Traffic signal timing optimization method and system, storage medium and program product

Abstract

The invention discloses a traffic signal timing optimization method and system, a storage medium and a program product, and relates to the field of traffic management data processing systems.The method comprises the steps that a plurality of adjacent intersection nodes are determined based on position identification information of local intersection nodes; collecting vehicle passing state data and vehicle characteristic data of a local intersection node, and generating traffic flow characteristic parameters; identifying a characteristic vehicle type and vehicle position information of the target characteristic vehicle; calculating the traffic influence score of the target characteristic vehicle on the traffic flow of each intersection; constructing a traffic signal optimization model, and calculating based on the traffic signal optimization model to obtain an initial timing scheme; sending the traffic flow output data to the adjacent intersection node, and receiving traffic flow input data sent by the adjacent intersection node; and updating the traffic signal optimization model, generating a final timing scheme, and adjusting signal control parameters of local intersection nodes. According to the invention, the accuracy of traffic signal timing can be improved.

Description

Technical Field

[0001] The present application relates to the field of traffic management data processing systems, and in particular to a traffic signal timing optimization method, system, storage medium and program product. Background Art

[0002] With the acceleration of urbanization and the continuous growth of motor vehicle ownership, urban traffic congestion has become increasingly prominent. In order to improve road traffic efficiency and reduce traffic congestion, traffic management departments in various places generally use intelligent traffic signal control systems for traffic management. In the complex urban road network environment, how to achieve intelligent timing control of traffic lights has become an important issue in current urban traffic management.

[0003] In the related technologies, common traffic signal timing control schemes mainly adopt two methods: fixed timing and inductive timing. The fixed timing scheme is based on historical traffic flow data and pre-sets the signal cycle and phase ratio for different time periods. The inductive timing scheme installs vehicle detectors at intersections and dynamically adjusts the signal light timing parameters according to the real-time detected vehicle queue length and waiting time. Some advanced traffic management systems also use regional coordinated timing technology to coordinate the signal timing of multiple intersections in the control area by setting green wave bands on the main roads.

[0004] However, the regional coordinated timing technology in related technologies mainly optimizes the traffic flow characteristics of main roads. The timing accuracy is not enough when dealing with different vehicles, and the optimization effect is not good. For example, when special vehicles such as large vehicles and new vehicles appear in the area, due to their long start-up time and slow travel speed, they can easily disrupt the original traffic rhythm, making it difficult for vehicles behind to follow and causing secondary congestion. Summary of the invention

[0005] The present application provides a traffic signal timing optimization method, system, storage medium and program product for improving the accuracy of traffic signal timing.

[0006] In a first aspect, the present application provides a traffic signal timing optimization method, which is applied to a traffic management system, and the method includes: determining multiple adjacent intersection nodes interconnected with the local intersection node based on location identification information of the local intersection node; collecting vehicle traffic status data and vehicle feature data of the local intersection node, and generating traffic flow feature parameters based on the vehicle traffic status data and vehicle feature data; performing feature vehicle identification based on the vehicle feature data to obtain the feature vehicle type and vehicle position information of the target feature vehicle; calculating the traffic impact score of the target feature vehicle on the traffic flow at each intersection based on the traffic flow feature parameters, the feature vehicle type and the vehicle position information; constructing a traffic signal optimization model based on the traffic flow feature parameters and the traffic impact score, and calculating an initial timing plan based on the traffic signal optimization model; sending the traffic output data including the traffic impact score and the initial timing plan to the adjacent intersection node, and receiving the traffic input data sent by the adjacent intersection node; updating the traffic signal optimization model based on the traffic input data and the traffic output data, generating a final timing plan, and adjusting the signal control parameters of the local intersection node based on the final timing plan.

[0007] In the above embodiment, the traffic management system realizes accurate identification of characteristic vehicles and traffic impact assessment based on the interconnection and coordination between local intersection nodes and adjacent intersection nodes, combined with vehicle feature recognition and traffic impact scoring; the traffic management system dynamically adjusts the timing plan according to vehicle type, location and other information, and realizes regional collaborative optimization through data interaction between nodes, thereby improving the flexibility and accuracy of signal timing, effectively alleviating the impact of characteristic vehicles on traffic flow, and improving the overall traffic efficiency of the road network.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the step of calculating the traffic impact score of the target characteristic vehicle on the traffic flow at each intersection based on the traffic flow characteristic parameters, characteristic vehicle type and vehicle position information specifically includes: calculating the traffic impact factor of the target characteristic vehicle based on the vehicle size parameters and acceleration performance parameters corresponding to the characteristic vehicle type; determining the relative distance and travel direction of the target characteristic vehicle from each intersection based on the vehicle position information; and calculating the traffic impact score of the target characteristic vehicle on the traffic flow at each intersection based on the relative distance, travel direction and traffic impact factor.

[0009] In the above embodiment, the traffic management system establishes a more accurate traffic impact evaluation mechanism by comprehensively considering the size parameters, acceleration performance parameters and relative position relationship of characteristic vehicles with each intersection. It can accurately quantify the actual impact of different types of vehicles on traffic flow and improve the pertinence and effectiveness of signal control.

[0010] In combination with some embodiments of the first aspect, in some embodiments, after the step of calculating the traffic impact factor of the target characteristic vehicle based on the vehicle size parameters and acceleration performance parameters corresponding to the characteristic vehicle type, the method also includes: obtaining the license plate information of the target characteristic vehicle, and querying the driving years data, historical violation data and traffic accident records of the corresponding target characteristic vehicle based on the license plate information; determining the driving risk level of the target characteristic vehicle based on the driving years data, historical violation data and traffic accident records; and adjusting the traffic impact factor based on the driving risk level.

[0011] In the above embodiment, the traffic management system introduces vehicle historical data analysis, including information such as driving years, violation records and accident records, and dynamically adjusts the traffic impact factor according to the risk level, thereby improving the ability to control high-risk vehicles and enhancing traffic safety.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the steps of sending traffic output data including a traffic impact score and an initial timing plan to an adjacent intersection node, and receiving traffic input data sent by the adjacent intersection node specifically include: calculating the estimated arrival time and estimated travel time of target characteristic vehicles at each entrance lane of the local intersection node based on the initial timing plan; performing weighted correction on the estimated arrival time and estimated travel time according to the traffic impact score of the target characteristic vehicle to obtain corrected time data; integrating the corrected time data, the traffic impact score and the initial timing plan into traffic output data, and sending the traffic output data to the adjacent intersection node; and receiving traffic input data fed back by the adjacent intersection node.

[0013] In the above embodiment, the traffic management system achieves more accurate traffic flow prediction by weighted correction of the arrival time and travel time of characteristic vehicles, and can timely transmit optimized traffic information to adjacent intersections, thereby promoting coordinated timing optimization within the road network and improving regional traffic organization efficiency.

[0014] In combination with some embodiments of the first aspect, in some embodiments, before performing characteristic vehicle identification based on vehicle characteristic data to obtain the characteristic vehicle type and vehicle position information of the target characteristic vehicle, the method also includes: collecting intersection environmental parameters of local intersection nodes; calculating environmental impact factors based on the intersection environmental parameters; and adjusting the confidence threshold of vehicle characteristic identification according to the environmental impact factors.

[0015] In the above embodiment, the traffic management system establishes a more reliable vehicle feature recognition mechanism by introducing intersection environmental parameters and environmental influencing factors, adaptively adjusts the recognition threshold according to environmental conditions, improves the accuracy and reliability of feature vehicle recognition, and provides a better data basis for subsequent timing optimization.

[0016] In combination with some embodiments of the first aspect, in some embodiments, the step of calculating the environmental impact factor based on the intersection environmental parameters specifically includes: extracting weather conditions, light intensity, road conditions and visibility conditions from the intersection environmental parameters; identifying abnormal features in the road conditions including water accumulation, ice, and obstruction, and determining the duration of the abnormal features in combination with the weather conditions; calculating the lighting impact coefficient based on the light intensity, abnormal features and visibility conditions; and calculating the environmental impact factor based on the abnormal features, duration and lighting impact coefficient.

