A traffic light control optimization method and system for multi-source data interaction
By constructing real-time traffic twin models and traffic prediction models, traffic light signals are dynamically optimized, solving the problem that traditional systems cannot adapt to changes in traffic flow in real time, and improving traffic flow efficiency and emergency vehicle passage capacity.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- INTELLIGENT INTER CONNECTION TECH CO LTD
- Filing Date
- 2025-02-28
- Publication Date
- 2026-04-24
AI Technical Summary
Traditional traffic light control systems cannot adapt to changes in traffic flow in real time, resulting in a lack of flexibility in signal timing and an inability to accurately predict changes in traffic flow, causing traffic congestion and obstruction of emergency vehicle passage.
By acquiring real-time traffic information data, a real-time traffic twin model is constructed and embedded into a traffic prediction model to dynamically determine vehicle traffic priorities, randomly simulate signal timing schemes, select the optimal control strategy, and adjust traffic light signals.
It enables dynamic optimization of traffic light signals, improves traffic flow efficiency, reduces congestion, and enhances the passage capacity of emergency vehicles.
Smart Images
Figure CN120183214B_ABST
Abstract
Description
Technical Field
[0001] This invention relates to the field of intelligent transportation technology, specifically to a method and system for optimizing traffic light control through multi-source data interaction. Background Technology
[0002] With the continuous expansion of urban transportation and the increasing traffic flow, traditional traffic light control systems can no longer effectively cope with the rapid changes in daily traffic volume. Existing traffic light timing methods are mostly fixed patterns or simple periodic adjustments, resulting in a lack of flexibility and an inability to adapt to traffic demands at different times of day. Furthermore, traditional methods fail to fully utilize real-time traffic data and cannot accurately predict changes in traffic flow, leading to traffic congestion, increased delays, and difficulties in the passage of emergency vehicles, thus affecting the efficiency and safety of traffic management. Summary of the Invention
[0003] This application provides a traffic light control optimization method and system with multi-source data interaction, which is used to address the technical problems of inflexible traffic light signal timing and inability to adapt to traffic flow changes in real time in the prior art.
[0004] In view of the above problems, this application provides a traffic light control optimization method and system based on multi-source data interaction.
[0005] The first aspect of this application provides a traffic light control optimization method based on multi-source data interaction, the method comprising:
[0006] A real-time traffic information dataset for the target road segment is obtained, including at least real-time traffic flow data and vehicle speed. Based on the real-time traffic information dataset, a real-time traffic twin model is constructed and embedded into a traffic prediction model to predict traffic congestion and obtain congestion results. Vehicle passage priorities are dynamically determined, and traffic light control targets are generated based on the congestion results and vehicle passage priorities. Different signal timing schemes are randomly simulated using the real-time traffic twin model to determine a traffic efficiency index evaluation result set. The optimal control strategy is selected based on the traffic efficiency index evaluation result set. The traffic light signals for the target road segment are adjusted based on the optimal control strategy.
[0007] A second aspect of this application provides a traffic light control optimization system based on multi-source data interaction, the system comprising:
[0008] The system includes a data acquisition module for acquiring a real-time traffic information dataset for the target road segment, which includes at least real-time traffic flow data and vehicle speed; a congestion result acquisition module for constructing a real-time traffic twin model based on the real-time traffic information dataset and embedding it into a traffic prediction model to predict traffic congestion and obtain congestion results; a control target generation module for dynamically determining vehicle passage priorities and generating traffic light control targets based on the congestion results and vehicle passage priorities; an optimal control strategy determination module for randomly simulating different signal timing schemes through the real-time traffic twin model to determine a traffic efficiency index evaluation result set and selecting the optimal control strategy based on the traffic efficiency index evaluation result set; and a traffic light signal control module for adjusting the traffic light signals of the target road segment based on the optimal control strategy.
[0009] One or more technical solutions provided in this application have at least the following technical effects or advantages:
[0010] This application acquires a real-time traffic information dataset for a target road segment, which includes at least real-time traffic flow data and vehicle speed. Based on the real-time traffic information dataset, a real-time traffic twin model is constructed and embedded with a traffic prediction model to predict traffic congestion and obtain congestion results. Vehicle passage priorities are dynamically determined, and traffic light control targets are generated based on the congestion results and vehicle passage priorities. Different signal timing schemes are randomly simulated using the real-time traffic twin model to determine a traffic efficiency index evaluation result set. The optimal control strategy is selected based on the traffic efficiency index evaluation result set. The traffic light signals for the target road segment are adjusted based on the optimal control strategy. This invention solves the technical problems of inflexible traffic light signal timing and inability to adapt to real-time traffic flow changes in the prior art. By collecting multi-source traffic data in real time and combining it with a traffic prediction model for dynamic optimization, it achieves the technical effects of improving traffic flow efficiency, reducing traffic congestion, and enhancing the priority passage capacity of emergency vehicles. 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 a traffic light control optimization method with multi-source data interaction is provided in an embodiment of this application.
