A method and system for guiding highway traffic flow based on unmanned aerial vehicles (UAVs)

Abstract

This invention discloses a method and system for guiding highway traffic flow based on unmanned aerial vehicles (UAVs), including a data acquisition and processing subsystem, a cloud-based traffic flow simulation subsystem, and a guidance subsystem. By dividing the frequently congested sections of the highway (ramp-ends, service areas) and their upstream areas into control zones, data is collected, and a cloud-based simulation model of the corresponding control zone is constructed. Simultaneously, the model is updated in real-time according to a certain time window. After the system identifies congestion, it calculates a speed limit strategy using the cloud model and sends the speed limit strategy to an airport. The UAV departs from the designated airport, arrives at a designated location on the road segment, and provides speed limit prompts to guide traffic flow, thereby alleviating traffic conditions on congested sections. It can also be used for incidental accident alerts.

Description

Technical Field

[0001] This invention relates to a method for guiding traffic flow on highways, and more particularly to a method and system for guiding traffic flow on highways based on unmanned aerial vehicles (UAVs). Background Technology

[0002] With the increase in car ownership, traffic congestion on highways is becoming increasingly serious, especially near special sections (ramp-offs, service areas), where concentrated traffic flow often leads to severe congestion during certain time periods. Traditional traffic flow guidance systems often react slowly and cannot respond to sudden traffic situations in real time. Drones, due to their mobility and flexibility, can perform speed limit control at any location, making them a new option for solving such problems. Patent CN202311188583 discloses a traffic control system and method for preventing traffic congestion, including a road anomaly unit, a climate anomaly unit, and a historical comparison unit. The road anomaly unit reports information on urban road maintenance and repair via a network. When maintenance and construction are carried out near the intersection of an expressway underpass, the maintenance and construction information is sent and reported. By collecting climate and road condition information in advance, and judging that congestion will occur at the underpass intersection later, the road condition preprocessing unit disperses parked vehicles near the underpass intersection. However, this plan can only prevent traffic congestion caused by static factors such as road maintenance, and cannot respond in a timely manner to sudden and occasional dynamic factors such as car accidents. Furthermore, the plan can only provide warnings to vehicles crossing underpasses and cannot guarantee a guiding effect, thus it has its shortcomings. Summary of the Invention

[0003] Purpose of the invention: To address the above problems, this invention proposes a method and system for guiding highway traffic flow based on unmanned aerial vehicles (UAVs), which can ensure the real-time performance of the road guidance system, improve the timeliness of guidance, and enhance the guidance effect.

[0004] Technical Solution: The technical solution adopted in this invention is a highway traffic flow guidance method based on unmanned aerial vehicles (UAVs), comprising the following steps:

[0005] S1. Designate the area from the ramp or service area to the upstream expressway main toll station as the control zone, and divide the control zone into three control sections in sequence: congestion monitoring area, traffic flow guidance area and toll station area.

[0006] S2. Based on the road geometry data and historical traffic flow data of the control section, construct a traffic flow simulation model for the control area; obtain the current real-time traffic flow information of the control section within the time window t, upload it to the traffic flow simulation model, and update the traffic flow simulation model in real time; the traffic flow simulation model outputs traffic status information at each location in the control area.

[0007] S3. Based on the traffic status information at each location in the control area, determine the real-time congestion status of the congestion monitoring area; based on the traffic status information at each location in the control area and the real-time traffic flow data of the toll station area, predict the traffic status information of the congestion monitoring area through a traffic flow simulation model, and then predict the congestion status of the congestion monitoring area in the future based on the predicted traffic status information of the congestion monitoring area.

[0008] S4. When congestion is detected or predicted in the congestion monitoring area, the speed limit strategy for the traffic flow guidance area is calculated using machine learning algorithms.

[0009] S5. Send the speed limit policy task to the UAV airport control center. The UAVs belonging to the airport will receive the task, take off, cruise to the task target location, and then issue a speed limit warning.