[0017] In the above embodiment, the traffic management system can accurately identify and evaluate the impact of abnormal road conditions on traffic operations by analyzing multi-dimensional environmental factors such as weather conditions, light intensity, and road conditions, thereby improving the adaptability of the timing plan to environmental changes.

[0018] In combination with some embodiments of the first aspect, in some embodiments, after the steps of constructing a traffic signal optimization model based on traffic flow characteristic parameters and traffic impact scores, and calculating an initial timing plan based on the traffic signal optimization model, the method further includes: inputting the initial timing plan into a timing evaluation model to obtain a plan feasibility score; determining similar scenarios of the initial timing plan in a historical timing plan library based on traffic flow characteristic parameters and target characteristic vehicles; and when the plan feasibility score is lower than a preset feasible threshold, adjusting the initial timing plan based on timing plans for similar scenarios.

[0019] In the above embodiment, the traffic management system realizes the feasibility verification of the timing plan by introducing the timing evaluation model and historical plan analysis, and can optimize and adjust the plan based on similar scenario experience, thereby improving the reliability of the timing plan.

[0020] In a second aspect, an embodiment of the present application provides a traffic management system, which includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the traffic management system to execute the method described in the first aspect and any possible implementation method of the first aspect.

[0021] In a third aspect, an embodiment of the present application provides a computer program product comprising instructions, which, when executed on a traffic management system, enables the traffic management system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0022] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium, comprising instructions, which, when executed on a traffic management system, enables the traffic management system to execute the method described in the first aspect and any possible implementation of the first aspect.

[0023] It is understandable that the traffic management system provided in the second aspect, the computer program product provided in the third aspect, and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiment of the present application. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method, which will not be repeated here.

[0024] One or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. Due to the adoption of a coordinated timing mechanism based on the interconnection of local intersection nodes and adjacent intersection nodes, as well as an accurate calculation method for characteristic vehicle identification and traffic impact scoring, it is possible to achieve differentiated management and precise timing control of different types of vehicles, effectively solving the problem that a single fixed timing scheme in the prior art cannot cope with complex traffic flow characteristics. Through multi-dimensional data collection and analysis, the impact of characteristic vehicles on traffic flow can be accurately identified and evaluated, and regional coordinated optimization can be achieved through data interaction between nodes, thereby improving the accuracy and adaptability of signal timing and effectively improving the overall traffic efficiency of the road network.

[0025] 2. Due to the adoption of historical data analysis methods based on license plate information query, combined with multi-dimensional information such as driving years, violation records and accident records for driving risk assessment, it is possible to comprehensively and accurately assess the actual operating risks of vehicles, effectively solving the limitation of the existing technology that only relies on road conditions to conduct impact assessment. By establishing a driving risk level assessment system and dynamically adjusting the traffic impact factors, the timing plan can adapt to the traffic needs of vehicles with different risk levels, thereby improving the level of traffic safety management.

[0026] 3. Due to the adoption of an adaptive recognition mechanism based on intersection environmental parameters and the dynamic adjustment of recognition thresholds through environmental influencing factors, it is possible to maintain a stable and reliable feature vehicle recognition effect under different environmental conditions, effectively solving the problem in the prior art that recognition performance is easily affected by the environment. By comprehensively analyzing the impact of various environmental factors on recognition accuracy, a more robust recognition mechanism is established, which improves the reliability of feature vehicle recognition and the accuracy of signal timing. BRIEF DESCRIPTION OF THE DRAWINGS

[0027] Figure 1 It is a flow chart of a traffic signal timing optimization method in an embodiment of the present application; Figure 2 is another flow chart of the traffic signal timing optimization method in the embodiment of the present application; Figure 3 It is a schematic diagram of the structure of a physical device of the traffic management system in an embodiment of the present application. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to be used as limitations to the present application. As used in the specification of the present application, the singular expressions "one", "a kind of", "above", "the" and "this" are intended to also include plural expressions, unless there is a clear indication to the contrary in the context. It should also be understood that the term "and / or" used in the present application refers to any or all possible combinations comprising one or more of the listed items.

[0029] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as suggesting or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features, and in the description of the embodiments of the present application, unless otherwise specified, "plurality" means two or more.

[0030] For ease of understanding, the application scenarios of the embodiments of the present application are introduced below.

[0031] On the roads near the ring road of a certain megacity, traffic congestion is becoming increasingly serious as the number of vehicles increases rapidly. Especially during peak hours in the morning and evening, a large number of social vehicles mix with buses, trucks and other special vehicles. Due to the differences in the traffic characteristics of different vehicles, the traffic flow is inefficient. For example, when a large truck starts at an intersection, it often takes a longer time to pass due to its poor acceleration performance, making it difficult for vehicles behind to follow. At the same time, novice vehicles often have a slow speed and are not skilled in operation when passing through intersections, which will also affect the overall traffic efficiency.

[0032] In the related art, real-time control of traffic signals can be achieved by adopting fixed timing and basic induction timing schemes. Such schemes mainly rely on vehicle detectors to obtain real-time traffic flow data and adjust signal timing based on preset control rules. The following introduces the scenario of using the traffic signal timing optimization method in the related art.

[0033] At present, the traffic signal control system used in a certain city is mainly based on two methods: fixed timing and inductive timing. On the main roads, the system switches different timing schemes through a pre-set time schedule, extending the green light time of the main roads during peak hours in the morning and evening, and using a shorter cycle length during off-peak hours. At intersections equipped with coil detectors, the system can dynamically adjust the green light time according to the length of the queue. However, this solution is difficult to effectively deal with the impact of characteristic vehicles. For example, when a fully loaded truck is at the front of the fleet, the system cannot identify its vehicle model characteristics and the degree of impact, and still allocates green light time according to the travel time of ordinary vehicles, resulting in the vehicle failing to pass the intersection within one cycle, causing a serious backlog of vehicles behind.

[0034] By using the traffic signal timing optimization method in the embodiment of the present application, by accurately identifying and classifying different types of characteristic vehicles, it is not only possible to accurately assess the actual impact of characteristic vehicles on traffic flow, but also to predict potential risks based on historical data analysis, providing a more reliable decision-making basis for signal optimization. The following introduces the scenarios in which the traffic signal timing optimization method in the present application is used.

[0035] After adopting the traffic signal timing optimization method provided by this application, the system can accurately identify and evaluate the impact of characteristic vehicles. At a certain intersection, the system identified a large transport vehicle through multi-source data fusion technology, and calculated a higher traffic impact score based on its vehicle size, acceleration performance and other parameters. At the same time, by querying the historical driving data of the vehicle, it was found that it had multiple violation records, and the system increased its risk level accordingly. Based on these evaluation results, the system adjusts the timing plan in advance when the vehicle approaches the intersection, appropriately extends the green light time, and ensures that it can pass the intersection safely. In addition, the system also sends the vehicle's traffic information to downstream intersections so that it can prepare in advance. This collaborative optimization mechanism improves the traffic efficiency of characteristic vehicles while also reducing the impact on other vehicles.

[0036] It can be seen that the traffic signal timing optimization method in the embodiment of the present application can not only realize the accurate identification and evaluation of the impact of characteristic vehicles, but also effectively solve the problem of insufficient response of traditional fixed timing schemes to the traffic needs of characteristic vehicles, thereby realizing more flexible and efficient traffic signal control.

[0037] For ease of understanding, the following describes the process of the method provided by this implementation in combination with the above scenario. Figure 1 , which is a flow chart of the traffic signal timing optimization method in an embodiment of the present application.

[0038] S101. Determine a plurality of adjacent intersection nodes interconnected with the local intersection node based on location identification information of the local intersection node.

[0039] Among them, the local intersection node represents the intersection that currently needs to be optimized for traffic signal timing; the location identification information refers to the data used to uniquely identify the geographical location of the intersection, including longitude and latitude coordinates, intersection number, etc.; the adjacent intersection node represents the adjacent intersection that has direct traffic interaction with the local intersection node, usually located on the same main road or secondary road; interconnection means that there are information exchange channels between each intersection node, which can carry out real-time data transmission and sharing.