[0013] Figure 2 This is a schematic diagram of a traffic light control optimization system with multi-source data interaction, provided as an embodiment of this application.
[0014] Figure labeling: Data acquisition module 11, congestion result acquisition module 12, control target generation module 13, optimal control strategy determination module 14, traffic light signal control module 15. Detailed Implementation
[0015] This application provides a traffic light control optimization method and system based on multi-source data interaction. It addresses the technical problems of inflexible traffic light signal timing and inability to adapt to traffic flow changes in real time in existing technologies. By collecting multi-source traffic data in real time and combining it with a traffic prediction model for dynamic optimization, it achieves the technical effects of improving traffic flow efficiency, reducing traffic congestion, and enhancing the priority passage capability of emergency vehicles.
[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 control optimization method based on multi-source data interaction, the method comprising:
[0019] Step S100: Obtain the real-time traffic information dataset of the target road segment, wherein the real-time traffic information dataset includes at least real-time traffic flow data and vehicle speed.
[0020] In this embodiment, a series of traffic monitoring devices, such as cameras, sensors, or radar, deployed on the target road segment first collect traffic flow data and vehicle speed information in real time. Traffic flow data represents the number of vehicles passing through a certain road segment per unit time, while vehicle speed reflects the speed at which vehicles travel on that road segment. By collecting data from the target road segment through the traffic monitoring devices, a real-time traffic information dataset is obtained.
[0021] Step S200: Based on the real-time traffic information dataset, construct a real-time traffic twin model and embed it into a traffic prediction model to predict traffic congestion and obtain congestion results.
[0022] In this embodiment, a real-time traffic twin model is constructed based on a real-time traffic information dataset through data preprocessing and fusion with a target road segment topology map. Next, a traffic prediction model is embedded within this model to predict future traffic congestion using real-time traffic state features and generate a congestion probability. Finally, the congestion probability is compared with a preset threshold to obtain a congestion result that includes the location and level of congestion.
[0023] Furthermore, in the method provided in the application embodiments, based on the real-time traffic information dataset, a real-time traffic twin model is constructed and embedded into a traffic prediction model to predict traffic congestion, which further includes:
[0024] The real-time traffic information dataset is preprocessed, and then fused with the target road segment topology map to construct a real-time traffic twin model. The real-time traffic twin model includes a road network topology structure and a vehicle trajectory simulation module. The traffic prediction model is embedded into the real-time traffic twin model, wherein the traffic prediction model is used to receive the real-time traffic state features output by the real-time traffic twin model and generate the congestion probability of each direction in the future time period. The congestion probability is compared with a preset threshold to generate a congestion result including congestion location and level.
[0025] In this embodiment, to construct a real-time traffic twin model based on a real-time traffic information dataset and predict traffic congestion, the real-time traffic information dataset is first preprocessed. Data preprocessing includes steps such as removing invalid or outlier data, filling in missing values, and standardizing the data. This process typically uses data cleaning methods, such as outlier detection and interpolation, to handle incomplete or erroneous traffic data. After preprocessing, key data such as traffic flow and vehicle speed are formatted for further analysis. Subsequently, the real-time traffic information dataset is fused with the target road segment topology map. This step employs a fusion algorithm, such as Dynamic Time Warping (DTW), to combine the real-time traffic data with the road network topology map of the target road segment. The road network topology map describes the structure of the road network, including intersections, number of lanes, and connections between roads, and is predetermined. This fusion allows the model to simultaneously consider traffic flow, vehicle speed, and the spatial characteristics of the road network, ensuring the coordination between traffic data and the road network structure. This process allows traffic data to be combined with the spatial layout of roads to form a complete description of the traffic system.
[0026] Next, based on the fused data, a real-time traffic twin modeling method is used to construct a traffic twin model. The twin model is a digital replica of the real-world traffic environment, designed to simulate changes in real traffic flow. During construction, the twin model includes a road network topology and a vehicle trajectory simulation module. The road network topology reflects the road layout, while the vehicle trajectory simulation module dynamically simulates changes and developments in traffic flow by simulating factors such as vehicle travel paths and speed variations.
