Urban intersection intelligent signal timing optimization method based on vehicle-road cooperation

Abstract

The invention discloses an urban intersection intelligent signal timing optimization method based on vehicle-road cooperation, and relates to the technical field of intelligent traffic, and the method comprises the specific steps: collecting intersection traffic data in real time through an Internet of Things equipment linkage system such as a geomagnetic sensor, and transmitting the intersection traffic data to a regional center after localization processing; fusing the data to predict the traffic flow situation, and generating a single-point dynamic timing scheme; a regional scheme is obtained through multi-intersection collaborative optimization; issuing and executing the scheme, collecting feedback data, and forming a closed-loop optimization system based on error optimization algorithm parameters; according to the invention, multi-source data acquisition and traffic flow prediction are fused, the traffic signal timing refinement and foresight are realized, and the traffic efficiency of a single-point intersection is improved; meanwhile, through a multi-intersection collaborative optimization and closed loop iteration mechanism, a global optimization system is constructed, single-point timing isolation is broken, regional traffic flow is efficiently scheduled, vehicle delay is reduced, and an effective scheme is provided for urban traffic control.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent transportation, in particular to a city intersection intelligent signal timing optimization method based on vehicle-road cooperation. BACKGROUND

[0002] With the acceleration of urbanization, the number of motor vehicles in cities continues to grow. As a key node where traffic flows converge, the rationality of intersection signal timing directly affects the overall traffic efficiency. Traditional traffic signal timing relies on fixed cycle schemes, which are difficult to adapt to dynamically changing traffic flow demands. The development of vehicle-road cooperation technology provides new technical support for traffic signal optimization. Currently, Internet of Things devices such as geomagnetic sensors and high-definition cameras have been widely used in traffic data collection. 5G edge computing technology enables rapid local processing and transmission of data. Regional data processing centers can integrate and analyze multi-source traffic data. The gradual maturity of these technologies creates conditions for building a more intelligent and efficient intersection signal timing optimization system, promoting the transformation of traffic management from passive response to active prediction and dynamic adjustment.

[0003] Traditional traffic signal timing technology has many shortcomings and cannot meet the complex and changing urban traffic demands. Some timing schemes are based on historical traffic data to set fixed cycles, which cannot respond to instantaneous changes in traffic flow in real time, leading to increased congestion at peak hours and wasted resources at off-peak hours. Some single-point dynamic timing techniques can consider the current traffic state at the intersection, but they ignore the correlation between adjacent intersections and lack regional coordination optimization mechanisms, which can easily cause traffic flow imbalance in the region and form new congestion nodes. In addition, traditional techniques do not adequately consider pedestrian crossing needs, and conflicts between pedestrians and vehicles occur frequently. There is also a lack of effective scheme iteration and optimization mechanism, and the timing strategy is difficult to self-improve with long-term changes in traffic flow, resulting in poor overall adaptability and flexibility. SUMMARY

[0004] The present application aims to overcome the shortcomings of the prior art and provides a city intersection intelligent signal timing optimization method based on vehicle-road cooperation. This method collects and processes multi-source data through geomagnetic sensors, AI high-definition cameras, and 5G edge computing gateways, fuses real-time and historical data to predict traffic flow patterns, generates single-point timing schemes using dynamic multi-factor algorithms, and achieves multi-intersection cooperative optimization through flow-phase difference coupling conflict index algorithms. Finally, through a closed-loop iterative optimization mechanism, algorithm parameters are adjusted based on actual execution results to form an efficient and accurate regional traffic signal timing system.

[0005] To solve the above technical problems, the present application provides the following technical solution: a city intersection intelligent signal timing optimization method based on vehicle-road cooperation. The specific steps of this method are as follows: S100, multi-source data acquisition and processing: through the linkage system of the Internet of Things equipment composed of a geomagnetic sensor, an AI high-definition camera and a 5G edge computing gateway, real-time intersection traffic data is collected, and after being processed locally by the 5G edge computing gateway, the data is transmitted to a regional data processing center; S200, traffic flow trend prediction: the standardized real-time traffic data set is fused with the data in the historical traffic flow database, the spatial correlation characteristics of the traffic flow are extracted, the traffic state comprehensive evaluation algorithm is used to evaluate the traffic state of the current intersection and the associated area, and the traffic flow trend in a preset time period in the future is predicted through the space-time decay flow prediction algorithm; S300, single-point dynamic timing generation: based on the traffic state evaluation result and the short-term prediction trend, combined with real-time traffic data, the dynamic multi-factor green light duration calculation algorithm is used to dynamically adjust the duration of the intersection signal, and an initial timing scheme of a single point is generated; S400, multi-intersection cooperative optimization: the initial timing scheme of a single point of each intersection is taken as input, adjacent intersections are divided into independent agents, the flow-phase difference coupling conflict index algorithm is used to calculate the cooperative conflict index between intersections, and the profit function of each agent is constructed, and the regional cooperative timing scheme is determined through a distributed negotiation mechanism; S500, closed-loop iterative optimization: the regional cooperative timing scheme is issued to each intersection signal controller for execution, real-time traffic data after the execution of the scheme is collected by the Internet of Things equipment and fed back to the regional data processing center; based on the error between the actual data and the predicted data, the algorithm parameters are optimized, and a closed-loop optimization system is formed.