[0040] When the traffic management system receives an instruction to optimize the signal timing of a certain intersection, it needs to first determine the range of surrounding intersections related to the intersection in order to achieve regional collaborative control. Specifically, the traffic management system first obtains the location information of the local intersection node, and then queries the road network topology database to obtain all road information directly connected to the local intersection node. Subsequently, the traffic management system searches for adjacent intersection nodes along these roads, and selects the set of adjacent intersection nodes that need to be collaboratively optimized based on factors such as actual distance, traffic correlation, and communication conditions. For intersections on main roads, the search range can be extended to a longer distance to ensure the continuity of the green wave belt; for intersections on secondary roads, the coordination of adjacent nodes is mainly considered.

[0041] In some embodiments, the process of determining nodes near intersections can be achieved in a variety of ways: Optionally, the traffic management system can calculate the physical distance between nodes based on Euclidean distance, set different distance thresholds in combination with road grade and traffic capacity, and include them in the collaborative scope when the distance between nodes is less than the corresponding threshold. For example, the distance threshold of adjacent intersections on the main road can be set to 1000 meters, and it can be set to 500 meters on the secondary road; Optionally, the traffic management system can analyze the flow correlation between nodes based on historical traffic data, calculate the correlation coefficient, and determine the traffic correlation between nodes when the correlation coefficient is greater than 0.7, and include them in the collaborative optimization scope. It is understandable that other methods such as network connectivity analysis based on graph theory and path selection model based on traffic demand distribution can also be used to determine nodes near intersections, which are not limited here.

[0042] S102: Collect vehicle traffic status data and vehicle characteristic data of local intersection nodes, and generate traffic flow characteristic parameters according to the vehicle traffic status data and vehicle characteristic data.

[0043] Among them, vehicle traffic status data represents the real-time operating status information of vehicles in all directions at the intersection, including dynamic parameters such as vehicle speed, acceleration, and headway between vehicles; vehicle characteristic data refers to data used to describe the physical properties and behavioral characteristics of vehicles, including vehicle type, vehicle length, driving trajectory, etc.; traffic flow characteristic parameters represent statistical indicators used to describe the overall traffic operation status of the intersection, including flow rate, density, average speed, queue length, etc.

[0044] After determining the nodes near the intersection, the traffic management system needs to obtain real-time traffic operation data at the intersection to evaluate the current traffic status. Specifically, the traffic management system continuously collects vehicle traffic data at each entrance through a variety of sensor devices such as video detectors, geomagnetic detectors, and radars deployed at the intersection. For traffic status data, the system records vehicle location and speed information every 100ms; for vehicle feature data, the system identifies vehicle types and extracts key dimensional parameters through video image analysis. Subsequently, the traffic management system cleans and aggregates the raw data, calculates various traffic flow parameters, and stores them by lane and time dimensions.

[0045] In some embodiments, the generation of traffic flow characteristic parameters can be achieved in a variety of ways: Optionally, the traffic management system can use the moving time window method, set a 5-minute statistical cycle, count the number of various types of vehicles and calculate the flow ratio, calculate the average lane speed and occupancy rate in combination with speed measurement data, and evaluate the current service level through the density-flow relationship model; Optionally, the traffic management system can calculate the number of stops, delay time, passing time and other performance indicators based on vehicle trajectory data, and establish an intersection operation status evaluation model in combination with historical data. It is understandable that other methods such as traffic status prediction based on deep learning and delay analysis based on queuing theory can also be used to generate traffic flow characteristic parameters, which are not limited here.

[0046] S103, performing characteristic vehicle identification based on the vehicle characteristic data to obtain characteristic vehicle type and vehicle position information of the target characteristic vehicle.

[0047] Among them, characteristic vehicles refer to vehicle types that have an impact on traffic flow, including large trucks, buses, new cars, etc.; characteristic vehicle types refer to the identification for classifying characteristic vehicles, which is used to distinguish the impact characteristics of different vehicles; vehicle position information represents the real-time spatial position of characteristic vehicles in the road network, including latitude and longitude coordinates, lane, distance from the stop line, etc.

[0048] After obtaining the characteristic parameters of traffic flow, the traffic management system needs to identify vehicles that have a special impact on traffic operation. Specifically, the traffic management system first pre-processes the vehicle characteristic data, including image enhancement, noise removal and other operations. Then, based on the deep learning model, the vehicles are classified and characteristic vehicles including large trucks, school buses, and dangerous goods transport vehicles are identified. For the identified characteristic vehicles, the system further extracts their detailed characteristic parameters, including body size, wheelbase, load, etc. At the same time, the system tracks the position changes of these vehicles in real time and records their movement trajectories to provide a basis for subsequent impact assessment.

[0049] In some embodiments, the identification and tracking of characteristic vehicles can be achieved in a variety of ways: Optionally, the traffic management system can use a target detection algorithm based on YOLOv5 to identify vehicle types, extract vehicle information in combination with license plate recognition technology, and implement vehicle trajectory tracking through a multi-target tracking algorithm, and calculate the position coordinates of the vehicle in the road network in real time; Optionally, the traffic management system can be based on multi-sensor data fusion, combined with geomagnetic, video, lidar and other multi-source data, and use the Kalman filter algorithm to estimate the vehicle state to improve the accuracy of identification and positioning. It is understandable that other methods such as vehicle type identification based on image semantic segmentation and collaborative perception based on the Internet of Vehicles can also be used to achieve characteristic vehicle identification, which is not limited here.

[0050] S104. Calculate the traffic impact score of the target characteristic vehicle on the traffic flow at each intersection based on the traffic flow characteristic parameters, characteristic vehicle type and vehicle position information.

[0051] Among them, the traffic impact score represents the quantitative value of the impact of characteristic vehicles on the traffic operation at the intersection, with a score range of 0-100; the target characteristic vehicle refers to the specific characteristic vehicle whose impact needs to be evaluated at the moment; the traffic impact of traffic flow refers to the impact of characteristic vehicles on performance indicators such as intersection capacity, delay time, and queue length.

[0052] After identifying characteristic vehicles, the traffic management system needs to evaluate the actual impact of these vehicles on traffic operations. Specifically, the traffic management system first establishes a multi-dimensional scoring index system, including space occupancy, starting characteristics, and vehicle following characteristics. Then, the basic influencing factor is determined according to the type attributes of the characteristic vehicle, and the spatial attenuation coefficient is calculated based on the distance relationship between the vehicle and each intersection. For characteristic vehicles whose driving trajectories cross multiple intersections, the system needs to evaluate the degree of their impact on each intersection along the route separately. Finally, the system comprehensively considers the current traffic flow operation status and calculates a normalized impact score.

[0053] In some embodiments, the calculation of the traffic impact score can be achieved in a variety of ways: Optionally, the traffic management system can use a weighted summation method to linearly combine the vehicle type impact factor, distance attenuation factor and traffic flow state factor, where the weight coefficient is obtained through historical data training and dynamically adjusted according to the actual scenario, such as increasing the weight of the type impact factor under congestion conditions; Optionally, the traffic management system can establish a scoring model based on fuzzy logic, using various influencing factors as input variables, and obtaining a comprehensive score through fuzzy reasoning rules. It is understandable that other methods such as influence degree prediction based on neural networks and traffic flow impact simulation based on cellular automata can also be used to calculate the traffic impact score, which is not limited here.

[0054] S105. Construct a traffic signal optimization model according to traffic flow characteristic parameters and traffic impact scores, and calculate an initial timing plan based on the traffic signal optimization model.

[0055] Among them, the traffic signal optimization model represents the mathematical model used to solve the optimal signal timing plan, including the objective function and constraints; the initial timing plan refers to the initial solution of the optimization model, including parameters such as cycle length, phase difference, and green-to-signal ratio; the signal control parameters represent the specific parameter settings used to implement traffic signal control.

[0056] After obtaining the traffic impact score, the traffic management system needs to establish a signal optimization model that takes into account the impact of characteristic vehicles. Specifically, the traffic management system first constructs an optimization objective function, comprehensively considering the minimization of overall delays and the satisfaction of the traffic needs of characteristic vehicles. Then various constraints are set, including the minimum green light time, the maximum cycle length, the safety interval time, etc. The system uses a heuristic algorithm to solve the optimization model and obtain an initial timing plan that meets the constraints. During the solution process, the system needs to consider the spatial distribution of characteristic vehicles and appropriately adjust the green-to-signal ratio distribution of the affected intersections.