[0027] After constructing the traffic twin model, the next step is to embed a traffic prediction model into it. The function of the traffic prediction model is to predict future traffic conditions using real-time traffic status features (such as flow rate and vehicle speed) provided by the traffic twin model. Through algorithms such as machine learning or deep learning, the traffic prediction model can analyze and learn from historical and real-time data to predict changes in traffic flow on various road segments over a future period, thereby calculating the probability of congestion in each direction. By processing real-time data, the prediction model generates congestion information for different directions or road segments, thus providing information on potential future traffic congestion.
[0028] Finally, the predicted congestion probability is compared with preset congestion thresholds. These thresholds, set by technical experts based on historical traffic data and the system, include multiple thresholds, each corresponding to a different congestion level. For example, a low congestion probability corresponds to a threshold of less than 30%, indicating only light traffic flow on the road segment; a medium congestion probability corresponds to a threshold of 30% to 70%, indicating that the road segment may experience moderate congestion; and a high congestion probability corresponds to a threshold greater than 70%, indicating that the road segment may experience severe congestion. By comparing the probability with these different thresholds, specific congestion results are generated, clearly identifying the location and level of congestion. The congestion levels are divided into three levels: Level 1 for light congestion, Level 2 for moderate congestion, and Level 3 for severe congestion.
[0029] Furthermore, the method provided in the application embodiments, which embeds a traffic prediction model, further includes:
[0030] Obtain the sample traffic state feature set and the corresponding sample congestion probability set; use the sample traffic state feature set as input and the sample congestion probability set as output to train the model and obtain the traffic prediction model; embed the traffic prediction model into the real-time traffic twin model.
[0031] In this embodiment, a sample traffic state feature set and a corresponding sample congestion probability set are first obtained from a preset database. The sample traffic state feature set includes various key data extracted from real-time traffic information, such as traffic flow, vehicle speed, and vehicle density, which are characteristic variables reflecting the current traffic conditions of the road. The sample congestion probability set consists of congestion probability values corresponding to these traffic state features, labeled by technical experts, representing the probability of congestion occurring on a road segment under specific traffic conditions.
[0032] Next, the model is trained using a set of sample traffic state features as input and a set of sample congestion probabilities as output. This process employs supervised learning methods, using machine learning or deep learning algorithms such as decision trees, support vector machines (SVM), and neural networks. During training, the model learns the relationship between traffic state features and congestion probabilities by analyzing a large amount of sample data. By optimizing the model parameters, the model can accurately predict future traffic congestion probabilities based on the input traffic state features. The goal of model training is to minimize the error between the predicted and actual values, thereby obtaining a traffic prediction model that can effectively predict traffic congestion probabilities.
[0033] Once the traffic prediction model is trained, the next step is to embed it into a real-time traffic twin model. During this process, the traffic prediction model will interface with the output of the real-time traffic twin model (such as traffic flow, vehicle speed, and other traffic state features). The real-time traffic twin model provides simulated traffic state features that reflect the current traffic conditions on the road in real time. The traffic prediction model then uses these real-time features to predict the probability of traffic congestion in future time periods. By embedding the traffic prediction model, the twin model can not only reflect the current traffic conditions but also dynamically predict future congestion, thereby providing decision support for traffic management.
[0034] Step S300: Dynamically determine vehicle passage priority, and generate traffic light control targets based on the congestion results and vehicle passage priorities.
[0035] In this embodiment, the process begins by acquiring the real-time vehicle types of the target road segment. Different types of vehicles, such as emergency vehicles, public transportation vehicles, and regular vehicles, are identified in real time using sensors or other traffic monitoring equipment. Preset priorities are assigned to different vehicle types, and corresponding weights are set for each type, where the weights are pre-defined.
[0036] After acquiring real-time vehicle priority information, priority control targets are calculated based on congestion results. Congestion results reflect the traffic conditions of each road segment. By analyzing the location and level of congestion, it is determined which road segments have high traffic volume and which segments are likely to experience severe congestion. Based on this information, priority is given to road segments that require allowing emergency or high-priority vehicles to pass, preventing traffic delays or further exacerbation of congestion. Finally, based on real-time vehicle priority information and congestion results, traffic light control targets are generated.
[0037] Furthermore, in the method provided in the application embodiments, dynamically determining vehicle traffic priority and generating traffic light control targets based on the congestion results and vehicle traffic priorities also includes:
[0038] The system obtains real-time vehicle types for the target road segment, assigns preset priorities and sets predetermined weights based on the real-time vehicle types, and obtains real-time vehicle priority information. Based on the real-time vehicle priority information and congestion results, it calculates priority control targets and determines the traffic light control targets.