[0010] The congestion assessment in step S3 includes: traffic status information at each location in the control area; calculating the average vehicle speed and average headway in the traffic flow guidance area; and using a threshold method to assess the congestion based on the average vehicle speed and average headway.

[0011] The speed limit strategy for the traffic flow guidance area calculated by the machine learning algorithm in step S4 is a reinforcement learning algorithm. The environmental state variables of the reinforcement learning algorithm include the traffic state information of each location in the control area. The action space is the speed limit value of each location in the traffic flow guidance area. The reward function is negatively correlated with the traffic congestion status of the congestion monitoring area and the traffic pressure generated by the traffic flow of the upstream toll station in the traffic flow guidance area. The more severe the congestion, the greater the traffic flow, the greater the traffic pressure, and the lower the reward value.

[0012] The speed limit strategy for traffic flow guidance areas is calculated using a reinforcement learning algorithm, including the following steps:

[0013] S41. Obtain traffic status information at various locations within the control area through a real-time updated traffic flow simulation model;

[0014] S42. Initialize the reinforcement learning environment, action space, and reward function, wherein the environment state includes traffic state information of each location in the control area obtained through a real-time updated traffic flow simulation model;

[0015] S43. Train the agent using a deep Q-network (DQN), calculate the action and corresponding reward function value based on the real-time traffic conditions, and select the optimal speed limit location and speed limit value based on the reward function value.

[0016] S44. Input the speed limit location and speed limit value decided by the agent into the traffic flow simulation model, obtain feedback from the traffic flow simulation model, and update the agent.

[0017] S45. Repeat steps S43-S44 until the model converges, and output the rate limiting policy corresponding to the current agent.

[0018] The reward function takes the following specific form:

[0019] R t =-α·T congestion +β·V average -γ·P pressure

[0020] Among them, T congestion V represents the congestion time in the congestion monitoring area. average P represents the average speed of traffic flow in the traffic flow guidance area. pressure The traffic pressure generated by the traffic flow at the upstream toll station on the traffic flow guidance area is represented by α, β, and γ, which are the corresponding weight coefficients for each item.

[0021]

[0022] D delay Indicates upstream delay, referring to the additional waiting time for vehicles queuing to enter the toll station due to congestion or speed limits; V input Traffic flow in the upstream toll station area;

[0023] D delay =T actual -T expected

[0024] Where T actual T is the actual time required for a vehicle to travel a certain distance R near the toll station. expected It is the estimated travel time for route R under conditions of no congestion.

[0025] The agent is trained using a Deep Q-Network (DQN), and the Q-value update formula during training is:

[0026]

[0027] Where Q represents the current state S t Next, execute action a t Q value, S t and S t+1 Let a represent the current state and the next state, respectively. t and a t+1 R represents the currently executing action and the next action, respectively. t This indicates that at time step t, the agent is in state S. t Perform action a t The immediate reward obtained afterward, α0 is the learning rate, γ0 is the discount factor, and Q is the value of Q. target The Q-value is calculated for the target network.

[0028] After receiving the mission, the airport's drones take off and cruise to the mission target location, then issue a speed limit warning, including the following steps:

[0029] S51. The airport receives the speed limit policy and performs task allocation, airspace allocation, and path planning for idle drones.

[0030] S52. The UAV receives the mission, goes through the takeoff decision process, and ascends to the target altitude. The UAV then cruises to the target location according to the preset path plan.

[0031] S53. The drone arrives at the mission location, automatically calibrates its positioning, lands above the highway guardrail, and performs a capture and fixation operation; the drone's guidance information screen rises and displays speed limit information.

[0032] This invention proposes a UAV-based highway traffic flow guidance system, comprising a data acquisition and processing subsystem, a cloud-based traffic flow simulation subsystem, a memory, and a guidance subsystem. The cloud-based traffic flow simulation subsystem includes a processor and a computer program stored in the memory and executable on the processor. The data acquisition and processing subsystem includes three data acquisition modules: a first data acquisition module located near the ramp entrance or service area of ​​the congestion monitoring area; a second data acquisition module located in the traffic flow guidance area; and a third data acquisition module located near the main toll station. When the processor in the cloud-based traffic flow simulation subsystem executes the computer program, it implements the UAV-based highway traffic flow guidance method.