[0057] In some embodiments, the signal optimization model can be solved in a variety of ways: Optionally, the traffic management system can use a genetic algorithm to encode the timing scheme into chromosomes, perform population evolution through operations such as selection, crossover, and mutation, and select the individual with the highest fitness as the optimal solution, where the fitness function needs to consider the impact score of the characteristic vehicle; Optionally, the traffic management system can use a particle swarm optimization algorithm to search for the optimal solution by simulating the movement of particles in the solution space, and the speed and position update of the particles needs to consider the impact factors of the characteristic vehicles. It is understandable that other methods such as adaptive optimization based on reinforcement learning and collaborative optimization based on ant colony algorithm can also be used to achieve timing scheme optimization, which is not limited here.

[0058] S106: Send the traffic flow output data including the traffic impact score and the initial timing plan to the adjacent intersection node, and receive the traffic flow input data sent by the adjacent intersection node.

[0059] Among them, the traffic output data refers to the traffic flow prediction information flowing from the local intersection to the adjacent intersection, including flow rate, vehicle type composition, arrival time, etc.; the traffic input data refers to the traffic flow prediction information flowing from the adjacent intersection to the local intersection.

[0060] After generating the initial timing plan, the traffic management system needs to exchange data with adjacent intersections to achieve regional coordination. Specifically, the traffic management system first calculates the capacity of each exit and the expected traffic arrival time distribution based on the initial timing plan. Then, combined with the traffic impact score of the characteristic vehicle, the prediction result is corrected to generate more accurate traffic output data. The system sends this data to the relevant adjacent intersection nodes through a dedicated communication network, and receives the traffic input data fed back by these nodes. The data interaction process needs to consider real-time requirements, and usually requires that the transmission delay does not exceed 100ms.

[0061] In some embodiments, data interaction between intersections can be achieved in a variety of ways: Optionally, the traffic management system can use a real-time communication mechanism based on WebSocket to establish a persistent connection between intersection nodes, encapsulate traffic flow data in JSON format, implement two-way data push, and set up a heartbeat detection mechanism to ensure connection reliability; Optionally, the traffic management system can build a data exchange platform based on a distributed message queue, use a publish-subscribe model to achieve data sharing between multiple intersections, and support data caching and retransmission. It is understandable that other methods such as vehicle-road cooperative communication based on 5G networks and distributed data sharing based on blockchain can also be used to achieve intersection data interaction, which is not limited here.

[0062] S107: Update the traffic signal optimization model based on the traffic flow input data and the traffic flow output data, generate a final timing plan, and adjust the signal control parameters of the local intersection node based on the final timing plan.

[0063] Among them, the final timing plan represents the optimized timing plan after considering the regional synergy effect; the signal control parameters include specific control parameters such as cycle length, phase difference, green-to-signal ratio, etc.; parameter adjustment refers to converting the optimized timing plan into actual signal control instructions.

[0064] After completing the data exchange, the traffic management system needs to comprehensively consider the regional traffic demand for final optimization. Specifically, the traffic management system first integrates the received traffic input data into the optimization model and updates the demand forecast and constraints. Then, based on the updated model, the optimization calculation is re-performed to obtain the final timing plan that takes into account the influence of upstream and downstream intersections. The system converts the final plan into specific control parameters, including the start time and duration of each phase, and issues instructions through the signal machine to achieve real-time control. During the parameter adjustment process, the system will ensure the smoothness of phase switching to avoid sudden changes that may cause impacts on traffic operations.

[0065] In some embodiments, the final optimization of the timing scheme can be achieved in a variety of ways: Optionally, the traffic management system can adopt a rolling time domain optimization strategy, set an optimization cycle of 5-15 minutes, update the model parameters based on the latest traffic data at the beginning of each cycle, and quickly solve the phased optimal solution through an online optimization algorithm; Optionally, the traffic management system can establish a scenario matching-based solution optimization mechanism, match the current traffic status with typical scenarios in the historical database, and fine-tune the matching solution in combination with real-time data. It is understandable that other methods such as adaptive optimization based on predictive control and collaborative decision-making based on multi-agents can also be used to achieve the final optimization of the timing scheme, which is not limited here.

[0066] In the above embodiment, the traffic management system realizes differentiated management of different types of vehicles by establishing a characteristic vehicle impact assessment model. In practical applications, the system can also dynamically adjust the assessment parameters according to real-time road conditions and environmental conditions to further improve the adaptability of timing optimization. The following is a supplement to the scenario of this embodiment.

[0067] In the regional traffic coordination control of a certain urban agglomeration, the system can not only handle the identification and evaluation of characteristic vehicles at a single intersection, but also establish a cross-regional collaborative optimization network. For example, when a school bus fleet needs to pass through multiple intersections, the system will dynamically calculate the optimal route based on real-time road conditions, weather conditions and other environmental factors. Through real-time data interaction between the nodes at each intersection, a "green wave belt" is constructed to guide the orderly passage of the school bus fleet. At the same time, the system also predicts possible traffic risks based on historical data analysis and adjusts the timing plan for the affected areas in advance. This multi-dimensional optimization strategy not only improves the traffic efficiency of characteristic vehicles, but also realizes the overall coordination of regional traffic.

[0068] After combining the above scenarios, the following is a more detailed description of the process of the method provided by this implementation. Figure 2 , is another flow chart of the traffic signal timing optimization method in an embodiment of the present application.

[0069] S201. Determine a plurality of adjacent intersection nodes interconnected with the local intersection node based on location identification information of the local intersection node.

[0070] Referring to step S101 , the traffic management system determines a plurality of adjacent intersection nodes.

[0071] S202: Collect vehicle traffic status data and vehicle characteristic data of local intersection nodes, and generate traffic flow characteristic parameters according to the vehicle traffic status data and vehicle characteristic data.

[0072] Referring to step S102 , the traffic management system determines traffic flow characteristic parameters.

[0073] S203: Collect the intersection environment parameters of the local intersection node.

[0074] Among them, intersection environmental parameters represent data on external environmental factors that affect traffic operations and equipment detection performance; weather conditions refer to meteorological elements including precipitation, temperature, wind speed, etc.; light intensity represents the brightness of ambient light, measured in lumens / square meter; road surface conditions refer to the physical properties of the road surface, including friction coefficient, water depth, etc.; sight distance conditions represent visibility levels, which are affected by factors such as haze and dust.

[0075] Before performing feature vehicle identification, the traffic management system needs to evaluate the impact of the current environment on detection performance. Specifically, the traffic management system collects multi-dimensional environmental data in real time through environmental monitoring equipment deployed at intersections, including weather stations, light sensors, road surface condition detectors, etc. The system preprocesses and calibrates the collected data, removes outliers, and supplements missing values. For dynamically changing parameters, the system uses a sliding window method to calculate short-term trends and evaluate the potential impact of environmental changes on detection performance. The system also needs to consider the working characteristics of different detection equipment and collect key environmental parameters that affect its performance in a targeted manner.

[0076] In some embodiments, the collection and processing of environmental parameters can be achieved in a variety of ways: Optionally, the traffic management system can first start the self-check program of the environmental monitoring equipment to ensure data reliability, and then collect various sensor data according to the preset 1-minute sampling period, and then perform data standardization to unify the parameters of different dimensions to the same scale, and finally smooth the noise data through the Kalman filter algorithm; Optionally, the traffic management system can establish a multi-source data fusion mechanism, first obtain local monitoring equipment data, and then access the weather forecast data of the regional meteorological station, and then collect environmental monitoring data of surrounding intersections, and finally use DS evidence theory to perform data fusion to obtain more reliable environmental status assessment results. It is understandable that other methods such as environmental feature extraction based on deep learning and distributed environmental perception based on the Internet of Things can also be used to achieve the collection and processing of environmental parameters, which are not limited here.

[0077] S204: Calculate environmental impact factors based on intersection environmental parameters.

[0078] Among them, the environmental impact factor is a quantitative indicator of the degree of influence of different environmental conditions on the detection performance, and its value range is 0-1; the detection performance refers to the evaluation indicators such as the accuracy and recall rate of feature vehicle recognition; the quantitative indicators of the degree of influence include sub-indicators such as lighting influence factor, weather influence factor, and road surface influence factor; the weight coefficient represents the relative importance of each environmental parameter on the detection performance.