[0039] In this embodiment, the real-time vehicle types of the target road segment are first obtained, and different types of vehicles are identified in real time using sensors or other traffic monitoring equipment. Specifically, vehicles are identified using a pre-trained vehicle recognition model, which is trained using a convolutional neural network based on a set of vehicle sample photos and corresponding vehicle model labels.
[0040] Next, preset priorities are assigned to each type of vehicle based on real-time vehicle type. Emergency vehicles are given the highest priority (e.g., Level 3) to ensure priority passage in emergency situations. Public transportation vehicles typically receive medium priority (e.g., Level 2) depending on traffic strategies and actual demand. Ordinary private cars usually have lower priority (e.g., Level 1). In addition to assigning priorities, predetermined weights are also set to adjust the priority weights for each type of vehicle; these weights are pre-defined.
[0041] Subsequently, based on real-time vehicle priority information and congestion results, priority control targets are calculated. In this step, the real-time vehicle priority information for each road segment is combined with the congestion level in the congestion results. First, the number of vehicles on each route is multiplied by the corresponding vehicle priority, then multiplied by a preset weight, and finally weighted according to the congestion level of each road segment. In this way, the control level of each route is obtained. Then, all the calculated control levels are compared, and the road segment with the highest control level is selected as the priority control target, thereby determining the traffic light control target for that road segment.
[0042] Step S400: Randomly simulate different signal timing schemes using the real-time traffic twin model to determine the traffic efficiency index evaluation result set, and select the optimal control strategy based on the traffic efficiency index evaluation result set.
[0043] In this embodiment, a historical signal timing scheme set is first obtained, and a signal timing scheme space is established. This space contains all historically feasible signal timing schemes, each including parameters such as green light duration, red light duration, and signal switching sequence. Next, different signal timing schemes are iteratively and randomly extracted from this signal timing scheme space, and these schemes are applied to a real-time traffic twin model for simulation. Through this simulation, the actual impact of each signal timing scheme on traffic flow is evaluated, and relevant traffic efficiency indicators, such as average vehicle delay time, traffic volume, and congestion level, are calculated. These evaluation results form a traffic efficiency indicator evaluation result set. Based on these evaluation results, all schemes are compared, and the optimal control strategy is selected.
[0044] Furthermore, in the method provided in the application embodiments, the method for determining the traffic efficiency index evaluation result set by randomly simulating different signal timing schemes through the real-time traffic twin model further includes:
[0045] A set of historical signal timing schemes is obtained, and a signal timing scheme space is established. Iterative random signal timing schemes are extracted from the signal timing scheme space, applied to the real-time traffic twin model for simulation, and traffic efficiency index evaluation is performed to determine the traffic efficiency index evaluation result set.
[0046] In this embodiment, a set of historical signal timing schemes is first obtained. This includes different signal timing strategies implemented in the past, such as the green light time, red light time, and switching order between each phase at each intersection. The set of historical signal timing schemes is collected by traffic management departments based on actual road conditions and traffic flow data, and covers a variety of different traffic signal timing methods. These historical signal timing schemes constitute the signal timing scheme space.
[0047] Next, iterative random signal timing schemes are extracted from the signal timing scheme space and applied to a real-time traffic twin model for simulation, demonstrating the performance of each scheme under actual traffic conditions. Then, the performance of each signal timing scheme is quantitatively analyzed through traffic efficiency index evaluation. Traffic efficiency indicators include average vehicle delay time, traffic volume, and congestion level. Average vehicle delay time is an important indicator of vehicle waiting time at an intersection, representing the average time from entering the red light section to leaving the intersection. It is calculated by averaging the waiting times of all vehicles passing through the intersection. If a signal timing scheme results in a long waiting time, the delay time is long; conversely, a shorter waiting time indicates a more efficient scheme. For example, if one signal timing scheme results in an average waiting time of 15 seconds per vehicle, while another scheme results in 30 seconds, the former has a lower delay time and higher traffic efficiency. Traffic volume is an indicator of traffic flow, representing the number of vehicles passing through the intersection per unit time. A higher traffic volume indicates smoother traffic flow. For example, if one signal timing scheme results in 100 vehicles passing through an intersection per minute, while another scheme only allows 80, the former has a greater throughput, indicating that it is more favorable to traffic flow. Congestion level is an indicator of traffic conditions at a road segment or intersection, typically assessed through traffic flow density and queue length. High congestion levels mean dense traffic flow and prolonged vehicle delays. For instance, some signal timing schemes may lead to excessive traffic flow, resulting in long queues and congestion; while other schemes may effectively alleviate these problems and reduce congestion.
