A traffic operation scheduling method based on unmanned aerial vehicle cooperation

Abstract

The application discloses a traffic operation scheduling method based on unmanned aerial vehicle cooperation, and belongs to the technical field of traffic operation scheduling. In order to solve the problem of reasonable intervention of unmanned aerial vehicles in traffic operation management, the application constructs an intersection inflow prediction model, constructs a passing capacity prediction model of the intersection, constructs a traffic pressure index for evaluating the traffic pressure of the intersection, establishes a relationship strength model of mutual influence of the flow between intersections according to traffic behavior logic, constructs an action efficacy model of the unmanned aerial vehicle on the intersection, scores the value of mutual intervention between the intersections, constructs a constraint feasibility model of the unmanned aerial vehicle, and optimizes the traffic operation scheduling of the unmanned aerial vehicle cooperation, including unmanned aerial vehicle scheduling optimization, intersection intervention priority optimization, signal control optimization and unmanned aerial vehicle information broadcast strength optimization. The application effectively solves the problem of untimely or slow response in the traditional traffic management process, and improves the utilization efficiency of resource allocation.

Description

Technical Field

[0001] This invention belongs to the field of traffic operation scheduling technology, and specifically relates to a traffic operation scheduling method based on unmanned aerial vehicle (UAV) collaboration. Background Technology

[0002] In recent years, with the increasing burden of urban traffic, timely monitoring and rapid response to dynamic traffic conditions have become core challenges in urban management. Traditional traffic systems mostly rely on fixed ground-based sensing equipment and static control strategies, making it difficult to capture abnormal changes or sudden congestion information in a timely manner. For urban traffic, which is highly dynamic in both time and space, this approach has significant limitations and lags. Drones, with their ability to flexibly switch between high and low altitudes, conduct high-frequency patrols, and achieve precise positioning, are gradually being introduced into traffic assistance management to undertake key tasks such as temporary sensing, emergency navigation, and traffic guidance.

[0003] The use of drones is mostly concentrated in video surveillance and traffic incident investigation, lacking a systematic scheduling model and intervention control mechanism, making it difficult to effectively alleviate local congestion. Existing systems do not integrate drone sensing with key elements such as traffic light control and traffic flow prediction, making it difficult to efficiently transform the acquired data into decision-making value. Summary of the Invention

[0004] The problem this invention aims to solve is to enable drones to make reasonable interventions in traffic operation management, and proposes a traffic operation scheduling method based on drone collaboration.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A traffic operation scheduling method based on UAV collaboration includes the following steps:

[0007] S1. Construct a model for predicting traffic flow at intersections;

[0008] S2. Construct a traffic capacity prediction model for the intersection;

[0009] S3. Based on the intersection inflow intensity predicted in step S1 and the intersection capacity predicted in step S2, construct a traffic pressure index to assess the traffic pressure at the intersection.

[0010] S4. Based on the intersection inflow intensity predicted in step S1, and according to traffic behavior logic, establish a relationship intensity model of the mutual influence of traffic flow between intersections;

[0011] S5. Considering the varying degrees of intervention, guidance, and monitoring capabilities of drones at different locations and observation angles on ground traffic nodes, construct a model of the effectiveness of drones at intersections.

[0012] S6. Based on the traffic pressure index obtained in step S3 and the effectiveness model of the UAV at the intersection obtained in step S5, the value score of mutual intervention between intersections is calculated.

[0013] S7. To determine the feasibility constraints for implementing UAV missions, a UAV constraint feasibility model is constructed.

[0014] S8. Optimize traffic operation scheduling in collaboration with drones, including drone scheduling optimization, intersection intervention priority optimization, signal control optimization, and drone information broadcasting intensity optimization.

[0015] Furthermore, in step S1, the historical average traffic flow at the intersection, the real-time measured traffic flow sampling data from the drone, and the spatially coupled traffic flow information from surrounding intersections are fused to construct an intersection inflow prediction model, the expression of which is:

[0016]

[0017] in, Let r be the predicted inflow intensity at intersection r at time t; The historical average traffic flow at intersection r is obtained from city OD data; The measured traffic flow at intersection r at time t was obtained by a drone. For the intersection The measured traffic flow at time t was obtained by a drone; Let t be the time when the vehicle departs from the intersection. The percentage of traffic entering intersection r is calculated from historical trajectory data, intersection turning flow detection data, or drone trajectory recognition results; Nr is the total number of intersections that are linked to intersection r in terms of accessibility. , , These are the historical traffic influence factor, the real-time traffic weight coefficient, and the neighborhood propagation influence coefficient, all of which are obtained through historical sample calibration or determined by expert experience.