[0079] After obtaining environmental parameters, the traffic management system needs to evaluate the comprehensive impact of environmental conditions on the detection system. Specifically, the traffic management system first establishes a mapping relationship model between environmental parameters and detection performance, considering the individual impact and interaction of various environmental factors. Then the real-time environmental parameters are input into the model to calculate the impact of each environmental factor. For severe weather that affects performance, such as heavy rain and fog, the system needs to reduce the corresponding impact factor value; for normal weather with less impact, the system maintains a higher impact factor value. Finally, the system comprehensively considers various impacts and generates a comprehensive impact factor that reflects the current environmental conditions.

[0080] In some embodiments, the calculation of environmental impact factors can be achieved in a variety of ways: Optionally, the traffic management system can first build an environmental assessment model based on fuzzy reasoning, taking parameters such as light intensity, precipitation intensity, visibility, etc. as input variables, and then design a fuzzy rule base to define the relationship between parameters and the degree of influence, and then perform fuzzy reasoning to obtain the membership value of each influencing factor, and finally use the centroid method to defuzzify to obtain the accurate value of the influencing factor; Optionally, the traffic management system can use a neural network method, first use historical data to train the environment-performance mapping network, and then input the real-time environmental parameters into the trained network, and finally obtain the value of the influencing factor under the current environmental conditions through the network output. It is understandable that other methods such as regression prediction based on machine learning and rule reasoning based on expert systems can also be used to achieve the calculation of environmental impact factors, which are not limited here.

[0081] In some embodiments, the traffic management system will perform environmental impact calculations in combination with road conditions and weather, that is, the traffic management system will extract weather conditions, light intensity, road conditions, and visibility conditions from the intersection environmental parameters; identify abnormal features of road conditions including water accumulation, ice, and obstructions, and determine the duration of the abnormal features in combination with weather conditions; calculate the lighting impact coefficient based on the light intensity, abnormal features, and visibility conditions; and calculate the environmental impact factor based on the abnormal features, duration, and lighting impact coefficient.

[0082] Among them, intersection environmental parameters represent a data set of external environmental factors that affect traffic operation and equipment detection performance; weather conditions refer to real-time monitoring data of meteorological elements including precipitation, temperature, wind speed, etc.; light intensity represents the brightness of ambient light, measured in lumens / square meter; road surface conditions refer to the physical properties of the road surface, including friction coefficient, water depth, etc.; sight distance conditions represent visibility levels, which are affected by factors such as fog and haze; abnormal features refer to abnormal road conditions that affect driving safety; light impact coefficient represents the degree of influence of light conditions on detection performance; and environmental impact factor represents the quantitative value of the impact of comprehensive environmental conditions on system performance.

[0083] Before performing feature vehicle identification, the traffic management system needs to evaluate the impact of current environmental conditions on detection performance. Specifically, the traffic management system first collects real-time data through environmental monitoring equipment deployed at intersections, including weather stations, light sensors, road surface condition detectors, etc. The collected data is then preprocessed and feature extracted to identify abnormal road conditions and evaluate their duration. The system combines light intensity and visibility to establish an environmental impact assessment model, and obtains a comprehensive impact factor through multi-factor weighted calculation. This factor will be used to subsequently adjust the parameter threshold of feature vehicle identification to ensure the reliability of the identification results.

[0084] In some embodiments, the calculation of environmental impact factors can be achieved in a variety of ways: Optionally, the traffic management system can process environmental data based on a fuzzy reasoning system, first convert each environmental parameter into a language variable through fuzzification, then perform reasoning operations through an expert rule base to obtain the degree of each impact, and finally defuzzify the accurate impact factor value through the centroid method; Optionally, the traffic management system can use a multi-layer neural network structure, use the pre-processed environmental parameters as input layer data, capture the complex relationship between parameters through nonlinear transformation of the hidden layer, and finally obtain the normalized impact factor at the output layer. It is understandable that the calculation of environmental impact factors can also be achieved by other methods such as environmental impact modeling based on support vector regression and environmental feature extraction based on deep learning, which are not limited here.

[0085] S205: Adjust the confidence threshold of vehicle feature recognition according to environmental impact factors.

[0086] Among them, the confidence threshold represents the criterion for determining the credibility of the characteristic vehicle recognition result, and is used to filter the recognition results with low confidence; threshold adjustment refers to the process of dynamically changing the confidence judgment criterion according to environmental conditions.

[0087] After obtaining the environmental impact factors, the traffic management system needs to adjust the judgment criteria for feature recognition accordingly. Specifically, the traffic management system first determines the direction and amplitude of the threshold adjustment based on the size of the environmental impact factors. When the environmental conditions are poor, the system appropriately lowers the confidence threshold to avoid filtering too many valid recognition results; when the environmental conditions are good, the system increases the confidence threshold to ensure the high reliability of the recognition results. During the adjustment process, the system needs to ensure that the threshold is always within the preset reasonable range, and consider the differences in the difficulty of identifying vehicles with different characteristics, and adopt differentiated adjustment strategies.

[0088] In some embodiments, dynamic adjustment of the confidence threshold can be achieved in a variety of ways: Optionally, the traffic management system can first establish a linear mapping relationship between the environmental impact factor and the threshold adjustment amount, then set the baseline threshold and the allowable adjustment range, then calculate the specific adjustment amount based on the current environmental impact factor, and finally add the baseline threshold to the adjustment amount to obtain the threshold actually used; Optionally, the traffic management system can establish an adaptive adjustment mechanism based on performance feedback, and dynamically optimize the threshold adjustment strategy by monitoring the accuracy changes of the recognition results. It is understandable that other methods such as threshold optimization based on reinforcement learning and parameter optimization based on heuristic algorithms can also be used to achieve dynamic adjustment of the confidence threshold, which is not limited here.

[0089] S206. Perform characteristic vehicle identification based on the vehicle characteristic data to obtain characteristic vehicle type and vehicle position information of the target characteristic vehicle.

[0090] Referring to step S103 , the traffic management system will identify the target characteristic vehicle.

[0091] It should be noted that the traffic management system uses a deep convolutional neural network structure for feature vehicle recognition. The network adopts an improved YOLOv5 architecture and includes three main components: a backbone network, a feature pyramid network, and a detection head. The backbone network uses a CSPDarknet structure to extract multi-scale features, and improves the feature extraction capability through residual connections and cross-stage local networks. The feature pyramid network uses an adaptive feature fusion module to dynamically weight and fuse features of different scales through an attention mechanism to enhance the detection capability of vehicles of different sizes. The detection head part includes a classification branch and a regression branch. The classification branch outputs the probability of the vehicle type, and the regression branch predicts the position and size of the bounding box. The network training uses a multi-task loss function, including classification loss, positioning loss, and confidence loss. The classification loss uses a cross entropy function, the positioning loss uses a GIoU loss, and the confidence loss is based on binary cross entropy. In order to improve the detection accuracy, the system also introduces data enhancement strategies, including Mosaic and MixUp technologies, to enhance the model's adaptability to complex scenes. In practical applications, this method can achieve a feature vehicle detection accuracy of more than 95%.

[0092] The improved YOLOv5 architecture is mainly optimized for the feature extraction network and feature fusion mechanism. In the feature extraction network, a dual-branch attention module is introduced, which contains two submodules: spatial attention and channel attention. Spatial attention generates an attention map by calculating the response strength of each spatial position in the feature map, extracts spatial information using a parallel structure of adaptive average pooling and maximum pooling, and then obtains spatial weights through fusion of 1x1 convolutional layers. Channel attention calculates the importance of each channel through the SE module structure. First, the feature map is globally averaged pooled to obtain the channel descriptor, then the correlation between channels is learned through a two-layer fully connected network, and finally the channel weight is obtained through the sigmoid function. The outputs of the two attention branches are multiplied with the original feature map by element-wise multiplication to achieve adaptive enhancement of features. The feature fusion mechanism adopts a cross-scale adaptive fusion module, which dynamically adjusts the fusion ratio of features of different scales through learnable weight coefficients. The weight coefficient is normalized by the softmax function to ensure that the sum of the contributions of features of different scales is 1. This improved structure improves the detection performance of the model for vehicles with different size features, especially the detection accuracy under complex lighting and occlusion conditions.