[0048] By calculating these traffic efficiency indicators, an evaluation result for each signal timing scheme is generated. Specifically, based on simulation results, the average vehicle delay time, traffic volume, and congestion level under each signal timing scheme are calculated, and these indicators are combined to form an evaluation result set. Assume there are two signal timing schemes, Scheme A and Scheme B. Scheme A has an average vehicle delay time of 15 seconds, a traffic volume of 100 vehicles / minute, and a low congestion level; Scheme B has an average vehicle delay time of 30 seconds, a traffic volume of 80 vehicles / minute, and a high congestion level. Based on these two indicators, Scheme A is significantly better than Scheme B because it not only has a lower delay time but also ensures a higher traffic volume and lower congestion level.
[0049] Finally, based on the evaluation results of all calculated traffic efficiency indicators, these schemes are compared, and the optimal traffic light control strategy is selected.
[0050] Furthermore, in the method provided in the application embodiments, the iterative random signal timing scheme extraction from the signal timing scheme space and its application to the real-time traffic twin model for simulation further includes:
[0051] An initial population is generated based on the signal timing scheme space. Each individual represents a timing scheme, including the phase difference, green light duration ratio, and phase switching sequence encoding. Each timing scheme is input into a real-time traffic twin model for simulation, and traffic efficiency indicators, including average vehicle delay time, traffic volume, and emergency vehicle priority passage rate, are calculated to generate a fitness score. Individuals with high fitness scores are selected, and offspring schemes are generated through a single-point crossover operation, retaining the best parameter combinations from the parent schemes. Gaussian mutation is performed on the phase difference and green light duration ratio in the offspring schemes. When the fitness score improvement rate is less than a preset threshold for three consecutive iterations, the current optimal timing scheme is output as the optimal control strategy.
[0052] In this embodiment, an initial population is first generated based on the signal timing scheme space. Each individual represents a signal timing scheme, which includes phase difference (i.e., the time interval between different traffic signals), green light duration ratio (representing the proportion of green light duration to the entire cycle), and phase switching sequence encoding (determining the order of signal switching). These factors are fundamental parameters for optimizing signal timing, and adjusting these parameters can affect the efficiency of traffic flow.
[0053] Next, each signal timing scheme is input into a real-time traffic twin model for simulation. The traffic twin model is a digital replica of the actual traffic environment, capable of reflecting dynamic factors such as traffic flow, vehicle speed, and congestion levels in real time. By applying the signal timing schemes to the twin model, the effect of each scheme under real traffic conditions is simulated. During the simulation, multiple traffic efficiency indicators are calculated, including average vehicle delay time (i.e., the average waiting time of vehicles during red lights), throughput (i.e., the number of vehicles passing through the intersection per unit time), and emergency vehicle priority rate (i.e., the priority of emergency vehicles under traffic light control, evaluated by tracking the passage of emergency vehicles under various signal timing schemes and calculating the ratio of emergency vehicles passing through the intersection).
[0054] Subsequently, the performance indicators for each scheme are calculated based on the simulation results, thereby generating a fitness score. The fitness score measures the merits of the signal timing scheme by converting these efficiency indicators into a comprehensive score. Specifically, the fitness score is typically calculated using a weighted average to synthesize the results of various indicators. The fitness score is calculated using the formula... The calculations show that w1, w2, and w3 are preset weights.
[0055] Next, individuals with high fitness scores are selected. In genetic algorithms, selection is based on an individual's fitness; individuals with higher fitness have a greater probability of being selected. Through a single-point crossover operation, some parameters are exchanged between two parent individuals to generate offspring. For example, assuming two signal timing schemes have green light duration ratios and phase switching sequences A and B, respectively, the crossover operation combines the green light duration ratio of A with the phase switching sequence of B to generate a new offspring scheme. In this way, high-quality parameter combinations from the parent generation are passed to the offspring to further optimize the signal timing strategy.
[0056] Next, Gaussian mutation is applied to the generated offspring schemes. Mutation enhances the algorithm's exploratory capabilities and avoids getting trapped in local optima by making small, random adjustments to certain parameters (such as phase difference and green light duration ratio). Gaussian mutation refers to fine-tuning these parameters according to a normal distribution (Gaussian distribution), that is, generating new parameter values by adding a random number drawn from the normal distribution to the original parameters. Mutation can explore new signal timing combinations, increasing the diversity of the search.