[0006] Further, in the S100, multi-source data acquisition and processing, the collected traffic data includes: the vehicle presence state of each lane collected by the geomagnetic sensor, the number of vehicles passing per unit time, the instantaneous vehicle speed and the vehicle stay time; the vehicle type of each direction, the vehicle queue length starting from the stop line, the number of queued vehicles, the pedestrian position, the pedestrian waiting state and the pedestrian crossing state collected by the AI high-definition camera; after the above data is processed by the 5G edge computing gateway, a standardized data set containing device identification, collection timestamp, data type and corresponding value is formed.

[0007] Further, in the S200, traffic flow trend prediction, the expression of the traffic state comprehensive evaluation algorithm is: wherein, is the traffic state comprehensive index of the direction at the moment is the actual vehicle flow of the direction at the moment is the historical maximum vehicle flow of the direction is the traffic state comprehensive index of the direction at the moment is the actual vehicle flow of the direction at the moment is the historical maximum vehicle flow of the direction is the traffic state comprehensive index of the direction at the moment is the actual vehicle flow of the direction at the moment is the historical maximum vehicle flow of the direction Time Average speed of the direction, is Design maximum speed of the direction, is Time Actual queue length of the direction, is Maximum allowable queue length of the direction, is Time Pedestrian crossing demand normalized value of the direction, is the weight coefficient of each factor, and satisfies .

[0008] Further, in the traffic flow state prediction, the expression of the space-time decay flow prediction algorithm of the S200 is: wherein, is The traffic volume prediction value of the direction at the future time, is the historical data backtracking window size, is the data sampling interval, is the weight coefficient of the historical time data, is the time decay coefficient, is a set of adjacent intersection directions associated with the traffic flow of the direction, is The actual traffic volume of the adjacent direction at the time, denotes the adjacent intersection direction associated with the traffic flow of the direction, which is an element in the set of adjacent intersection directions , is The traffic flow correlation coefficient between the direction and the direction, is the spatial correlation weight coefficient, denotes the time step number of the historical data backtracking.

[0009] Further, in the single-point dynamic timing generation, the expression of the dynamic multi-factor green light duration calculation algorithm of the S300 is: wherein, is The green light duration of the direction at the time, is The actual queue length of the direction at the time , is The actual queue length of the direction at the time average speed of the direction, safety redundancy time, safety redundancy time, future 5-minute traffic flow prediction value of the direction, safety redundancy time, historical maximum traffic flow of the direction, traffic flow prediction sensitivity coefficient, optimized based on historical data, time traffic state comprehensive index of the direction, safety redundancy time, maximum traffic state comprehensive index of the direction, traffic state influence coefficient, optimized based on historical data.

[0010] Further, in the multi-intersection cooperative optimization, the S400, the independent intelligent agent is a city intersection, and the attributes of each intelligent agent include: an initial timing scheme of a single point itself, a spatial position relationship with adjacent intelligent agents, historical traffic flow data, a cooperative conflict index set with adjacent intelligent agents calculated by a traffic flow-phase difference coupling conflict index algorithm, and a benefit value calculated based on a benefit function; and each intelligent agent realizes information interaction of a timing strategy, a benefit value and a cooperative conflict index through a vehicle-road cooperative network.

[0011] Further, in the multi-intersection cooperative optimization, the S400, the expression of the traffic flow-phase difference coupling conflict index algorithm is: wherein, is the intersection is the cooperative conflict index of the intersection , is the actual green phase difference of the intersection and the intersection , is the optimal green wave band phase difference of the intersection and the intersection , are current traffic flows of the intersection and the intersection , are historical peak traffic flows of the intersection and the intersection , the cooperative area range formed by the adjacent intersections is 1-2 kilometers in radius, contains 3-5 intersections, and the distance between adjacent intersections is not more than 800 meters.

[0012] Further, in the multi-intersection cooperative optimization, the S400, the expression of the benefit function is: wherein, ​​​a benefit value of the intersection agent , a real-time volume of the intersection , a historical maximum volume of the intersection , a real-time average delay time of the intersection , a historical average delay time of the intersection , a set of intersections adjacent to the intersection , a cooperative conflict index of the intersection and the intersection , a weight coefficient of each benefit item, and satisfying .

[0013] Further, in the multi-intersection cooperative optimization, the weight coefficient of each benefit item is adjusted according to a time period: a first set of weight values is used in a peak period, a second set of weight values is used in a flat peak period, and a third set of weight values is used in a special event period; the number of iterations of the distributed negotiation mechanism is 3-5 rounds, and each round of iteration takes no more than 3 seconds; when the cooperative conflict index , forced coordination is triggered to preferentially guarantee the green wave phase difference of the trunk road, and the deviation is not more than 10 seconds.