[0018] Furthermore, step S2 dynamically estimates the intersection's capacity based on the vehicle density and phase ratio obtained by the drone, expressed as:

[0019]

[0020] in, Let r be the traffic capacity of intersection r at time t; Let r be the maximum theoretical throughput capacity of intersection r; The vehicle density at time t is obtained from the traffic monitoring platform; The percentage of green time at time t is obtained by the signal control system. The average vehicle spacing is obtained through simulation or actual measurement. It serves as a congestion-sensitive adjustment coefficient and is also used to offset the dimensions generated within the exponential function, thus making the exponential term dimensionless. This is a minimum constant corresponding to the percentage of green light time, used to prevent the denominator from being zero.

[0021] Furthermore, step S3 combines the predicted inflow intensity and instantaneous capacity of the intersection to construct a traffic pressure index, expressed as:

[0022]

[0023] in, Let r be the traffic pressure index at intersection r at time t.

[0024] Furthermore, step S4 considers factors such as vehicle route selection preferences, local traffic speed, and connection distance to establish a relationship strength model of the mutual influence of traffic flow between intersections, expressed as:

[0025]

[0026] in, Let r1 be the strength of the connectivity between the first intersection r1 and the second intersection r2 at time t. Let r1 be the predicted inflow per unit time at intersection; This is the road network connectivity coefficient, and it is also used for matching dimensions. The percentage of traffic flow from intersection r1 to r2 is determined by simulation or experimentation. The average speed from r1 to r2 at time t is obtained by the UAV data collection and measurement. The length from r1 to r2 is obtained from the GIS system.

[0027] Furthermore, step S5 considers the inverse square relationship between the spatial distance between the UAV and the target intersection, introduces the viewpoint cosine value to correct the effectiveness of the intervention direction, and adds a regularization parameter to prevent singularities, thus obtaining the UAV's effect on the intersection effectiveness model, expressed as:

[0028]

[0029] in, Let the effect weight of the drone u on the intersection r at time t be denoted as . This is the adjustment coefficient for maximum perceived intensity; This is the drone's top-down view, obtained by the drone monitoring platform, with a value range of 0-90 degrees. The sensing range radius of the drone is determined by the drone's parameters; The distance between the drone and the intersection is obtained from GNSS. To prevent the denominator from being zero, the minimum constant corresponding to the distance has the same dimensions as... Maintain consistency.

[0030] Furthermore, the specific implementation method of step S6 includes the following steps:

[0031] S6.1. Calculate the degree of linkage between each intersection and the traffic flow of adjacent intersections. The expression is:

[0032]

[0033] in, The spillover intervention value score of the drone u at intersection r1 at time t for the adjacent intersection r2; , The traffic pressure indices for r1 and r2 at time t are respectively. This is the pressure reduction factor at adjacent intersections, used to measure the spatial attenuation of the intervention effect; Let u be the influence factor of UAV on r1 at time t;

[0034] S6.2. The comprehensive score for the intervention value of UAV u at intersection r is calculated as follows:

[0035]

[0036] in, The comprehensive score of the intervention value of the drone u at intersection r at time t; The intervention value score for the area from intersection r to its adjacent intersection r2 at time t; This represents the total number of intersections that are linked to intersection r in terms of accessibility.

[0037] Furthermore, the expression for the UAV constrained feasibility model in step S7 is as follows:

[0038]

[0039] in, Let be an indicator function for whether the drone u can reach the intersection r at time t; This is a standard indicator function; it takes the value 1 when the condition inside the parentheses is true, and 0 otherwise. The remaining energy of the drone u at time t is obtained from the drone monitoring platform; The predicted energy consumption required for the UAV u to perform the minimum task cycle for the intersection r at time t includes outbound flight energy consumption, target area hovering energy consumption, and safe return energy consumption. This value is greater than 0 and is obtained from UAV simulation paths or experiments. The energy redundancy safety factor is set by technicians to ensure that the drone has sufficient margin for return and hovering.

[0040] Furthermore, the specific implementation method of step S8 includes the following steps:

[0041] S8.1. UAV scheduling optimization: Taking the comprehensive score of the intervention value between UAV and target intersection as the objective, and combining it with the UAV constraint feasibility model and corresponding task constraints, the scheduling variable is designed as a binary decision quantity to obtain the UAV scheduling optimization scheme;

[0042] S8.2. Intersection intervention priority optimization: Treat all drones dispatched to the target intersection as the service set of the target intersection, and sum the comprehensive scores of the intervention value of the drones to the target intersection to obtain the overall intervention priority score of the intersection.