[0093] S207. Calculate the traffic impact factor of the target characteristic vehicle based on the vehicle size parameters and acceleration performance parameters corresponding to the characteristic vehicle type.

[0094] Among them, vehicle size parameters represent the physical size characteristics of characteristic vehicles, including vehicle length, vehicle width, vehicle height, etc.; acceleration performance parameters refer to indicators that describe the dynamic characteristics of vehicles, including maximum acceleration, starting acceleration, maximum speed, etc.; traffic impact factors represent the basic quantitative values ​​of the impact of the inherent properties of characteristic vehicles on traffic flow operation; vehicle dynamic characteristics refer to the movement characteristics of vehicles during acceleration, deceleration, turning, etc.

[0095] After identifying the characteristic vehicle type, the traffic management system needs to quantitatively evaluate the basic impact of this type of vehicle. Specifically, the traffic management system first extracts the standard parameters of this type of vehicle from the vehicle characteristic database, including typical dimensions and performance indicators. Then it analyzes the impact of the vehicle's physical characteristics on road capacity, focusing on factors such as vehicle space occupation, starting characteristics, and following characteristics. For large vehicles, the system focuses on evaluating the impact of their space occupation and turning radius; for performance-limited vehicles, the system focuses on evaluating the impact of their acceleration and deceleration characteristics on traffic flow. Finally, the system weights and combines various impacts to generate a traffic impact factor that reflects the overall impact of this type of vehicle.

[0096] In some embodiments, the calculation of traffic impact factors can be achieved in a variety of ways: Optionally, the traffic management system can first establish an impact assessment model based on the physical characteristics of the vehicle, convert the vehicle size into a standard vehicle equivalent coefficient, and then calculate the average travel time increment based on the acceleration performance parameters, and then evaluate the increase in safety spacing caused by the vehicle's operating characteristics, and finally obtain a comprehensive impact factor through multi-factor weighting; Optionally, the traffic management system can use a data-driven method to establish a correlation model between vehicle characteristics and traffic flow parameters based on historical observation data, and determine the impact weight of each characteristic parameter through regression analysis. It is understandable that the calculation of traffic impact factors can also be achieved by other methods such as analytical calculation based on traffic flow theory and impact assessment based on micro-simulation, which are not limited here.

[0097] In some embodiments, the traffic management system will perform impact calculations based on license plate information, that is, the traffic management system will obtain the license plate information of the target feature vehicle, and query the driving years data, historical violation data and traffic accident records of the corresponding target feature vehicle based on the license plate information; determine the driving risk level of the target feature vehicle based on the driving years data, historical violation data and traffic accident records; and adjust the traffic impact factor based on the driving risk level.

[0098] Among them, the license plate information represents the coding data used to uniquely identify the vehicle; the driving years data refers to the length of time the vehicle has been in use; the historical violation data represents the vehicle's past traffic violation records; the traffic accident record refers to the traffic accidents involving the vehicle; the driving risk level represents the comprehensive assessment result of the vehicle's safety; the traffic impact factor refers to a quantitative indicator that reflects the degree of impact of the vehicle on traffic operations.

[0099] After obtaining the basic information of the characteristic vehicle, the traffic management system needs to evaluate the potential risk level of the vehicle. Specifically, the traffic management system first obtains the unique identification of the target vehicle through the license plate recognition system, and then accesses the traffic management database to obtain the historical data of the vehicle. The system analyzes the historical data, considers factors such as violation frequency and accident severity, and constructs a risk assessment model. Based on the assessment results, the system dynamically adjusts the vehicle's traffic impact factor and imposes stricter traffic control on high-risk vehicles.

[0100] In some embodiments, driving risk assessment can be achieved in a variety of ways: Optionally, the traffic management system can establish a risk reasoning model based on the Bayesian network, use historical data as prior knowledge, calculate the risk level through conditional probability, and dynamically update the model parameters according to the newly added data; Optionally, the traffic management system can use a time series analysis method to model the vehicle's violation and accident data in time series to predict future risk trends. It is understandable that driving risk assessment can also be achieved by other methods such as risk classification based on decision trees and risk identification based on anomaly detection, which are not limited here.

[0101] S208. Determine the relative distance and driving direction between the target feature vehicle and each intersection based on the vehicle position information.

[0102] Among them, relative distance represents the actual driving distance from the characteristic vehicle to each intersection node; driving direction refers to the movement trend of the characteristic vehicle, including the forward direction and turning intention; path prediction refers to the possible driving path of the vehicle based on the current state; spatial relationship analysis represents the evaluation of the position correlation between the vehicle and the road network elements.

[0103] After obtaining the real-time location of the vehicle, the traffic management system needs to analyze its spatial relationship with the surrounding intersections. Specifically, the traffic management system first locates the target vehicle on the road network topology map and obtains the location information of the surrounding intersection nodes. Then, based on the road network connection relationship, the actual driving distance of the vehicle to each intersection is calculated, taking into account the road shape and turning restrictions. The system also needs to analyze the vehicle's historical trajectory data and predict its possible driving direction based on the lane selection situation. For vehicles near the intersection, the system also needs to judge its turning intention based on the type of lane it is in and the status of the turn signal.

[0104] In some embodiments, the determination of spatial relationships can be achieved in a variety of ways: Optionally, the traffic management system can first build a shortest path calculation model based on the Dijkstra algorithm, calculate the shortest driving path from the vehicle to each intersection, then extract the distance and direction information of the path, and then estimate the arrival time based on the real-time speed of the vehicle, and finally generate a spatiotemporal relationship matrix between the vehicle and each intersection; Optionally, the traffic management system can establish a dynamic analysis model based on trajectory prediction, and use a sequence prediction algorithm to predict the movement trajectory of the vehicle for a period of time in the future. It is understandable that other methods such as path probability estimation based on Bayesian reasoning and accessibility analysis based on graph theory can also be used to achieve the determination of spatial relationships, which are not limited here.

[0105] S209: Calculate the traffic impact score of the target characteristic vehicle on the traffic flow at each intersection based on the relative distance, driving direction and traffic impact factor.

[0106] Among them, the traffic impact score represents a comprehensive evaluation indicator of the actual impact of characteristic vehicles on specific intersections.

[0107] After determining the spatial relationship, the traffic management system needs to evaluate the actual impact of the characteristic vehicles on each intersection. Specifically, the traffic management system first establishes an impact propagation model and defines the functional relationship between the impact intensity and the distance attenuation. Then the relative relationship between the vehicle's driving direction and the position of each intersection is analyzed, and the directional correlation coefficient is calculated. For intersections on the vehicle's driving path, the system assigns a higher directional correlation; for intersections that deviate from the path, the correlation coefficient is reduced. The system multiplies the traffic impact factor with the distance attenuation value and the directional correlation coefficient to obtain the impact score for each intersection. Finally, the system normalizes the score to ensure that the scores between different intersections are comparable.

[0108] In some embodiments, the calculation of the traffic impact score can be achieved in a variety of ways: Optionally, the traffic management system can first construct an impact propagation function based on the gravity model, taking the traffic impact factor as the quality term and the relative distance as the impedance term, and then introduce a direction adjustment coefficient to correct the basic impact, and then consider the modulation effect of the road network topology on the impact propagation, and finally map the score to the range of 0-100 through the sigmoid function; Optionally, the traffic management system can establish a scoring model based on a fuzzy neural network, taking the relative distance, driving direction and traffic impact factor as input variables, and directly outputting a standardized impact score through the trained network. It is understandable that the calculation of the traffic impact score can also be achieved by other methods such as the impact diffusion model based on spatiotemporal analysis and the traffic flow evolution model based on cellular automata, which are not limited here.

[0109] S210: construct a traffic signal optimization model based on traffic flow characteristic parameters and traffic impact scores, and calculate an initial timing plan based on the traffic signal optimization model.

[0110] Referring to step S105 , the traffic management system generates an initial timing plan.