[0014] Further, in the multi-intersection cooperative optimization, the specific steps for determining the regional cooperative timing scheme by the distributed negotiation mechanism are as follows: each intersection agent calculates an initial benefit value based on the initial timing scheme of the single point itself, and shares the initial strategy and the benefit value through the vehicle-road cooperative network; each agent iterates through the candidate timing schemes in the strategy space of itself, calculates the expected benefit value after switching to the candidate scheme, and screens out the candidate schemes that can improve the benefit of itself and have an acceptable impact on the benefits of adjacent agents; each agent sends the screened candidate schemes to adjacent agents through the vehicle-road cooperative network, and the adjacent agents evaluate the impact of the schemes on their own benefits. If both sides agree, the strategy is updated, otherwise, it is rolled back to the original scheme; steps (2) and (3) are repeated until the strategies of all agents no longer change in continuous 3 rounds of iteration. At this time, the combination of the timing schemes is the regional cooperative timing scheme; (5) consistency check is performed on the regional cooperative timing scheme to ensure that the deviation of the green wave phase difference of adjacent intersections is not more than a preset threshold of 10 seconds. If there is a deviation, the timing parameters of the related intersections are fine-tuned, and finally the executable regional cooperative timing scheme is determined.

[0015] Compared with the prior art, the intelligent signal timing optimization method for urban intersection based on vehicle-road cooperation has the following beneficial effects: Firstly, the present application realizes the fine and forward-looking of traffic signal timing through the deep fusion of multi-source data acquisition and processing and traffic flow state accurate prediction. The comprehensive traffic data is collected by the linkage of multi-source Internet of Things devices, and the data transmission efficiency and quality are guaranteed after local processing. The spatial correlation characteristics are extracted by combining historical data fusion, the current traffic state is accurately evaluated and the future trend is predicted, and the signal length of the intersection is dynamically adjusted based on these accurate data support. The generated single-point initial timing scheme can match the traffic flow changes in real time, avoid the hysteresis and limitations of traditional timing schemes, effectively improve the traffic efficiency of single-point intersection, reduce the waiting time of vehicle queue, and at the same time, the needs of pedestrians crossing the street are considered, and the dynamic balance of human and vehicle traffic is realized.

[0016] Secondly, the present application constructs a global optimization system of regional traffic signal timing through multi-intersection cooperative optimization and closed-loop iterative optimization mechanism. The adjacent intersections are divided into independent agents, the cooperative conflict index is calculated through a specific algorithm, and the benefit function is constructed. The dynamic optimization of regional timing scheme is realized by means of distributed negotiation mechanism, the cooperative linkage between adjacent intersections is guaranteed, the overall traffic efficiency of the region is improved, the actual data after the implementation of the scheme is continuously collected by the closed-loop iterative optimization system, the algorithm parameters are compared with the predicted data, and the timing scheme is continuously improved in the cycle optimization, which adapts to the dynamic changes of traffic flow. This mode breaks the isolation of single-point timing and realizes the efficient scheduling of regional traffic flow, significantly reduces the average delay of vehicles in the region, improves the utilization rate of road resources, and provides an effective solution to urban traffic congestion.

[0017] Other advantages, objects, and features of the present application will be apparent to those skilled in the art from the following specification, and will be learned from the practice of the present application. BRIEF DESCRIPTION OF DRAWINGS

[0018] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed to be used in the embodiment or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creating laborious work.

[0019] Figure 1 The flowchart of the intelligent signal timing optimization method for urban intersection based on vehicle-road cooperation; Figure 2 The step disassembly diagram of the intelligent signal timing optimization method for urban intersection based on vehicle-road cooperation. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] Example 1: Coordinated optimization of main roads and adjacent intersections during the morning rush hour.

[0022] S100, Multi-source data acquisition and processing: An IoT device linkage system comprised of geomagnetic sensors, AI high-definition cameras, and a 5G edge computing gateway collects real-time traffic data at a core intersection on a major urban road and two adjacent intersections. The geomagnetic sensors collect data on vehicle presence in each lane, the number of vehicles passing per unit time, instantaneous speed, and vehicle dwell time, accurately capturing real-time vehicle dynamics within each lane and providing foundational data for subsequent lane congestion assessment. The AI ​​high-definition cameras capture vehicle types in each direction, queue lengths starting from the stop line, the number of vehicles in the queue, pedestrian positions, pedestrian waiting status, and pedestrian crossing status, providing a comprehensive understanding of the intersection's vehicle and pedestrian traffic demands and preventing inappropriate timing due to missing information. This data is then locally processed by the 5G edge computing gateway to form a standardized dataset containing device identifiers, collection timestamps, data types, and corresponding values, ensuring a consistent data format for subsequent integration with historical data and algorithmic calculations. This standardized dataset is then transmitted to a regional data processing center, providing a complete and accurate data source for traffic flow prediction. Figure 1 As shown.