[0043] S8.3. Signal Control Optimization: Drones are introduced to assist in optimizing the green light phase time at intersection r. The expression for the adjustment strategy is:

[0044]

[0045]

[0046] in, The adjustment amount for the percentage of green light time at intersection r; To maximize control over the adjustment range, it is set by professionals based on historical data, and is also used to match the units of measurement; The preset baseline intervention scoring constant is set by professionals based on historical data; The benchmark threshold for traffic pressure at the intersection is determined by the average pressure under historical smooth conditions. When the current pressure is greater than this threshold, the adjustment is positive, increasing the proportion of green light time on the main road; conversely, it is negative, releasing excess green light time to secondary phases. Update the signal phase ratio value; This is the original value of the signal phase ratio, obtained from the control system; , These are the lower and upper limits of the maximum green light phase ratio allowed for the main road at the intersection, determined by experts with reference to road geometry, to prevent extreme timing that could lead to traffic jams.

[0047] S8.4. Optimization of Drone Information Broadcasting Intensity: A guidance intensity output function is established by fusing the traffic pressure index and intervention priority score at intersections. This function is used to set the broadcast intensity for drones deployed at each specific intersection. The expression is:

[0048]

[0049] in, The drone broadcast guidance intensity at intersection r is defined as the corresponding broadcast control level value, which is used to map to preset volume levels, broadcast intervals, or broadcast frequencies. It is a broadcast signal strength conversion coefficient, determined by technicians in combination with broadcast parameters and historical data, and is also used for matching dimensions; The maximum allowable broadcast intensity limit, to prevent an overall limitless effect due to excessively high priority scores, is determined by technical personnel in conjunction with broadcast parameters.

[0050] The beneficial effects of this invention are:

[0051] The present invention discloses a traffic operation scheduling method based on UAV collaboration, which addresses the problem that UAVs only perceive but do not intervene in traffic operation management, resulting in chaotic scheduling and lack of control. It constructs a traffic control method that includes data fusion, node pressure assessment, and linkage between UAV scheduling and signal optimization. It can respond accurately in space, update dynamically in time, and optimize in a closed loop mechanism to achieve multi-objective collaborative control.

[0052] This invention discloses a traffic operation scheduling method based on UAV collaboration, which includes flow prediction, capacity analysis, traffic pressure evolution, intersection coupling propagation, scheduling control, and signal control. It can improve the real-time, refined, and responsiveness of urban traffic management. Historical data is integrated with UAV-measured data to dynamically predict vehicle inflow at each intersection and quantify its capacity. A traffic pressure index is defined by combining inflow and capacity, establishing the interaction relationships and coupling weights between intersections. An efficiency model for UAV intervention at intersections is established, and the value score and feasibility assessment of each UAV-intersection combination are performed. Based on this, dynamic optimal scheduling is achieved. Considering the needs of on-site traffic guidance, mechanisms for signal timing optimization and UAV information broadcasting are incorporated.

[0053] The present invention discloses a traffic operation scheduling method based on UAV collaboration, which can integrate the optimal intervention scheduling and signal guidance scheme of UAVs for urban traffic operation, effectively solve the problems of untimely or slow response speed in the traditional traffic management process, and improve the utilization efficiency of resource allocation. Attached Figure Description

[0054] Figure 1 This is a flowchart of a traffic operation scheduling method based on unmanned aerial vehicle (UAV) collaboration according to the present invention;

[0055] Figure 2 This is a comparison chart of the traffic capacity of the intersection before and after the adjustment of the present invention;

[0056] Figure 3 This is a comparison chart of the pressure indices before and after the adjustment in this invention. Detailed Implementation

[0057] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described specific embodiments are merely a part of the embodiments of the invention, and not all of them. The components of the specific embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations, and the invention may also have other embodiments.

[0058] Therefore, the following detailed description of specific embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected specific embodiments of the invention. All other specific embodiments obtained by those skilled in the art based on these specific embodiments without inventive effort are within the scope of protection of this invention.

[0059] To further understand the invention's content, features, and effects, the following specific embodiments are provided, along with accompanying drawings. Figure 1 -Appendix Figure 3 Detailed explanation is as follows:

[0060] Example 1:

[0061] A traffic operation scheduling method based on UAV collaboration includes the following steps:

[0062] S1. Construct a model for predicting traffic flow at intersections;

[0063] In traffic management, accurately grasping the vehicle inflow volume at each intersection at different times is fundamental for identifying changes in network conditions and implementing response strategies. Traditional methods rely excessively on historical OD matrices or fixed-point monitoring equipment, making it difficult to promptly understand sudden changes in on-site traffic conditions. Therefore, this model integrates historical average traffic flow, current real-time sampled data, and spatially coupled traffic information from surrounding nodes, and incorporates dynamic data streams from the high-frequency sensing capabilities of UAVs. The resulting comprehensive prediction formula considers both long-term trends and short-term changes, enabling UAVs to perceive changes in traffic supply and demand with higher timeliness and accuracy.

