediction. The system and method embodiments in this application are based on the same inventive concept. The system includes:

[0070] The system includes: a real-time traffic flow data acquisition module 11, used to obtain real-time traffic flow data of the target intersection based on the intersection traffic perception module; a traffic light control strategy acquisition module 12, used to obtain the current traffic light control strategy of the target intersection; a traffic flow prediction module 13, used to predict the traffic flow of the target intersection based on the real-time traffic flow data and the future time zone, and obtain a traffic flow prediction result; a change degree evaluation module 14, used to evaluate the change degree based on the real-time traffic flow data and the traffic flow prediction result, and obtain a traffic flow change coefficient; a judgment module 15, used to determine whether the traffic flow change coefficient is greater than or equal to a traffic flow change threshold; an adjustment instruction generation module 16, used to generate a traffic light adjustment instruction if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold; and an adaptive control module 17, used to adjust the current traffic light control strategy based on the traffic light adjustment instruction and the traffic flow prediction result, generate a future traffic light control strategy, and execute adaptive traffic light control of the target intersection according to the future traffic light control strategy.

[0071] Furthermore, the system is also used to implement the following functions:

[0072] Based on the intersection traffic perception module, the real-time traffic perception dataset of the target intersection is obtained; the real-time traffic perception dataset is filtered and denoised to generate a traffic perception denoised dataset; outlier detection is performed on the traffic perception denoised dataset to obtain perception data outlier detection results; the traffic perception denoised dataset is corrected based on the perception data outlier detection results to obtain the real-time traffic flow data.

[0073] Furthermore, the system is also used to implement the following functions:

[0074] Obtain a traffic flow record set for the target intersection; perform time-series processing on the traffic flow record set to obtain a traffic flow summary set; perform supervised learning on a long short-term memory neural network based on the traffic flow summary set to obtain a traffic flow prediction model; input the real-time traffic flow data into the traffic flow prediction model and output the traffic flow prediction result.

[0075] Furthermore, the system is also used to implement the following functions:

[0076] The traffic flow analysis set is divided into a traffic flow training set, a traffic flow test set, and a traffic flow validation set. A long short-term memory neural network is trained under supervision using the traffic flow training set to obtain a traffic flow prediction network. The traffic flow prediction network is tested using the traffic flow test set to obtain the traffic flow prediction accuracy. Based on the traffic flow prediction accuracy, the hyperparameters of the traffic flow prediction network are tuned using the traffic flow validation set to generate the traffic flow prediction model.

[0077] Furthermore, the system is also used to implement the following functions:

[0078] Based on the real-time traffic flow data, the traffic flow prediction results are analyzed to identify the trend of traffic flow changes, thereby obtaining the traffic flow change trend; based on the traffic flow change trend, the degree of change is assessed to generate the traffic flow change coefficient.

[0079] Furthermore, the system is also used to implement the following functions:

[0080] Based on the traffic light adjustment instructions, the current traffic light control strategy is adjusted according to the traffic flow prediction results to obtain a set of traffic light control adjustment schemes; traffic efficiency is predicted based on the set of traffic light control adjustment schemes to obtain multiple predicted traffic efficiencies; the set of traffic light control adjustment schemes is then filtered to maximize traffic efficiency based on the multiple predicted traffic efficiencies to obtain the future traffic light control strategy.

[0081] Furthermore, the system is also used to implement the following functions:

[0082] If the traffic flow change coefficient is less than the traffic flow change threshold, the current traffic light control strategy is mapped to the future time zone, and adaptive traffic light control in the future time zone is executed according to the current traffic light control strategy.

[0083] It should be noted that the order of the embodiments described above is merely for descriptive purposes and does not represent the superiority or inferiority of the embodiments. Furthermore, the above description focuses on specific embodiments of this specification. The processes depicted in the accompanying drawings do not necessarily require a specific or sequential order to achieve the desired results. In some implementations, multitasking and parallel processing are possible or may be advantageous.

[0084] The above description is only a preferred embodiment of this application and is not intended to limit this application. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application.

[0085] This specification and accompanying drawings are merely illustrative examples of this application and are intended to cover any and all modifications, variations, combinations, or equivalents within the scope of this application. Clearly, those skilled in the art can make various alterations and modifications to this application without departing from its scope. Therefore, if such modifications and modifications fall within the scope of this application and its equivalents, this application intends to include such modifications and modifications.

Claims

1. A traffic light adaptive control method integrating traffic flow prediction, characterized in that, The method includes: Based on the intersection traffic perception module, obtain real-time traffic flow data for the target intersection; Obtain the current traffic light control strategy for the target intersection; Based on the real-time traffic flow data, traffic flow is predicted for the target intersection according to the future time zone to obtain the traffic flow prediction result. Based on the real-time traffic flow data and the traffic flow prediction results, the degree of change is evaluated to obtain the traffic flow change coefficient; Determine whether the traffic flow change coefficient is greater than or equal to the traffic flow change threshold; If the traffic flow change coefficient is greater than or equal to the traffic flow change threshold, a traffic light adjustment instruction is generated. Based on the traffic light adjustment instructions, the current traffic light control strategy is adjusted according to the traffic flow prediction results to generate a future traffic light control strategy, and the traffic light adaptive control of the target intersection is executed according to the future traffic light control strategy. Based on the real-time traffic flow data and the traffic flow prediction results, a change rate evaluation is performed to obtain the traffic flow change coefficient, including: Based on the real-time traffic flow data, the traffic flow prediction results are analyzed to identify the trend of traffic flow changes, thereby obtaining the traffic flow change trend. The degree of change is assessed based on the traffic flow change trend, and the traffic flow change coefficient is generated. Specifically, the difference method and the sliding window method are used to calculate the traffic flow difference between the real-time traffic flow data and the traffic flow prediction results at each time point to determine the traffic flow change trend. The fluctuation degree of traffic flow change is quantified by the ratio of the standard deviation of the traffic flow difference to the mean, and the traffic flow change coefficient is obtained.