[0023] S200, Traffic Flow Forecast: The standardized real-time traffic dataset, after collection and processing, is fused with data from the historical traffic flow database. This effectively combines current traffic conditions with past traffic patterns, reducing the limitations of single data dimensions. Simultaneously, spatial correlation features of traffic flow are extracted to clarify the traffic flow impact relationships between intersections and directions. A comprehensive traffic condition assessment algorithm is then applied to comprehensively evaluate the traffic conditions of the current main road's core intersections and adjacent intersections in all directions, based on four dimensions: traffic volume, average vehicle speed, queue length, and pedestrian crossing demand. The expression for the comprehensive traffic condition assessment algorithm is: ,in, for time The comprehensive traffic condition index for each direction. for time Actual traffic flow in the direction for The highest historical traffic volume in this direction is the moment the average speed of vehicles in the direction, is the design maximum speed of vehicles in the direction, is the moment the actual queue length in the direction, is the maximum allowable queue length in the direction, is the moment the pedestrian crossing demand normalized value in the direction, is the weight coefficient of each factor, and satisfies Let the staff clearly understand the congestion degree and traffic demand priority of each direction, and through the space-time decay flow prediction algorithm, combined with historical data backtracking and adjacent intersection traffic flow correlation, fully consider the reference value of historical data in the time dimension and the mutual influence of adjacent intersections in the space dimension, accurately predict the traffic flow change trend of each intersection in each direction within the next 15 minutes, and the expression of the space-time decay flow prediction algorithm is: wherein, is the vehicle flow prediction value of the direction in the future moment, is the historical data backtracking window size, is the data sampling interval, is the weight coefficient of the th historical moment data, is the time decay coefficient, is the adjacent intersection direction set that exists traffic flow correlation with the direction, is the actual vehicle flow of the adjacent direction at the moment, indicates the adjacent intersection direction that exists traffic flow correlation with the direction, which is an element in the adjacent intersection direction set , is the traffic flow correlation coefficient between the direction and the direction, is the space correlation weight coefficient, indicates the time step number of historical data backtracking, successfully predicts that the city-bound direction of the main road will appear a continuous vehicle flow peak, provides a forward-looking basis for subsequent single-point dynamic timing generation, and avoids that the timing scheme lags behind the traffic flow change.

[0024] S300, single-point dynamic timing generation: Based on traffic condition assessment results and short-term traffic flow forecast trends, and fully utilizing the comprehensive understanding of current traffic conditions and predictions of future traffic flow changes, combined with real-time traffic data, the timing scheme is designed to accurately reflect current traffic conditions. A dynamic multi-factor green light duration calculation algorithm is employed, comprehensively considering factors such as queue length, average vehicle speed, and future traffic volume forecasts to dynamically adjust the green light duration for each intersection and direction. The expression for the dynamic multi-factor green light duration calculation algorithm is as follows: ,in, for time Green light duration for each direction for time The actual queue length in the direction. for time Average vehicle speed in the direction, For safety redundancy time, for Traffic flow forecast for the next 5 minutes in the direction for The highest historical traffic volume in this direction The sensitivity coefficient for traffic flow prediction is optimized based on historical data. The result obtained through the traffic condition comprehensive assessment index algorithm time Comprehensive traffic condition index for directions for The maximum comprehensive traffic condition index for a given direction. The traffic condition impact coefficient is optimized based on historical data to avoid timing imbalances caused by a single factor determining the green light duration. For key intersections on main roads with long queues in the direction of traffic entering the city and a continued increase in traffic flow in the future, the green light duration is appropriately extended to effectively alleviate queuing pressure in that direction and improve vehicle traffic efficiency. At the same time, the green light duration of other directions is reasonably allocated to ensure the basic traffic needs of each direction. Finally, the initial timing scheme for each intersection is generated, laying the foundation for multi-intersection collaborative optimization.