[0064] Furthermore, in step S1, the historical average traffic flow at the intersection, the real-time measured traffic flow sampling data from the drone, and the spatially coupled traffic flow information from surrounding intersections are fused to construct an intersection inflow prediction model, the expression of which is:

[0065]

[0066] in, Let r be the predicted inflow intensity at intersection r at time t; The historical average traffic flow at intersection r is obtained from city OD data; The measured traffic flow at intersection r at time t was obtained by a drone. For the intersection The measured traffic flow at time t was obtained by a drone; Let t be the time when the vehicle departs from the intersection. The percentage of traffic entering intersection r is calculated from historical trajectory data, intersection turning flow detection data, or drone trajectory recognition results; Nr is the total number of intersections that are linked to intersection r in terms of accessibility. , , These are the historical traffic influence factor, the real-time traffic weight coefficient, and the neighborhood propagation influence coefficient, all of which are obtained through historical sample calibration or determined by expert experience.

[0067] S2. Construct a traffic capacity prediction model for the intersection;

[0068] The capacity of an intersection is not a fixed value, but a dynamic variable influenced by various factors such as traffic density, traffic light ratios, and weather conditions. Traditional traffic management often uses fixed theoretical values ​​as the upper limit of capacity, which is difficult to reflect the current real road load level. This model incorporates vehicle density and phase ratio data obtained from drones to dynamically estimate the available capacity at different times, greatly improving the model's responsiveness to real-time changes in traffic flow. This invention does not arbitrarily choose exponential decay, but rather considers that when vehicle density approaches a critical value, the car-following effect between vehicles can cause a non-linear, avalanche-like decrease in capacity, which conventional linear decay cannot fit. Using this exponential model can accurately capture the phase transition characteristics of the congestion threshold, making it more realistically explanatory in congested scenarios.

[0069] Furthermore, step S2 dynamically estimates the intersection's capacity based on the vehicle density and phase ratio obtained by the drone, expressed as:

[0070]

[0071] in, Let r be the traffic capacity of intersection r at time t; Let r be the maximum theoretical throughput capacity of intersection r; The vehicle density at time t is obtained from the traffic monitoring platform; The percentage of green time at time t is obtained by the signal control system. The average vehicle spacing is obtained through simulation or actual measurement. It serves as a congestion-sensitive adjustment coefficient and is also used to offset the dimensions generated within the exponential function, thus making the exponential term dimensionless. This is a minimum constant corresponding to the percentage of green light time, used to prevent the denominator from being zero.

[0072] S3. Based on the intersection inflow intensity predicted in step S1 and the intersection capacity predicted in step S2, construct a traffic pressure index to assess the traffic pressure at the intersection.

[0073] To define a more effective real-time diagnostic indicator for traffic pressure based on inflow forecasts and capacity, and to assess the operational status of the road network, it's crucial to understand that relying solely on inflow volume or outflow capacity is insufficient to accurately determine congestion levels on a road segment. Some high-traffic areas may not experience significant congestion due to well-developed road infrastructure. A reasonable assessment of traffic pressure requires considering the deviation between supply and demand. By proposing a traffic pressure index that combines predicted inflow intensity with instantaneous capacity, it can reflect the most risky nodes in the urban road network and can also assist in drone dispatching and signal control.

[0074] Furthermore, step S3 combines the predicted inflow intensity and instantaneous capacity of the intersection to construct a traffic pressure index, expressed as:

[0075]

[0076] in, Let r be the traffic pressure index at intersection r at time t.

[0077] S4. Based on the intersection inflow intensity predicted in step S1, and according to traffic behavior logic, establish a relationship intensity model of the mutual influence of traffic flow between intersections;

[0078] Considering the intensity of mutual influence between intersections, a quantitative model of the traffic propagation mechanism between regions is established. In urban transportation systems, the mutual influence between nodes is significant; changes at main road intersections generally trigger a chain reaction on adjacent branch roads, leading to the spread of traffic congestion. This step's model, based on traffic behavior logic, uniformly models factors such as vehicle route selection preferences, local travel speeds, and connection distances, constructing a time-varying coupled weight function to capture the intensity characteristics of traffic flow interference between regions.

[0079] Furthermore, step S4 considers factors such as vehicle route selection preferences, local traffic speed, and connection distance to establish a relationship strength model of the mutual influence of traffic flow between intersections, expressed as:

[0080]

[0081] in, Let r1 be the strength of the connectivity between the first intersection r1 and the second intersection r2 at time t. Let r1 be the predicted inflow per unit time at intersection; This is the road network connectivity coefficient, and it is also used for matching dimensions. The percentage of traffic flow from intersection r1 to r2 is determined by simulation or experimentation. The average speed from r1 to r2 at time t is obtained by the UAV data collection and measurement. The length from r1 to r2 is obtained from the GIS system.

[0082] S5. Considering the varying degrees of intervention, guidance, and monitoring capabilities of drones at different locations and observation angles on ground traffic nodes, construct a model of the effectiveness of drones at intersections.