[0111] It should be noted that the traffic management system uses a mixed integer programming model based on queuing theory for signal optimization. The objective function of the model consists of two parts: minimizing the total delay of general vehicles and minimizing the weighted delay of characteristic vehicles. The delay calculation is based on a deterministic queuing model, taking into account factors such as arrival flow, saturated flow, and effective green light time. The constraints include cycle length constraints, minimum green light time constraints, phase sequence constraints, and traffic demand constraints. Among them, the cycle length is limited to 40-180 seconds, the minimum green light time is set to 10-15 seconds according to the lane function, and the phase switching time considers the yellow light time and the full red interval. The influence of characteristic vehicles is introduced into the model by correcting the saturated flow and weight coefficients, and the weight coefficients are dynamically adjusted according to the traffic impact score. The model solution adopts a branch and bound algorithm to search for the optimal solution in the feasible solution space through linear relaxation and branching strategies. In order to improve the solution efficiency, the system uses heuristic rules to prune the solution space, and uses the historical optimal solution as the initial solution to accelerate convergence. In practical applications, the model can be solved within 2 seconds to meet real-time control requirements.

[0112] In some embodiments, the objective function of the model can be: D=Σαidi+Σβisi+γ(1-C / Cmax); wherein D represents the comprehensive optimization target value, which is used to evaluate the pros and cons of the overall timing plan; di is the average delay time of vehicles in the i-th entrance lane, reflecting the traffic efficiency of each entrance lane; si is the number of stops of vehicles in the i-th entrance lane, reflecting the continuity of vehicle traffic; C is the current traffic capacity of the intersection, Cmax is the maximum traffic capacity of the intersection, and the C / Cmax ratio reflects the capacity utilization of the intersection; αi, βi, and γ are corresponding weight coefficients, which can be dynamically adjusted according to real-time traffic conditions.

[0113] It should be noted that when the weight coefficient is dynamically adjusted according to the traffic impact score, it will be based on an adaptive learning algorithm. The system first constructs an evaluation index system based on vehicle characteristics, which includes three dimensions: vehicle type influence, space occupancy rate, and delay sensitivity. For each characteristic vehicle, the basic weight value is calculated based on its traffic impact score, and then dynamically corrected according to the current traffic status. The correction process uses an exponential smoothing method. The new weight value is equal to the weighted sum of the historical weight and the current calculated weight. The smoothing coefficient is optimized based on historical data using the least squares method. When traffic flow is close to saturation, the system will increase the weight coefficient of the characteristic vehicle to ensure that it has sufficient travel time; when the traffic flow is small, the weight will be appropriately reduced to balance the traffic needs of general vehicles. The weight adjustment range is controlled by upper and lower limit constraints to avoid excessive bias towards a certain type of vehicle.

[0114] During the correction process, the smoothing coefficient is determined by an adaptive optimization method based on historical data. The system first constructs a time series forecast error model and expresses the forecast error as a function of the smoothing coefficient. The optimization objective function is established by minimizing the sum of squares of the forecast error. The smoothing coefficient is iteratively optimized using the gradient descent method, in which the gradient calculation takes into account the sensitivity of the forecast error to the smoothing coefficient. To avoid overfitting, a regularization term is introduced to limit the range of change of the smoothing coefficient. During the iteration process, the system dynamically adjusts the learning rate, and reduces the learning rate to ensure convergence when the error decreases slowly. The final smoothing coefficient can not only ensure the accuracy of the forecast, but also maintain the stability of the weight adjustment. Specifically, assuming that the historical weight sequence is {0.6, 0.65, 0.55, 0.7}, when the smoothing coefficient obtained by optimization is 0.3, the new weight calculation will give the historical weight a weight of 0.3 and the current calculation weight a weight of 0.7, thereby obtaining the smoothed weight value.

[0115] S211, sending the traffic flow output data including the traffic impact score and the initial timing plan to the adjacent intersection node, and receiving the traffic flow input data sent by the adjacent intersection node.

[0116] Referring to step S106, the traffic management system will perform data interconnection.

[0117] In some embodiments, the traffic management system will interconnect data based on time data, that is, the traffic management system will calculate the estimated arrival time and estimated travel time of target characteristic vehicles at each entrance of the local intersection node based on the initial timing plan; perform weighted correction on the estimated arrival time and estimated travel time according to the traffic impact score of the target characteristic vehicle to obtain corrected time data; integrate the corrected time data, traffic impact score and initial timing plan into traffic output data, and send the traffic output data to adjacent intersection nodes; receive traffic input data fed back by adjacent intersection nodes.

[0118] Among them, the estimated arrival time indicates the time when the vehicle is expected to arrive at the intersection; the estimated passing time refers to the time required for the vehicle to pass through the intersection; the corrected time data represents the time prediction value after considering the influence of characteristic vehicles; the traffic output data refers to the traffic flow prediction information transmitted to the adjacent intersection; the traffic input data represents the traffic flow information received from the adjacent intersection.

[0119] After generating the initial timing plan, the traffic management system needs to exchange data with surrounding intersections to achieve collaborative control. Specifically, the traffic management system first predicts the arrival and travel time of each vehicle based on the current vehicle position and speed. Then, the prediction results are corrected considering the influence of characteristic vehicles. The system integrates the corrected data with the timing plan, sends it to relevant intersections through a dedicated communication network, and receives feedback data to provide a basis for subsequent optimization.

[0120] In some embodiments, the prediction and interaction of traffic flow data can be achieved in a variety of ways: Optionally, the traffic management system can use a Kalman filter to predict vehicle trajectories, build a state transition model based on historical data, and continuously update the prediction results through real-time observation data; Optionally, the traffic management system can establish a time series prediction model based on a long short-term memory network to capture the evolution of traffic flow. It is understandable that the prediction and interaction of traffic flow data can also be achieved by other methods such as traffic flow simulation based on cellular automata and delay prediction based on queuing theory, which are not limited here.

[0121] S212: Update the traffic signal optimization model based on the traffic flow input data and the traffic flow output data, generate a final timing plan, and adjust the signal control parameters of the local intersection node based on the final timing plan.

[0122] Referring to step S107, the traffic management system will generate a final timing plan and perform signal control.

[0123] In some embodiments, the traffic management system will verify the timing data, that is, the traffic management system will input the initial timing plan into the timing evaluation model to obtain a feasibility score for the plan; based on the traffic flow characteristic parameters and the target characteristic vehicles, the similar scenarios of the initial timing plan in the historical timing plan library are determined; when the feasibility score of the plan is lower than a preset feasible threshold, the initial timing plan is adjusted based on the timing plans of similar scenarios.

[0124] Among them, the timing evaluation model refers to the mathematical model used to evaluate the feasibility of the timing plan; the plan feasibility score refers to the practical evaluation index of the timing plan; similar scenarios refer to historical states with similar traffic characteristics; the preset feasible threshold refers to the minimum requirement for the feasibility of the plan; the timing plan adjustment refers to the optimization and correction process of the initial plan.

[0125] After generating the initial timing plan, the traffic management system needs to evaluate its actual feasibility. Specifically, the traffic management system first inputs the timing plan into the evaluation model and calculates its feasibility score. Then it searches for similar scenarios in historical data as a reference for adjusting the plan. When the score does not meet the requirements, the system optimizes the plan based on the successful experience of similar scenarios to ensure the practicality of the final plan.

[0126] In some embodiments, the evaluation and adjustment of the timing scheme can be achieved in a variety of ways: Optionally, the traffic management system can establish a feasibility analysis model based on fuzzy comprehensive evaluation, taking into account multiple evaluation indicators such as delays, number of stops, and queue length; Optionally, the traffic management system can use case reasoning methods to find the most matching historical cases through similarity calculations, extract their timing features for scheme optimization. It is understandable that the evaluation and adjustment of the timing scheme can also be achieved by other methods such as scheme evaluation based on reinforcement learning and parameter optimization based on genetic algorithms, which are not limited here.