[0033] This invention proposes a computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the aforementioned UAV-based highway traffic flow guidance method.

[0034] This invention proposes a computer program product, including a computer program and / or instructions, which, when executed by a processor, implement the aforementioned UAV-based highway traffic flow guidance method.

[0035] Beneficial Effects: Compared with existing technologies, this invention has the following advantages: This invention divides frequently congested sections of highways (ramp-ends, service areas) and upstream areas into control zones, collects data, constructs cloud-based simulation models of the corresponding control zones, and updates the models in real-time according to a specific time window, ensuring the real-time performance of the guidance system. Simultaneously, after the system identifies congestion, it calculates a speed limit strategy through the cloud model, sends the speed limit strategy to the airport, and a drone departs from the designated airport, arrives at the designated location on the road segment, and provides speed limit prompts to guide traffic flow, thereby alleviating traffic conditions on congested sections. It can also be used for incidental accident alerts, improving the timeliness and effectiveness of guidance. The guidance method and system of this invention can not only handle traffic congestion caused by static factors such as road maintenance, but also respond promptly to sudden and occasional dynamic factors such as traffic accidents. This invention utilizes the mobility and flexibility of drones to achieve real-time monitoring and guidance of highway traffic flow. Through precise speed limit prompts, congestion is effectively alleviated, and highway traffic efficiency is improved. This system has broad application prospects and significant socio-economic benefits. Attached Figure Description

[0036] Figure 1 The specific division of the road control area;

[0037] Figure 2 This is a flowchart illustrating the speed limit strategy calculation for the traffic flow guidance area described in this invention.

[0038] Figure 3 This is a schematic diagram of the UAV-based highway traffic flow guidance system described in this invention.

[0039] Figure 4 This is a schematic diagram of the overall workflow of the UAV-based highway traffic flow guidance system described in this invention. Detailed Implementation

[0040] The technical solution of the present invention will be further described below with reference to the accompanying drawings and embodiments.

[0041] Example 1

[0042] The UAV-based highway traffic flow guidance method of this invention utilizes the mobility of UAVs to guide traffic flow (speed limit reminders) on highways based on data collection and traffic modeling. The method includes the following steps:

[0043] S1. The area from the ramp (service area) to the upstream highway toll station is designated as a control zone, which is further divided into a congestion monitoring area, a traffic flow guidance area, and a toll station area.

[0044] Depending on their location, highway toll stations can be divided into mainline toll stations and ramp toll stations. Mainline toll stations are generally located on the main line of the highway, where traffic volume is high, traffic speed is fast, and high capacity requirements are needed; ramp toll stations are located on highway ramps or along the road, where traffic volume is relatively low. The target control area is defined on the main line of the highway.

[0045] Congestion Monitoring Area: A congestion monitoring area will be designated 200 meters upstream of the ramp entrance (service area). Cameras and geomagnetic induction coils will be deployed there for congestion identification and detection. The specific length of the congestion monitoring area will be based on historical traffic flow data.

[0046] Toll Station Area: The area upstream of the highway toll station at the ramp (service area) and extending 150 meters downstream is designated as the toll station area. This is the area where vehicles accelerate after exiting the toll station, and no traffic flow guidance measures are implemented here. Vehicle flow data is obtained using toll station information. For comfortable acceleration, the vehicle's acceleration is 1 m / s². According to regulations, when there are three or more lanes in the same direction, the minimum speed in the leftmost lane is 110 km / h, the middle lane is 90 km / h, and the rightmost lane is 60 km / h. The minimum acceleration distance for a vehicle exiting the toll station is approximately 138 meters; therefore, 150 meters is taken as the acceleration zone distance.

[0047] Traffic flow guidance area: The section of road between the toll plaza area and the congestion monitoring area serves as the primary implementation area for drone speed limit operations. By guiding and controlling traffic flow in this area, congestion in the downstream congestion monitoring area can be alleviated.