[0083] Furthermore, step S5 considers the inverse square relationship between the spatial distance between the UAV and the target intersection, introduces the viewpoint cosine value to correct the effectiveness of the intervention direction, and adds a regularization parameter to prevent singularities, thus obtaining the UAV's effect on the intersection effectiveness model, expressed as:

[0084]

[0085] in, Let the effect weight of the drone u on the intersection r at time t be denoted as . This is the adjustment coefficient for maximum perceived intensity; This is the drone's top-down view, obtained by the drone monitoring platform, with a value range of 0-90 degrees. The sensing range radius of the drone is determined by the drone's parameters; The distance between the drone and the intersection is obtained from GNSS. To prevent the denominator from being zero, the minimum constant corresponding to the distance has the same dimensions as... Maintain consistency. This function can dynamically represent the value density of drone resources under different deployment postures, assisting multiple drones in task decomposition and priority ranking when competing for tasks.

[0086] S6. Based on the traffic pressure index obtained in step S3 and the effectiveness model of the UAV at the intersection obtained in step S5, the value score of mutual intervention between intersections is calculated.

[0087] The interference relationships between urban intersections are complex. A traffic bottleneck often affects the surrounding areas connected to it. This step relies on calculating the degree of linkage between each intersection and the traffic flow of neighboring intersections, and weights and integrates the criticality, accessibility to intervention, and local congestion of the area to construct a critical comprehensive evaluation value that can be used for decision feedback.

[0088] Furthermore, the specific implementation method of step S6 includes the following steps:

[0089] S6.1. Calculate the degree of linkage between each intersection and the traffic flow of adjacent intersections. The expression is:

[0090]

[0091] in, The spillover intervention value score of the drone u at intersection r1 at time t for the adjacent intersection r2; , The traffic pressure indices for r1 and r2 at time t are respectively. This is the pressure reduction factor at adjacent intersections, used to measure the spatial attenuation of the intervention effect; Let u be the influence factor of UAV on r1 at time t;

[0092] S6.2. The comprehensive score for the intervention value of UAV u at intersection r is calculated as follows:

[0093]

[0094] in, The comprehensive score of the intervention value of the drone u at intersection r at time t; The intervention value score for the area from intersection r to its adjacent intersection r2 at time t; This represents the total number of intersections that are linked to intersection r in terms of accessibility.

[0095] S7. To determine the feasibility constraints for implementing UAV missions, a UAV constraint feasibility model is constructed.

[0096] To ensure the continuous reachability of UAVs during scheduling tasks, a mechanism for verifying the ratio of energy consumption to remaining range is introduced to constrain the feasibility of task implementation. This mechanism is expressed using an indicator function. By setting a lower limit for intervention time, it ensures that the selected deployment scheme has a practical basis for engineering execution. Especially when multiple tasks compete for limited resources simultaneously, eliminating options with insufficient power helps improve the reliability of system response and the stability of the scheduling system, preventing control failures caused by mid-task interruptions.

[0097] Furthermore, the expression for the UAV constrained feasibility model in step S7 is as follows:

[0098]

[0099] in, Let be an indicator function for whether the drone u can reach the intersection r at time t; This is a standard indicator function; it takes the value 1 when the condition inside the parentheses is true, and 0 otherwise. The remaining energy of the drone u at time t is obtained from the drone monitoring platform; The predicted energy consumption required for the UAV u to perform the minimum task cycle for the intersection r at time t includes outbound flight energy consumption, target area hovering energy consumption, and safe return energy consumption. This value is greater than 0 and is obtained from UAV simulation paths or experiments. The energy redundancy safety factor is set by technicians to ensure that the drone has sufficient margin for return and hovering.

[0100] S8. Optimize traffic operation scheduling in collaboration with drones, including drone scheduling optimization, intersection intervention priority optimization, signal control optimization, and drone information broadcasting intensity optimization.

[0101] Furthermore, the specific implementation method of step S8 includes the following steps:

[0102] S8.1. Drone Scheduling Optimization: Taking the comprehensive score of the intervention value between the drone and the target intersection as the objective, and combining it with the drone constraint feasibility model and corresponding task constraints, the scheduling variable is designed as a binary decision quantity, resulting in the drone scheduling optimization scheme, expressed as:

[0103]

[0104]

[0105]

[0106]

[0107]

[0108] in, Whether drone u is dispatched to intersection r, with a value of 0 or 1, is determined by the optimization solution; N is the number of intersections; W is the number of drones; The maximum number of drone services that intersection r can accept simultaneously;

[0109] Scheduling optimization requires identifying the most profitable resource allocation scheme within a limited number of drones and a finite response radius. This step involves constructing the optimization allocation logic based on the objective function, treating the intervention value score between the drone and the target intersection as the objective, and combining it with the task feasibility state and the unique task constraint to design a reasonable deployment scheme. Each drone is responsible for only one control objective; the scheduling variable is a binary decision quantity, and solving this function aims to achieve the optimal drone distribution at the current moment.