2. The traffic light adaptive control method integrating traffic flow prediction as described in claim 1, characterized in that, Based on the intersection traffic perception module, real-time traffic flow data for the target intersection is obtained, including: Based on the intersection traffic perception module, obtain the real-time traffic perception dataset of the target intersection; The real-time traffic perception dataset is filtered and denoised to generate a traffic perception denoised dataset. Outlier detection is performed on the traffic perception denoising dataset to obtain outlier detection results for the perception data; The traffic perception denoising dataset is corrected based on the outlier detection results of the perception data to obtain the real-time traffic flow data.

3. The traffic light adaptive control method integrating traffic flow prediction as described in claim 1, characterized in that, Based on the real-time traffic flow data, traffic flow is predicted for the target intersection according to the future time zone to obtain traffic flow prediction results, including: Obtain the traffic flow record set for the target intersection; The traffic flow record set is processed in a time sequence to obtain a traffic flow analysis set. A traffic flow prediction model is obtained by supervising the learning of the long short-term memory neural network based on the traffic flow data set. The real-time traffic flow data is input into the traffic flow prediction model, and the traffic flow prediction result is output.

4. The traffic light adaptive control method integrating traffic flow prediction as described in claim 3, characterized in that, Supervised learning of a long short-term memory neural network is performed on the traffic flow analysis set to obtain a traffic flow prediction model, including: The traffic flow analysis set is divided to obtain a traffic flow training set, a traffic flow test set, and a traffic flow validation set. The long short-term memory neural network is trained under supervision based on the traffic flow training set to obtain a traffic flow prediction network. The traffic flow prediction network is tested using the traffic flow test set to obtain the traffic flow prediction accuracy. Based on the traffic flow prediction accuracy, the hyperparameters of the traffic flow prediction network are tuned according to the traffic flow validation set to generate the traffic flow prediction model.

5. The traffic light adaptive control method integrating traffic flow prediction as described in claim 1, characterized in that, Based on the traffic light adjustment instructions, the current traffic light control strategy is adjusted according to the traffic flow prediction results to generate a future traffic light control strategy, including: Based on the traffic light adjustment instructions, the current traffic light control strategy is adjusted according to the traffic flow prediction results to obtain a set of traffic light control adjustment schemes; Based on the set of traffic light control and adjustment schemes, traffic efficiency is predicted to obtain multiple predicted traffic efficiencies. Based on the multiple predicted traffic efficiencies, the traffic light control adjustment scheme set is filtered to maximize traffic efficiency, thereby obtaining the future traffic light control strategy.

6. The traffic light adaptive control method integrating traffic flow prediction as described in claim 1, characterized in that, If the traffic flow change coefficient is less than the traffic flow change threshold, the current traffic light control strategy is mapped to the future time zone, and adaptive traffic light control in the future time zone is executed according to the current traffic light control strategy.

7. A traffic light adaptive control system integrating traffic flow prediction, characterized in that, The system includes: The real-time traffic flow data acquisition module is used to obtain real-time traffic flow data of the target intersection based on the intersection traffic perception module. The traffic light control strategy acquisition module is used to obtain the current traffic light control strategy of the target intersection. The traffic flow prediction module is used to predict the traffic flow of the target intersection based on the real-time traffic flow data and the future time zone, and obtain the traffic flow prediction result. The variability evaluation module is used to evaluate the variability based on the real-time traffic flow data and the traffic flow prediction results, and obtain the traffic flow variability coefficient. The judgment module is used to determine whether the traffic flow change coefficient is greater than or equal to the traffic flow change threshold. The adjustment instruction generation module is used to generate a traffic light adjustment instruction if the traffic flow change coefficient is greater than or equal to the traffic flow change threshold. An adaptive control module is used to adjust the current traffic light control strategy based on the traffic light adjustment command and the traffic flow prediction result, generate a future traffic light control strategy, and execute adaptive traffic light control at the target intersection according to the future traffic light control strategy. The variability evaluation module is also used to evaluate the variability based on the real-time traffic flow data and the traffic flow prediction results, and to obtain the traffic flow variability coefficient, including: Based on the real-time traffic flow data, the traffic flow prediction results are analyzed to identify the trend of traffic flow changes, thereby obtaining the traffic flow change trend. The degree of change is assessed based on the traffic flow change trend, and the traffic flow change coefficient is generated. Specifically, the difference method and the sliding window method are used to calculate the traffic flow difference between the real-time traffic flow data and the traffic flow prediction results at each time point to determine the traffic flow change trend. The fluctuation degree of traffic flow change is quantified by the ratio of the standard deviation of the traffic flow difference to the mean, and the traffic flow change coefficient is obtained.

Images