[0057] When the fitness score improvement rate falls below a preset threshold for three consecutive iterations, convergence is considered achieved, and the search process is close to the optimal solution. At this point, the timing scheme with the highest current fitness score is output as the optimal control strategy, and the optimization process ends. This optimal scheme provides the best signal timing strategy in terms of reducing traffic delays, increasing throughput, and prioritizing emergency vehicles.
[0058] Step S500: Adjust the traffic light signals of the target road segment based on the optimal control strategy.
[0059] In this embodiment, the traffic light timing of the road segment is adjusted according to the signal timing scheme in the optimal control strategy (e.g., green light duration, red light duration, phase switching sequence, etc.). Through this process, road traffic efficiency is optimized, traffic congestion is reduced, and the priority passage of emergency vehicles is improved, ultimately achieving intelligent and dynamic traffic management.
[0060] Furthermore, in the method provided in the application embodiment, after adjusting the traffic light signals of the target road segment based on the optimal control strategy, it further includes:
[0061] Real-time traffic flow data after adjustment is collected, and the actual traffic efficiency index under the current control strategy is calculated. The deviation between the actual traffic efficiency index and the simulated prediction index is analyzed. If the deviation exceeds the preset deviation threshold, a dynamic correction mechanism is triggered to re-extract the real-time traffic information dataset and update the real-time traffic twin model. Based on the updated twin model, a corrected control strategy is generated to replace the current traffic light timing scheme.
[0062] In this embodiment, traffic monitoring equipment installed on the road section, such as geomagnetic sensors, video surveillance systems, and radar sensors, first acquires real-time information on traffic flow, vehicle speed, and vehicle density. Then, the actual traffic efficiency index under the current control strategy is calculated using the same method as described above.
[0063] Next, a deviation analysis will be performed between the calculated actual traffic efficiency index and the traffic efficiency index previously simulated and predicted using a real-time traffic twin model. The real-time traffic twin model is a digital replica of the target road segment, predicting traffic performance under different traffic light timing schemes by simulating factors such as traffic flow and signal timing. Prediction results typically include vehicle delays, traffic volume, and congestion predictions. By comparing the actual traffic efficiency index with the simulated prediction index, the deviation value is calculated, usually through error analysis methods such as mean squared error. If the deviation exceeds a preset deviation threshold, it means that the current traffic light timing scheme has failed to achieve the expected results, triggering a dynamic correction mechanism.
[0064] After the dynamic correction mechanism is triggered, the real-time traffic information dataset is re-extracted, and the latest data on traffic flow, vehicle speed, and vehicle type are collected and processed again. This data is used to update the real-time traffic twin model, which is then corrected based on the new traffic information to re-evaluate the effectiveness of different signal timing schemes. Through real-time feedback and data adjustment, the traffic twin model can simulate the current traffic environment and provide more accurate traffic prediction results.
[0065] Based on the updated twin model, a new corrected control strategy is generated, and the optimal traffic light control strategy is recalculated using signal timing optimization algorithms (such as genetic algorithms and particle swarm optimization). Finally, the new control strategy replaces the current signal timing scheme, thereby achieving more efficient traffic management, further optimizing traffic flow, reducing congestion, and improving traffic efficiency.
[0066] In summary, the embodiments of this application have at least the following technical effects:
[0067] This application acquires a real-time traffic information dataset for a target road segment, which includes at least real-time traffic flow data and vehicle speed. Based on the real-time traffic information dataset, a real-time traffic twin model is constructed and embedded with a traffic prediction model to predict traffic congestion and obtain congestion results. Vehicle passage priorities are dynamically determined, and traffic light control targets are generated based on the congestion results and vehicle passage priorities. Different signal timing schemes are randomly simulated using the real-time traffic twin model to determine a traffic efficiency index evaluation result set. The optimal control strategy is selected based on the traffic efficiency index evaluation result set. The traffic light signals for the target road segment are adjusted based on the optimal control strategy. This invention solves the technical problems of inflexible traffic light signal timing and inability to adapt to real-time traffic flow changes in the prior art. By collecting multi-source traffic data in real time and combining it with a traffic prediction model for dynamic optimization, it achieves the technical effects of improving traffic flow efficiency, reducing traffic congestion, and enhancing the priority passage capacity of emergency vehicles.