[0110] S8.2. Intersection Intervention Priority Optimization: Treat all drones dispatched to the target intersection as the service set of the target intersection, and sum the comprehensive intervention value scores of the drones to obtain the overall intervention priority score of the intersection, expressed as:

[0111]

[0112] in, The overall intervention priority score for intersection r;

[0113] Based on the optimal scheduling strategy, it is necessary to conduct an intervention value assessment for each intersection to support actions such as signal control adjustments and induced broadcast intensity. This step treats all drones dispatched to a target intersection as a service set for that intersection, summing up their intervention results to comprehensively reflect the current level of attention and criticality of governance at that location.

[0114] S8.3. Signal Control Optimization: Drones are introduced to assist in optimizing the green light phase time at intersection r. The expression for the adjustment strategy is:

[0115]

[0116]

[0117] in, The adjustment amount for the percentage of green light time at intersection r; To maximize control over the adjustment range, it is set by professionals based on historical data, and is also used to match the units of measurement; The preset baseline intervention scoring constant is set by professionals based on historical data; The benchmark threshold for traffic pressure at the intersection is determined by the average pressure under historical smooth conditions. When the current pressure is greater than this threshold, the adjustment is positive, increasing the proportion of green light time on the main road; conversely, it is negative, releasing excess green light time to secondary phases. Update the signal phase ratio value; This is the original value of the signal phase ratio, obtained from the control system; , These are the lower and upper limits of the maximum green light phase ratio allowed for the main road at the intersection, determined by experts with reference to road geometry, to prevent extreme timing that could lead to traffic jams.

[0118] In traditional traffic control, traffic light adjustment strategies are mostly dominated by fixed cycles or historical rules. In the event of sudden congestion, the response is often untimely. This invention, however, only considers adjusting the signal after on-site confirmation by a drone. The signal phase adjustment amount is determined based on the drone intervention priority score and traffic pressure deviation, and timing stability is ensured through upper and lower limit truncation. The hyperbolic tangent function tanh is used instead of a conventional linear proportional or step function because traffic signal timing is prone to a "deadlock" effect when approaching saturation. The tanh function exhibits an approximately linear and rapid response when the pressure deviation threshold is small, while providing smooth asymptotic constraints under extreme pressure or extreme idleness, effectively avoiding phase oscillations caused by traditional linear adjustment—that is, excessive green light leading to rapid congestion in the other direction. Simultaneously, the introduction of a pressure benchmark threshold and upper and lower limit truncation mechanisms allows the green light time to be both increased and decreased, conforming to the true zero-sum game logic of phase resources.

[0119] S8.4. Optimization of Drone Information Broadcasting Intensity: A guidance intensity output function is established by fusing the traffic pressure index and intervention priority score at intersections. This function is used to set the broadcast intensity for drones deployed at each specific intersection. The expression is:

[0120]

[0121] in, The drone broadcast guidance intensity at intersection r is defined as the corresponding broadcast control level value, which is used to map to preset volume levels, broadcast intervals, or broadcast frequencies. It is a broadcast signal strength conversion coefficient, determined by technicians in combination with broadcast parameters and historical data, and is also used for matching dimensions; The maximum allowable broadcast intensity limit, to prevent an overall limitless effect due to excessively high priority scores, is determined by technical personnel in conjunction with broadcast parameters.

[0122] Beyond specific scheduling and signal phasing, drones play another crucial role in traffic management: they can proactively guide traffic flow through methods such as voice announcements. This step utilizes an exponential function to control the induction intensity because drivers experience an information overload saturation effect when perceiving voice announcements. When traffic pressure is low, the announcement intensity increases approximately linearly with pressure to attract attention; when pressure is extremely high, the exponential term approaches 0, and the announcement intensity smoothly transitions and approaches the upper limit of saturation. This effectively prevents excessive volume or frequency from triggering driver panic and resistance.

[0123] By leveraging the quantitative model system based on traffic pressure, coupling propagation, intervention effectiveness, and signal adjustment feedback constructed in this embodiment, and through collaborative optimization of aspects such as UAV scheduling, signal phase control, and information broadcasting intensity control, a closed-loop linkage control from multi-source dynamic perception to traffic intervention is achieved. This method has good system scalability, engineering feasibility, and practical robustness, and is suitable for use in various complex road network structures and highly time-varying traffic scenarios, providing scalable technical support for urban intelligent traffic management.