[0025] S400, multi-intersection collaborative optimization: The initial timing schemes at each intersection are used as input to provide an initial basis for collaborative optimization. Simultaneously, the core intersections on the main road and two adjacent intersections are divided into independent intelligent agents, allowing each intersection to autonomously participate in timing optimization decisions, thus improving optimization efficiency. Each intelligent agent's attributes include its own initial timing scheme, its spatial relationship with neighboring intelligent agents, historical traffic flow data, and a set of collaborative conflict indices with neighboring intelligent agents calculated using the flow-phase difference coupling conflict index algorithm. The expression for the flow-phase difference coupling conflict index algorithm is: ,in, For the intersection Intersection The collaborative conflict index, For the intersection Intersection The actual green light phase difference For the intersection Intersection The optimal green band phase difference, , Intersections ,intersection Current traffic flow , Intersections ,intersection The historical peak traffic flow, the collaborative area formed by the adjacent intersections has a radius of 1-2 kilometers, includes 3-5 intersections, and the distance between adjacent intersections does not exceed 800 meters, and the revenue value calculated based on the revenue function, the expression of which is: ,in, Intelligent agents at intersections The profit value, For the intersection The actual traffic volume per unit time. For the intersection The highest historical throughput per unit time For the intersection The actual average delay time, For the intersection The historical average delay time, To the intersection Gather at adjacent intersections. For the intersection Intersection The collaborative conflict index, Let be the weighting coefficients of each benefit item, and satisfy . These attributes comprehensively reflect the situation of the agent itself and the adjacent correlation, providing sufficient information for information interaction and scheme optimization. Each agent realizes information interaction through the vehicle-road cooperation network, ensuring that key information such as timing strategy, benefit value, and coordination conflict index is shared in real time, avoiding coordination failure caused by information silos. The traffic-phase difference coupling conflict index algorithm is used to calculate the coordination conflict index between intersections, accurately identifying timing conflict problems between intersections. The benefit function of each agent is constructed, taking into account traffic volume, delay time, and coordination conflict index to quantify the pros and cons of timing schemes. According to the first set of weight values corresponding to the peak period, the benefit item weight coefficient is set, which fits the traffic demand characteristics of the peak period and prioritizes the optimization of key indicators. Through the distributed negotiation mechanism, three rounds of iterative negotiation are carried out, with each iteration lasting less than 3 seconds, quickly advancing the timing scheme optimization. During this period, if the coordination conflict index between the core intersection of the main road and one of its adjacent intersections exceeds 0.3, triggering forced coordination, the main road green wave phase difference is prioritized to control the deviation within 10 seconds, ensuring smooth traffic on the main road during peak hours. Finally, the regional coordinated timing scheme is determined, achieving overall traffic efficiency improvement of multiple intersections.

[0026] S500, closed-loop iterative optimization: The regional coordinated timing scheme is issued to each intersection signal controller for execution, allowing the optimized timing scheme to be applied and effectively used in actual traffic control. Internet of Things devices collect real-time traffic data after the scheme is implemented, continuously monitor the implementation effect, and timely obtain traffic changes. The data is fed back to the regional data processing center to provide real data support for subsequent error analysis and parameter optimization. By comparing the errors between actual data and predicted data, the deviation between the timing scheme and actual traffic conditions is accurately identified. Based on the error, algorithm parameters are optimized, such as adjusting the weight coefficients in the traffic state comprehensive evaluation algorithm and the decay coefficients in the spatiotemporal decay flow prediction algorithm, continuously improving algorithm accuracy and optimizing the timing scheme. A closed-loop optimization system is formed to continuously improve the rationality and effectiveness of signal timing and adapt to dynamic changes in traffic flow.

[0027] In summary, in the cooperative optimization of early morning peak main road and adjacent intersections, multi-dimensional traffic data is collected through the Internet of Things device linkage system, providing accurate data sources for subsequent links after processing. Real-time and historical data are fused to determine traffic status and future trends based on traffic state comprehensive evaluation algorithm and spatiotemporal decay flow prediction algorithm. Based on this, a single-point initial timing scheme is generated using dynamic multi-factor green light duration calculation algorithm. Then, in the form of independent agents, the regional coordinated timing scheme is determined by combining the traffic-phase difference coupling conflict index algorithm, benefit function, and distributed negotiation mechanism. Finally, through closed-loop iterative optimization, the rationality of timing is continuously improved to effectively deal with the early morning peak traffic peak.

[0028] Example Two: Intersection optimization in flat peak period business district.

[0029] S100, multi-source data acquisition and processing: At a certain intersection in the urban business district and its two adjacent intersections, an Internet of Things equipment linkage system composed of geomagnetic sensors, AI high-definition cameras, and 5G edge computing gateways is deployed to form a real-time data acquisition network covering the core traffic nodes of the business district. The geomagnetic sensors can accurately capture the lane occupancy during the flat peak period when vehicles are scattered, avoiding misjudgment of timing caused by uneven vehicle distribution by collecting real-time vehicle presence status, vehicle passing quantity per unit time, instantaneous vehicle speed, and vehicle stay duration. The AI high-definition cameras can fully adapt to the characteristics of the business district with high demand for mixed traffic and pedestrians, fully grasp the pedestrian crossing rules, and prevent traffic safety problems caused by neglecting pedestrian demand by collecting vehicle types, vehicle queue length, and the number of queued vehicles in each direction, as well as the location, waiting state, and crossing state of pedestrians in pedestrian-intensive areas of the business district. After the data is processed by the 5G edge computing gateway, a standardized data set containing device identification, collection timestamp, data type, and corresponding numerical value is generated to ensure the uniformity of data formats collected by different devices, facilitate efficient fusion with historical data, and then transmit the standardized data set to the regional data processing center to provide complete and reliable basic data support for subsequent traffic flow trend prediction, as shown in FIG. 1. Figure 1