[0068] Example 2, based on the same inventive concept as the traffic light control optimization method with multi-source data interaction in the foregoing examples, such as... Figure 2 As shown, this application provides a traffic light control optimization system with multi-source data interaction. The system and method embodiments in this application are based on the same inventive concept. The system includes:
[0069] The data acquisition module 11 is used to acquire a real-time traffic information dataset of the target road segment, wherein the real-time traffic information dataset includes at least real-time traffic flow data and vehicle speed; the congestion result acquisition module 12 is used to construct a real-time traffic twin model based on the real-time traffic information dataset, embed a traffic prediction model, perform traffic congestion prediction, and obtain congestion results; the control target generation module 13 is used to dynamically determine vehicle passage priority, and generate traffic light control targets based on the congestion results and vehicle passage priority; the optimal control strategy determination module 14 is used to randomly simulate different signal timing schemes through the real-time traffic twin model, determine a traffic efficiency index evaluation result set, and select the optimal control strategy based on the traffic efficiency index evaluation result set; the traffic light signal control module 15 is used to adjust the traffic light signals of the target road segment based on the optimal control strategy.
[0070] Furthermore, the system is also used to implement the following functions:
[0071] The real-time traffic information dataset is preprocessed, and then fused with the target road segment topology map to construct a real-time traffic twin model. The real-time traffic twin model includes a road network topology structure and a vehicle trajectory simulation module. The traffic prediction model is embedded into the real-time traffic twin model, wherein the traffic prediction model is used to receive the real-time traffic state features output by the real-time traffic twin model and generate the congestion probability of each direction in the future time period. The congestion probability is compared with a preset threshold to generate a congestion result including congestion location and level.
[0072] Furthermore, the system is also used to implement the following functions:
[0073] Obtain the sample traffic state feature set and the corresponding sample congestion probability set; use the sample traffic state feature set as input and the sample congestion probability set as output to train the model and obtain the traffic prediction model; embed the traffic prediction model into the real-time traffic twin model.
[0074] Furthermore, the system is also used to implement the following functions:
[0075] The system obtains real-time vehicle types for the target road segment, assigns preset priorities and sets predetermined weights based on the real-time vehicle types, and obtains real-time vehicle priority information. Based on the real-time vehicle priority information and congestion results, it calculates priority control targets and determines the traffic light control targets.
[0076] Furthermore, the system is also used to implement the following functions:
[0077] A set of historical signal timing schemes is obtained, and a signal timing scheme space is established. Iterative random signal timing schemes are extracted from the signal timing scheme space, applied to the real-time traffic twin model for simulation, and traffic efficiency index evaluation is performed to determine the traffic efficiency index evaluation result set.
[0078] Furthermore, the system is also used to implement the following functions:
[0079] An initial population is generated based on the signal timing scheme space. Each individual represents a timing scheme, including the phase difference, green light duration ratio, and phase switching sequence encoding. Each timing scheme is input into a real-time traffic twin model for simulation, and traffic efficiency indicators, including average vehicle delay time, traffic volume, and emergency vehicle priority passage rate, are calculated to generate a fitness score. Individuals with high fitness scores are selected, and offspring schemes are generated through a single-point crossover operation, retaining the best parameter combinations from the parent schemes. Gaussian mutation is performed on the phase difference and green light duration ratio in the offspring schemes. When the fitness score improvement rate is less than a preset threshold for three consecutive iterations, the current optimal timing scheme is output as the optimal control strategy.
[0080] Furthermore, the system is also used to implement the following functions:
[0081] Real-time traffic flow data after adjustment is collected, and the actual traffic efficiency index under the current control strategy is calculated. The deviation between the actual traffic efficiency index and the simulated prediction index is analyzed. If the deviation exceeds the preset deviation threshold, a dynamic correction mechanism is triggered to re-extract the real-time traffic information dataset and update the real-time traffic twin model. Based on the updated twin model, a corrected control strategy is generated to replace the current traffic light timing scheme.
[0082] It should be noted that the order of the embodiments described above is for descriptive purposes only 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.
[0083] 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.