[0124] The application examples of this embodiment are illustrated below:

[0125] In a certain road network, there are two intersections connected to intersection r1, namely r2 and r3. Drones are deployed for collaborative management, with observation periods from 8:00 AM to 10:00 AM. The historical average traffic volumes for intersections r1 to r3 are 1200 vehicles / hour, 1000 vehicles / hour, and 800 vehicles / hour, respectively.

[0126] Intersection R1 experienced reduced traffic speed and localized congestion. This method was adopted, optimizing traffic flow by selecting one drone for scheduling. The green light time percentage was adjusted by 0.2, and the drone's broadcast guidance intensity was 44.5 dB. After optimization, the traffic capacity of intersection R1 increased from 1200 vehicles / hour to 1600 vehicles / hour, and the pressure index decreased from 1 to 0.81.

[0127] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0128] Although this application has been described above with reference to specific embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of this application. In particular, as long as there is no structural conflict, the features in the specific embodiments disclosed in this application can be combined with each other in any way. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, this application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A traffic operation scheduling method based on unmanned aerial vehicle (UAV) collaboration, characterized in that, Includes the following steps: S1. Construct a model for predicting traffic flow at intersections; S2. Construct a traffic capacity prediction model for the intersection; S3. Based on the intersection inflow intensity predicted in step S1 and the intersection capacity predicted in step S2, construct a traffic pressure index to assess the traffic pressure at the intersection. S4. Based on the intersection inflow intensity predicted in step S1, and according to traffic behavior logic, establish a relationship intensity model of the mutual influence of traffic flow between intersections; S5. Considering the varying degrees of intervention, guidance, and monitoring capabilities of drones at different locations and observation angles on ground traffic nodes, construct a model of the effectiveness of drones at intersections. S6. Based on the traffic pressure index obtained in step S3 and the effectiveness model of the UAV at the intersection obtained in step S5, the value score of mutual intervention between intersections is calculated. S7. To determine the feasibility constraints for implementing UAV missions, a UAV constraint feasibility model is constructed. S8. Optimize traffic operation scheduling in collaboration with drones, including drone scheduling optimization, intersection intervention priority optimization, signal control optimization, and drone information broadcasting intensity optimization.

2. The traffic operation scheduling method based on UAV collaboration according to claim 1, characterized in that, In step S1, the historical average traffic flow at the intersection, the real-time measured traffic flow data from the drone, and the spatially coupled traffic flow information from surrounding intersections are fused to construct an intersection inflow prediction model, the expression of which is: ； in, Let r be the predicted inflow intensity at intersection r at time t; The historical average traffic flow at intersection r is obtained from city OD data; The measured traffic flow at intersection r at time t was obtained by a drone. For the intersection The measured traffic flow at time t was obtained by a drone; Let t be the time when the vehicle departs from the intersection. The percentage of traffic entering intersection r is calculated from historical trajectory data, intersection turning flow detection data, or drone trajectory recognition results; Nr is the total number of intersections that are linked to intersection r in terms of accessibility. , , These are the historical traffic influence factor, the real-time traffic weight coefficient, and the neighborhood propagation influence coefficient, all of which are obtained through historical sample calibration or determined by expert experience.

3. The traffic operation scheduling method based on UAV collaboration according to claim 2, characterized in that, Step S2 dynamically estimates the intersection's capacity based on vehicle density and phase ratio data obtained from the drone, expressed as: ； in, Let r be the traffic capacity of intersection r at time t; Let r be the maximum theoretical throughput capacity of intersection r; The vehicle density at time t is obtained from the traffic monitoring platform; The percentage of green time at time t is obtained by the signal control system. The average vehicle spacing is obtained through simulation or actual measurement. It serves as a congestion-sensitive adjustment coefficient and is also used to offset the dimensions generated within the exponential function, thus making the exponential term dimensionless. This is a minimum constant corresponding to the percentage of green light time, used to prevent the denominator from being zero.

4. The traffic operation scheduling method based on UAV collaboration according to claim 3, characterized in that, Step S3 combines the predicted inflow intensity and instantaneous capacity of the intersection to construct a traffic pressure index, expressed as: ； in, Let r be the traffic pressure index at intersection r at time t.

5. A traffic operation scheduling method based on UAV collaboration according to claim 4, characterized in that, Step S4 considers factors such as vehicle route selection preferences, local traffic speed, and connection distance to establish a relationship strength model for the mutual influence of traffic flow between intersections, expressed as: ； in, Let r1 be the strength of the connectivity between the first intersection r1 and the second intersection r2 at time t. Let r1 be the predicted inflow per unit time at intersection; This is the road network connectivity coefficient, and it is also used for matching dimensions. The percentage of traffic flow from intersection r1 to r2 is determined by simulation or experimentation. The average speed from r1 to r2 at time t is obtained by the UAV data collection and measurement. The length from r1 to r2 is obtained from the GIS system.