[0030] S200, traffic flow trend prediction: The standardized real-time traffic data set is fused with historical traffic flow database data, the past traffic flow rules during the flat peak period are combined, the interference of accidental fluctuations in real-time data on judgment is reduced, the spatial correlation characteristics of traffic flow are extracted, the correlation logic between the intersections and adjacent intersections in the business district is clarified, the one-sidedness of timing caused by isolated analysis of a single intersection is avoided, the traffic state comprehensive evaluation algorithm is used, and four key dimensions of traffic volume, average speed, queue length, and pedestrian crossing demand are comprehensively evaluated to evaluate the current traffic state in each direction of each intersection. The expression of the traffic state comprehensive evaluation algorithm is: which clearly presents the traffic characteristics of "few vehicles but many pedestrians" in each direction during the flat peak period, providing accurate state basis for subsequent timing adjustment. Through the time-space decay flow prediction algorithm, combined with historical data backtracking and adjacent intersection traffic flow correlation, the stability of traffic flow in the time dimension and the mutual influence of adjacent intersections in the space dimension during the flat peak period are fully considered to accurately predict the traffic flow trend in each direction within the next 10 minutes. The expression of the time-space decay flow prediction algorithm is: which successfully discovers the rule that the traffic flow in each direction is relatively stable, but the pedestrian crossing demand in the business district is high, providing forward-looking guidance for single-point dynamic timing generation to ensure that the timing scheme can adapt to pedestrian traffic demand in advance.

[0031] S300, single-point dynamic timing generation:​ Based on the traffic state evaluation results and short-term prediction trends, the traffic characteristics of "few cars and many people" are fully utilized and the future changes in people flow are predicted. Combined with real-time traffic data, the timing scheme is ensured to closely match the actual traffic demand of the commercial area during the flat peak period. A dynamic multi-factor green light duration calculation algorithm is adopted, which comprehensively considers the queue length, average speed, future traffic flow prediction value, and pedestrian crossing demand related factors. The green light duration of each direction at each intersection is dynamically adjusted. The expression of the dynamic multi-factor green light duration calculation algorithm is: Avoiding the traffic jam caused by the single car flow-based timing logic, considering the high demand for pedestrian crossing in the commercial area, the green light duration for pedestrian crossing is appropriately increased under the premise of ensuring normal vehicle traffic, effectively reducing the waiting time of pedestrians and reducing the risk of pedestrian-vehicle conflict. At the same time, the green light duration for vehicles in each direction is reasonably allocated to avoid resource waste caused by excessive green light duration in a certain direction. Finally, the initial timing scheme for each intersection is generated, providing an initial basis for the multi-intersection cooperative optimization that meets the characteristics of the commercial area.

[0032] S400, multi-intersection cooperative optimization: The initial timing scheme for each intersection is taken as input to provide a basic direction for cooperative optimization. At the same time, the three intersections are divided into independent agents, allowing each intersection to independently participate in timing optimization decisions based on its own characteristics of "few cars and many people", improving optimization efficiency and adaptability. The properties of each agent include the initial timing scheme for itself, the spatial position relationship with adjacent agents, historical traffic flow data, the cooperative conflict index set with adjacent agents calculated by the traffic-phase difference coupling conflict index algorithm, the expression of which is: The cooperative area composed of adjacent intersections has a radius of 1-2 kilometers, contains 3-5 intersections, and the distance between adjacent intersections does not exceed 800 meters. The expression of the benefit function is: These attributes comprehensively reflect the situation of the agent itself and its adjacent association, especially highlighting the impact of pedestrians on timing, providing sufficient information support for information interaction and scheme optimization. Each agent realizes information interaction through a vehicle-road cooperation network, ensuring real-time sharing of key information such as timing strategy, benefit value, and coordination conflict index, avoiding timing conflicts between adjacent intersections due to lack of information exchange, and ensuring the overall traffic order in the business district. The traffic-flow-phase-difference coupling conflict index algorithm is used to calculate the coordination conflict index between intersections, accurately identifying the timing conflict problems that may be caused by pedestrian crossing adjustments during the flat peak period. A benefit function is constructed and the weight coefficient is set according to the second set of weight values corresponding to the flat peak period, which meets the demand of "balancing pedestrian and vehicle traffic, reducing delay" during the flat peak period, prioritizing the balance between pedestrian safety and vehicle traffic efficiency. Through a 4-round iteration negotiation mechanism, each round of iteration takes less than 3 seconds, quickly advancing the timing scheme optimization, and there is no situation of coordination conflict index exceeding the standard during the period. Each agent gradually optimizes the timing scheme, and finally determines the regional coordinated timing scheme, and ensures that the phase difference deviation of adjacent intersection green wave bands does not exceed 10 seconds through consistency verification, ensuring smooth traffic for vehicles in the business district, while considering the pedestrian crossing demand, achieving overall balance of pedestrian and vehicle traffic.