[0084] 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 control optimization method based on multi-source data interaction, characterized in that, The method includes: Obtain a real-time traffic information dataset for the target road segment, wherein the real-time traffic information dataset includes at least real-time traffic flow data and vehicle speed; Based on the real-time traffic information dataset, a real-time traffic twin model is constructed and embedded into a traffic prediction model to predict traffic congestion and obtain congestion results. Dynamically determine vehicle traffic priority, and generate traffic light control targets based on the congestion results and vehicle traffic priorities; By randomly simulating different signal timing schemes using the real-time traffic twin model, a set of traffic efficiency index evaluation results is determined, and the optimal control strategy is selected based on the set of traffic efficiency index evaluation results. Adjust the traffic light signals of the target road segment based on the optimal control strategy; By randomly simulating different signal timing schemes using the real-time traffic twin model, a set of traffic efficiency index evaluation results is determined, including: Obtain a set of historical signal timing schemes and establish a signal timing scheme space; Iterative random signal timing schemes are extracted from the signal timing scheme space, applied to the real-time traffic twin model for simulation, and traffic efficiency index evaluation is performed to determine the traffic efficiency index evaluation result set. Iterative random signal timing schemes are extracted from the signal timing scheme space and applied to the real-time traffic twin model for simulation, including: An initial population is generated based on the signal timing scheme spatial generation. Each individual represents a timing scheme, including the phase difference, green light duration ratio, and phase switching sequence encoding. Each timing scheme is input into a real-time traffic twin model for simulation, and traffic efficiency indicators are calculated, including average vehicle delay time, traffic volume and emergency vehicle priority passage rate, and an fitness score is generated. Individuals with high fitness scores are selected, and offspring schemes are generated through single-point crossover operations, while retaining high-quality parameter combinations from the parent generation. Gaussian mutation is performed on the phase difference and green light duration ratio in the offspring scheme. When the fitness score improvement rate of three consecutive iterations is less than a preset threshold, the current optimal timing scheme is output as the optimal control strategy.
2. The traffic light control optimization method with multi-source data interaction as described in claim 1, characterized in that, Based on the aforementioned real-time traffic information dataset, a real-time traffic twin model is constructed and embedded into a traffic prediction model to predict traffic congestion, including: The real-time traffic information dataset is preprocessed, and the preprocessed real-time traffic information dataset is fused with the target road segment topology map to construct a real-time traffic twin model. The real-time traffic twin model includes a road network topology structure and a vehicle trajectory simulation module. The traffic prediction model is embedded into the real-time traffic twin model, wherein the traffic prediction model is used to receive the real-time traffic state features output by the real-time traffic twin model and generate the congestion probability of each direction in the future time period. The congestion probability is compared with a preset threshold to generate a congestion result that includes the location and level of congestion.
3. The traffic light control optimization method with multi-source data interaction as described in claim 2, characterized in that, Embedded traffic prediction models include: Obtain the sample traffic state feature set and the corresponding sample congestion probability set; The traffic prediction model is obtained by using the sample traffic state feature set as input and the sample congestion probability set as output for model training. The traffic prediction model is embedded into the real-time traffic twin model.
4. The traffic light control optimization method with multi-source data interaction as described in claim 1, characterized in that, Dynamically determine vehicle traffic priority, and generate traffic light control targets based on the congestion results and vehicle traffic priorities, including: Obtain the real-time vehicle type of the target road segment, assign a preset priority based on the real-time vehicle type, set a predetermined weight, and obtain real-time vehicle priority information; Based on the real-time vehicle priority information and combined with the congestion results, the priority control target is calculated to determine the traffic light control target.
5. The traffic light control optimization method with multi-source data interaction as described in claim 1, characterized in that, After adjusting the traffic light signals of the target road segment based on the optimal control strategy, the following steps are included: Real-time collection of adjusted traffic flow data; calculation of actual traffic efficiency indicators under the current control strategy. The actual traffic efficiency index is compared with the simulated prediction index. If the deviation exceeds the preset deviation threshold, a dynamic correction mechanism is triggered to re-extract the real-time traffic information dataset and update the real-time traffic twin model. A revised control strategy is generated based on the updated twin model to replace the current traffic light signal timing scheme.
6. A traffic light control optimization system with multi-source data interaction, used to execute the method described in any one of claims 1 to 5, characterized in that, The system includes: The data acquisition module is used to acquire a real-time traffic information dataset of the target road segment, wherein the real-time traffic information dataset includes at least real-time traffic flow data and vehicle speed; The congestion result acquisition module is used to construct a real-time traffic twin model based on the real-time traffic information dataset, embed a traffic prediction model, perform traffic congestion prediction, and obtain congestion results. The control target generation module is used to dynamically determine vehicle passage priority and generate traffic light control targets based on the congestion results and vehicle passage priorities. The optimal control strategy determination module is used to randomly simulate different signal timing schemes through the real-time traffic twin model, determine the traffic efficiency index evaluation result set, and select the optimal control strategy based on the traffic efficiency index evaluation result set. The traffic light signal control module is used to adjust the traffic light signals of the target road segment based on the optimal control strategy.
Citation Information
Patent Citations
Automatic timing method and device based on intelligent traffic lights
CN118571044A