6. A traffic operation scheduling method based on unmanned aerial vehicle (UAV) collaboration according to claim 5, characterized in that, Step S5 considers the inverse square relationship between the spatial distance between the UAV and the target intersection, introduces the viewpoint cosine value to correct the effectiveness of the intervention direction, and adds a regularization parameter to prevent singularities, thus obtaining the UAV's effect on the intersection effectiveness model, expressed as: ； in, Let the effect weight of the drone u on the intersection r at time t be denoted as . This is the adjustment coefficient for maximum perceived intensity; This is the drone's top-down view, obtained by the drone monitoring platform, with a value range of 0-90 degrees. The sensing range radius of the drone is determined by the drone's parameters; The distance between the drone and the intersection is obtained from GNSS. To prevent the denominator from being zero, the minimum constant corresponding to the distance has the same dimensions as... Maintain consistency.

7. A traffic operation scheduling method based on UAV collaboration according to claim 6, characterized in that, The specific implementation method of step S6 includes the following steps: S6.

1. Calculate the degree of linkage between each intersection and the traffic flow of adjacent intersections. The expression is: ； in, The spillover intervention value score of the drone u at intersection r1 at time t for the adjacent intersection r2; , The traffic pressure indices for r1 and r2 at time t are respectively. This is the pressure reduction factor at adjacent intersections, used to measure the spatial attenuation of the intervention effect; Let u be the influence factor of UAV on r1 at time t; S6.

2. The comprehensive score for the intervention value of UAV u at intersection r is calculated as follows: ； in, The comprehensive score of the intervention value of the drone u at intersection r at time t; The intervention value score for the area from intersection r to its adjacent intersection r2 at time t; This represents the total number of intersections that are linked to intersection r in terms of accessibility.

8. A traffic operation scheduling method based on unmanned aerial vehicle (UAV) collaboration according to claim 7, characterized in that, The expression for the UAV constrained feasibility model in step S7 is: ； in, Let be an indicator function for whether the drone u can reach the intersection r at time t; This is a standard indicator function; it takes the value 1 when the condition inside the parentheses is true, and 0 otherwise. The remaining energy of the drone u at time t is obtained from the drone monitoring platform; The predicted energy consumption required for the UAV u to perform the minimum task cycle for the intersection r at time t includes outbound flight energy consumption, target area hovering energy consumption, and safe return energy consumption. This value is greater than 0 and is obtained from UAV simulation paths or experiments. The energy redundancy safety factor is set by technicians to ensure that the drone has sufficient margin for return and hovering.

9. A traffic operation scheduling method based on unmanned aerial vehicle (UAV) collaboration according to claim 8, characterized in that, The specific implementation method of step S8 includes the following steps: S8.

1. UAV scheduling optimization: Taking the comprehensive score of the intervention value between UAV and target intersection as the objective, and combining it with the UAV constraint feasibility model and corresponding task constraints, the scheduling variable is designed as a binary decision quantity to obtain the UAV scheduling optimization scheme; S8.

2. Intersection intervention priority optimization: Treat all drones dispatched to the target intersection as the service set of the target intersection, and sum the comprehensive scores of the intervention value of the drones to the target intersection to obtain the overall intervention priority score of the intersection. S8.

3. Signal Control Optimization: Drones are introduced to assist in optimizing the green light phase time at intersection r. The expression for the adjustment strategy is: ； ； in, The adjustment amount for the percentage of green light time at intersection r; To maximize control over the adjustment range, it is set by professionals based on historical data, and is also used to match the units of measurement; The preset baseline intervention scoring constant is set by professionals based on historical data; The benchmark threshold for traffic pressure at the intersection is determined by the average pressure under historical smooth conditions. When the current pressure is greater than this threshold, the adjustment is positive, increasing the proportion of green light time on the main road; conversely, it is negative, releasing excess green light time to secondary phases. Update the signal phase ratio value; This is the original value of the signal phase ratio, obtained from the control system; , These are the lower and upper limits of the maximum green light phase ratio allowed for the main road at the intersection, determined by experts with reference to road geometry, to prevent extreme timing that could lead to traffic jams. S8.

4. Optimization of Drone Information Broadcasting Intensity: A guidance intensity output function is established by fusing the traffic pressure index and intervention priority score at intersections. This function is used to set the broadcast intensity for drones deployed at each specific intersection. The expression is: ； in, The drone broadcast guidance intensity at intersection r is defined as the corresponding broadcast control level value, which is used to map to preset volume levels, broadcast intervals, or broadcast frequencies. It is a broadcast signal strength conversion coefficient, determined by technicians in combination with broadcast parameters and historical data, and is also used for matching dimensions; The maximum allowable broadcast intensity limit, to prevent an overall limitless effect due to excessively high priority scores, is determined by technical personnel in conjunction with broadcast parameters.

Images