[0127] In the embodiment of the present application, due to the use of deep learning-based feature vehicle identification technology and multi-dimensional impact assessment methods, combined with historical data analysis and environmental factor monitoring, a complete feature vehicle management mechanism has been established, so it is possible to achieve accurate identification and impact assessment of different types of feature vehicles, and realize multi-intersection linkage control through the regional collaborative optimization mechanism, effectively solving the problem that traditional fixed timing schemes are difficult to cope with complex traffic flow characteristics, and thus achieving more intelligent and efficient traffic signal control. The system not only improves the traffic efficiency of feature vehicles, but also enhances traffic safety through risk warning and active control mechanisms. At the same time, based on the environmental adaptive mechanism, it ensures the reliability of the control strategy under various weather conditions, and finally achieves the overall optimization of regional traffic organization.

[0128] The following describes the traffic management system in the embodiment of the present invention from the perspective of hardware processing. Figure 3 , which is a schematic diagram of the structure of a physical device of a traffic management system in an embodiment of the present application.

[0129] It should be noted that Figure 3 The structure of the traffic management system shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0130] like Figure 3 As shown, the traffic management system includes a CPU 301, which can perform various appropriate actions and processes according to the program stored in the ROM 302 or the program loaded from the storage part 308 to the RAM 303, such as executing the method described in the above embodiment. In the RAM 303, various programs and data required for the operation of the system are also stored. The CPU 301, the ROM 302 and the RAM 303 are connected to each other through a bus 304. The I / O interface 305 is also connected to the bus 304.

[0131] The following components are connected to the I / O interface 305: an input section 306 including an audio input device, a button switch, etc.; an output section 307 including a liquid crystal display (LCD) and an audio output device, an indicator light, etc.; a storage section 308 including a hard disk, etc.; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, etc. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as needed. A removable medium 311, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 310 as needed so that a computer program read therefrom is installed into the storage section 308 as needed.

[0132] In particular, according to an embodiment of the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer readable medium, and the computer program includes a computer program for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part 309, and / or installed from the removable medium 311. When the computer program is executed by the CPU 301, various functions defined in the present invention are executed.

[0133] The flowcharts and block diagrams in the accompanying drawings illustrate the possible architecture, functions and operations of the systems, methods and computer program products according to various embodiments of the present invention. Each box in the flowchart or block diagram may represent a module, a program segment, or a part of a code, and the above-mentioned module, program segment, or a part of a code contains one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions marked in the box may also occur in an order different from that marked in the accompanying drawings.

[0134] Specifically, the traffic management system of this embodiment includes a processor and a memory. The memory stores a computer program. When the computer program is executed by the processor, the traffic signal timing optimization method provided in the above embodiment is implemented.

[0135] As another aspect, the present invention further provides a computer-readable storage medium, which may be included in the traffic management system described in the above embodiment; or may exist independently without being assembled into the traffic management system. The above storage medium carries one or more computer programs, and when the above one or more computer programs are executed by a processor of the traffic management system, the traffic management system implements the traffic signal timing optimization method provided in the above embodiment.

[0136] As described above, the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present application.

[0137] As used in the above embodiments, the term "when..." may be interpreted to mean "if..." or "after..." or "in response to determining..." or "in response to detecting...", depending on the context. Similarly, the phrases "upon determining..." or "if (the stated condition or event) is detected" may be interpreted to mean "if determining..." or "in response to determining..." or "upon detecting (the stated condition or event)" or "in response to detecting (the stated condition or event)", depending on the context.

Claims

1. A traffic signal timing optimization method, characterized in that: Applied to a traffic management system, the method comprises: Based on the location identification information of the local intersection node, determining a plurality of adjacent intersection nodes interconnected with the local intersection node; Collecting vehicle traffic status data and vehicle characteristic data of the local intersection node, and generating traffic flow characteristic parameters according to the vehicle traffic status data and the vehicle characteristic data; Performing characteristic vehicle identification based on the vehicle characteristic data to obtain characteristic vehicle type and vehicle position information of the target characteristic vehicle; Calculate the traffic impact score of the target characteristic vehicle on the traffic flow at each intersection according to the traffic flow characteristic parameters, the characteristic vehicle type and the vehicle position information; According to the traffic flow characteristic parameters and the traffic impact score, a traffic signal optimization model is constructed, and an initial timing plan is calculated based on the traffic signal optimization model; Sending the traffic flow output data including the traffic impact score and the initial timing plan to the adjacent intersection node, and receiving the traffic flow input data sent by the adjacent intersection node; The traffic signal optimization model is updated based on the vehicle flow input data and the vehicle flow output data to generate a final timing plan, and the signal control parameters of the local intersection node are adjusted based on the final timing plan.

2. The method according to claim 1, characterized in that The step of calculating the traffic impact score of the target characteristic vehicle on the traffic flow at each intersection according to the traffic flow characteristic parameters, the characteristic vehicle type and the vehicle position information specifically includes: Calculating the traffic impact factor of the target characteristic vehicle based on the vehicle size parameters and acceleration performance parameters corresponding to the characteristic vehicle type; Determine the relative distance and driving direction of the target feature vehicle from each intersection according to the vehicle position information; Based on the relative distance, the driving direction and the traffic impact factor, the traffic impact score of the target feature vehicle on the traffic flow at each intersection is calculated.

3. The method according to claim 2, characterized in that After the step of calculating the traffic impact factor of the target characteristic vehicle based on the vehicle size parameter and the acceleration performance parameter corresponding to the characteristic vehicle type, the method further includes: Acquire the license plate information of the target characteristic vehicle, and query the driving age data, historical violation data and traffic accident records corresponding to the target characteristic vehicle based on the license plate information; Determining the driving risk level of the target characteristic vehicle according to the driving years data, the historical violation data and the traffic accident record; Based on the driving risk level, the traffic impact factor is adjusted.

4. The method according to claim 1, characterized in that The step of sending the traffic flow output data including the traffic impact score and the initial timing plan to the adjacent intersection node, and receiving the traffic flow input data sent by the adjacent intersection node, specifically includes: Based on the initial timing plan, the estimated arrival time and the estimated travel time of the target characteristic vehicles at each entrance road of the local intersection node are calculated; Performing weighted correction on the estimated arrival time and the estimated travel time according to the travel impact score of the target characteristic vehicle to obtain corrected time data; Integrate the corrected time data, the traffic impact score and the initial timing plan into traffic flow output data, and send the traffic flow output data to the adjacent intersection node; Receive the vehicle flow input data fed back by the adjacent intersection node.

5. The method according to claim 1, characterized in that Before the step of performing characteristic vehicle identification based on the vehicle characteristic data to obtain characteristic vehicle type and vehicle position information of the target characteristic vehicle, the method further includes: Collecting the intersection environment parameters of the local intersection node; Calculating environmental impact factors based on the intersection environmental parameters; The confidence threshold of vehicle feature recognition is adjusted according to the environmental impact factor.

6. The method according to claim 5, characterized in that The step of calculating the environmental impact factor based on the intersection environmental parameters specifically includes: Extracting weather conditions, light intensity, road surface conditions and sight distance conditions from the intersection environmental parameters; Identify abnormal features of the road condition including water accumulation, ice, and obstruction, and determine the duration of the abnormal features in combination with the weather conditions; Calculating a lighting influence coefficient according to the lighting intensity, the abnormal characteristics and the viewing distance; The environmental impact factor is calculated according to the abnormal characteristics, the duration and the light impact coefficient.

7. The method according to claim 1, characterized in that After the step of constructing a traffic signal optimization model according to the traffic flow characteristic parameters and the traffic impact score, and calculating an initial timing plan based on the traffic signal optimization model, the method further includes: Input the initial timing plan into the timing evaluation model to obtain a feasibility score of the plan; Based on the traffic flow characteristic parameters and the target characteristic vehicles, determining similar scenarios of the initial timing plan in a historical timing plan library; When the feasibility score of the scheme is lower than a preset feasible threshold, the initial timing scheme is adjusted based on the timing schemes of similar scenarios.

8. A traffic management system, characterized in that: The traffic management system includes: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors call the computer instructions to enable the traffic management system to execute the method described in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that: When the instruction is executed on a traffic management system, the traffic management system is caused to execute the method according to any one of claims 1 to 7.

10. A computer program product, characterized in that When the computer program product runs on a traffic management system, the traffic management system is enabled to perform the method according to any one of claims 1 to 7.

Images