[0033] S500, closed-loop iterative optimization: The regional coordinated timing scheme is issued to each intersection signal controller for execution, allowing the timing scheme adapted to the flat peak period of the business district to be applied and effectively used in actual traffic control, improving the efficiency and safety of pedestrian and vehicle traffic. Internet of Things devices collect real-time traffic data after the scheme is implemented, continuously monitor the implementation effect, focus on tracking key indicators such as pedestrian crossing waiting time and vehicle delay, and timely obtain traffic changes. The data is fed back to the regional data processing center to provide real and accurate data support for subsequent error analysis and parameter optimization. By comparing the errors between actual data and predicted data, the deviation between the timing scheme and the actual "few cars and many people" traffic situation is accurately found out, such as whether the pedestrian crossing green light duration still needs to be adjusted, whether the vehicle green light allocation is reasonable, etc. According to the error situation, optimize the algorithm parameters, such as adjusting the weight coefficient of pedestrian crossing demand in the traffic state comprehensive evaluation algorithm and the sensitive coefficient in the dynamic multi-factor green light duration calculation algorithm, etc. Continuously improve the adaptability of the algorithm to the traffic characteristics of the business district, and then optimize the timing scheme, forming a closed-loop optimization system, continuously improving the traffic efficiency and pedestrian crossing safety of the business district intersections, and adapting to the subtle dynamic changes of the flat peak period traffic flow.

[0034] In summary, during the optimization of the intersection in the commercial district during the flat peak period, the Internet of Things device is deployed to collect and adapt the data of people and vehicles in the commercial district, to ensure the availability of the data after processing, to fuse the data and extract the features, to master the traffic characteristics and trends of "few vehicles and many people" by using the traffic state comprehensive evaluation algorithm and the space-time decay flow prediction algorithm, to generate the single-point initial timing scheme considering the demand of people and vehicles by using the dynamic multi-factor green light duration calculation algorithm based on the information, to determine the collaborative timing scheme through the independent intelligent agent interaction, the related algorithm calculation and the distributed negotiation, to balance the traffic of people and vehicles in the commercial district through the closed-loop iterative optimization, and to improve the traffic efficiency and safety.

[0035] The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Although the present application has been disclosed as above with a preferred embodiment, it is not intended to limit the present application. Any person skilled in the art can make some changes or modifications to the above disclosed technical content to obtain equivalent embodiments with equivalent changes, without departing from the technical solution of the present application. Any modification, change, equivalent change and modification of the above embodiments based on the technical essence of the present application are still within the scope of the technical solution of the present application.

Claims

1. A method for optimizing intelligent signal timing at urban intersections based on vehicle-road cooperation, characterized in that, The specific steps of this method are as follows: S100, multi-source data acquisition and processing: Through an IoT device linkage system composed of geomagnetic sensors, AI high-definition cameras and 5G edge computing gateways, traffic data at intersections is collected in real time, processed locally by the 5G edge computing gateway and then transmitted to the regional data processing center. S200, Traffic Flow Situation Prediction: It integrates standardized real-time traffic datasets with data from historical traffic flow databases, extracts spatial correlation features of traffic flow, uses a comprehensive traffic condition assessment algorithm to assess the current traffic condition of intersections and related areas, and uses a spatiotemporal decay flow prediction algorithm to predict the traffic flow change trend within a preset time period in the future. S300, Single-point dynamic timing generation: Based on traffic condition assessment results and short-term forecast trends, combined with real-time traffic data, a dynamic multi-factor green light duration calculation algorithm is adopted to dynamically adjust the duration of intersection signals and generate a single-point initial timing scheme. S400, Multi-intersection Cooperative Optimization: Taking the initial timing scheme of each intersection as input, adjacent intersections are divided into independent intelligent agents. The flow-phase difference coupling conflict index algorithm is used to calculate the cooperative conflict index between intersections, and the revenue function of each intelligent agent is constructed. The regional cooperative timing scheme is determined through a distributed negotiation mechanism. S500, closed-loop iterative optimization: The regional coordinated timing scheme is distributed to the signal controllers at each intersection for execution. The actual traffic data after the scheme is executed is collected in real time through IoT devices and fed back to the regional data processing center. Based on the error between the actual data and the predicted data, the algorithm parameters are optimized to form a closed-loop optimization system.

2. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation as described in claim 1, characterized in that, In S100, the traffic data collected during multi-source data acquisition and processing includes: the vehicle presence status of each lane, the number of vehicles passing through per unit time, instantaneous vehicle speed, and vehicle dwell time collected by the geomagnetic sensor; the vehicle types in each direction, the vehicle queue length starting from the stop line, the number of vehicles in the queue, pedestrian positions, pedestrian waiting status, and pedestrian crossing status collected by the AI ​​high-definition camera; after the above data is processed by the 5G edge computing gateway, a standardized dataset containing device identifiers, acquisition timestamps, data types, and corresponding values ​​is formed.

3. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation as described in claim 1, characterized in that, In S200, the expression for the comprehensive traffic condition assessment algorithm in traffic flow situation prediction is: ,in, for time The comprehensive traffic condition index for each direction. for time Actual traffic flow in the direction for The highest historical traffic volume in this direction for time Average vehicle speed in the direction, for The design maximum speed of the direction, for time The actual queue length in the direction, for Maximum allowed queue length in any direction for time Normalized values ​​of pedestrian crossing demand in different directions. Let be the weight coefficients of each factor, and satisfy . .

4. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In S200, the expression for the spatiotemporal decay flow prediction algorithm in traffic flow situation prediction is: ,in, for Direction to the Future Traffic flow forecast at any given time For the size of the historical data backtracking window, The data sampling interval, For the first Weighting coefficients for historical data at each moment The time decay coefficient, To and The set of adjacent intersection directions that are related by traffic flow. for Adjacent directions at time Actual traffic volume Indicates and The directions of adjacent intersections that are related by traffic flow are the set of directions of adjacent intersections. The elements in for direction and Traffic flow correlation coefficient in different directions Spatial correlation weight coefficient, This indicates the time step number for historical data retrieval.

5. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In S300, during the single-point dynamic timing generation, the expression for the dynamic multi-factor green light duration calculation algorithm is as follows: ,in, for time Green light duration for each direction for time The actual queue length in the direction, for time Average vehicle speed in the direction, For safety redundancy time, for Traffic flow forecast for the next 5 minutes in the direction for The highest historical traffic volume in this direction The sensitivity coefficient for traffic flow prediction is optimized based on historical data. The result obtained through the traffic condition comprehensive assessment index algorithm time Comprehensive traffic condition index for directions for The maximum comprehensive traffic condition index for a given direction. The traffic condition impact coefficient is optimized based on historical data.

6. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In the S400 multi-intersection cooperative optimization, the independent intelligent agents are urban intersections. The attributes of each intelligent agent include: its own initial timing scheme, spatial position relationship with neighboring intelligent agents, historical traffic flow data, a set of cooperative conflict indices with neighboring intelligent agents calculated by the flow-phase difference coupling conflict index algorithm, and a benefit value calculated based on the benefit function. Each intelligent agent realizes the information exchange of timing strategy, benefit value and cooperative conflict index through the vehicle-road cooperative network.

7. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In the S400 multi-intersection cooperative optimization, the expression for the flow-phase difference coupling conflict index algorithm is: ,in, For the intersection Intersection The collaborative conflict index, For the intersection Intersection The actual green light phase difference For the intersection Intersection The optimal green band phase difference, , Intersections ,intersection Current traffic flow , Intersections ,intersection The historical peak traffic flow, the collaborative area formed by the adjacent intersections has a radius of 1-2 kilometers, includes 3-5 intersections, and the distance between adjacent intersections does not exceed 800 meters.

8. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In the S400 multi-intersection collaborative optimization, the expression for the revenue function is: ,in, Intelligent agents at intersections The profit value, For the intersection The actual traffic volume per unit time. For the intersection The highest historical throughput per unit time For the intersection The actual average delay time, For the intersection The historical average delay time, To the intersection Gather at adjacent intersections. For the intersection Intersection The collaborative conflict index, Let be the weighting coefficients of each benefit item, and satisfy . .

9. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In the S400 multi-intersection collaborative optimization, the weight coefficients of each benefit item are adjusted according to the time period: the first set of weight values ​​is used during peak hours, the second set of weight values ​​is used during off-peak hours, and the third set of weight values ​​is used during special event periods. The distributed negotiation mechanism iterates 3-5 times, with each iteration taking no more than 3 seconds; when the coordination conflict index... When this occurs, mandatory coordination is triggered to prioritize the green wave phase difference of the main road, with the deviation not exceeding 10 seconds.

10. The intelligent signal timing optimization method for urban intersections based on vehicle-road cooperation according to claim 1, characterized in that, In the S400 multi-intersection collaborative optimization, the specific steps for determining the regional collaborative timing scheme through a distributed negotiation mechanism are as follows: Each intersection agent calculates its initial revenue value based on its own initial timing scheme at each point. And share the initial strategy and revenue value through the vehicle-road cooperative network; Each agent iterates through the candidate timing schemes in its own policy space and calculates the expected reward value after switching to a candidate scheme. Candidate solutions that can improve their own benefits and whose impact on the benefits of neighboring agents is within an acceptable range are selected. Each agent sends the selected candidate solutions to neighboring agents through the vehicle-road cooperative network. The neighboring agents evaluate the impact of the solution on their own benefits. If both parties agree, the policy update is completed; otherwise, it reverts to the original solution. Repeat steps (2) and (3) until the policies of all agents no longer change in three consecutive iterations. The timing scheme combination at this point is the regional collaborative timing scheme. (5) Perform consistency verification on the regional coordinated timing scheme to ensure that the phase difference of the green wave band at adjacent intersections does not exceed the preset threshold of 10 seconds. If there is a deviation, fine-tune the timing parameters of the relevant intersections and finally determine the executable regional coordinated timing scheme.

Images