[0048] like Figure 1 The diagram shows the specific division of the road control area, including the congestion monitoring area, traffic flow guidance area, and toll station area. The first data acquisition module 1 in the congestion monitoring area is located near the ramp entrance (service area) and is used to collect traffic flow, average speed, average queue length, and headway near the ramp entrance (service area). It includes a video data acquisition unit and a geomagnetic coil detection auxiliary unit. The second data acquisition module 2 is located in the traffic flow guidance area and is mainly a video data acquisition unit used to collect the average speed of traffic flow in the traffic flow guidance area. The third data acquisition module 3 is located at the toll station entrance and exit and is a toll station gantry information unit used to collect the unit time traffic flow input status at the toll station.

[0049] S2. Acquire road geometry data and historical traffic flow data for the controlled section to initially construct a traffic flow simulation model for the controlled section. Define a time window length, acquire real-time traffic flow information for the current controlled section according to time window t, upload it to the cloud simulation model, and perform real-time updates to the simulation model.

[0050] The traffic flow simulation model is built using SUMO. The model is then validated using specific evaluation methods and parameters. Successful validation indicates the simulation model is reliable.

[0051] The specific steps for real-time updating of the simulation model are as follows:

[0052] S21. Obtain road geometry data, historical traffic flow data, and average speed data for the controlled section, initially construct a traffic flow simulation model, and use simulation parameters such as traffic flow volume, average speed, vehicle parameters, and vehicle following model to verify the model and ensure its feasibility and reliability.

[0053] S22. Using pre-deployed cameras as the main data acquisition unit, and through image recognition technology, acquire data such as vehicle type, traffic flow, average speed, and headway in the control area (focusing on the congestion monitoring area). At the same time, use geomagnetic induction coils to collect relevant data as a supplement to the main data acquisition unit, to repair missing data and correct abnormal data.

[0054] S23. Use the toll station gantry system to obtain the vehicle flow data input from the upstream of the current control area, and use it as the vehicle input for the simulation model.

[0055] S24. Update the data according to the above steps according to the time window t, and input the simulation model to update the model.

[0056] S3. Based on the real-time traffic flow information in step S2, calculate the average vehicle speed and average headway in the congestion identification area, and determine the congestion status based on the average speed and average headway; simultaneously, predict the congestion status in the congestion monitoring area based on the traffic flow input in the toll station area. The specific steps are as follows:

[0057] S31. Calculate the average vehicle speed and average headway using the collected traffic flow data. The specific formula is as follows:

[0058]

[0059] Get the speed v of the i-th vehicle within a certain time period. i The number of vehicles is n.

[0060] h i =t i -t i-1

[0061]

[0062] The time series of a vehicle passing through a certain point: t1, t2, t3, ..., t n , where t iIf the time it takes for the i-th car to pass a certain point is , then the headway between the i-th car and the (i-1)-th car is h. i =t i -t i-1 .

[0063] S32. The congestion identification method in the congestion monitoring area is the threshold method. Congestion is identified by setting thresholds for the average vehicle speed and average headway. and When a congestion is detected, the system is deemed to be in a congested state. Here, v1 and v2 represent the average speed in the traffic flow guidance area and the average speed in the congestion monitoring area, respectively, calculated from traffic flow information collected by data acquisition unit 2 and data acquisition unit 1. Simultaneously, the traffic conditions in the congestion identification area are predicted based on upstream traffic flow conditions.

[0064] S4. When congestion is detected or predicted in the congestion monitoring area, a speed limit strategy for the traffic flow guidance area is calculated using a real-time updated traffic flow simulation model and machine learning methods. For example... Figure 2 The diagram shows the calculation process for speed limit policies in traffic flow guidance areas, illustrating the specific steps involved in data collection, state vector construction, model training and optimization, and real-time policy generation. The specific steps for generating the speed limit policy are as follows:

[0065] S41. Obtain traffic status information at various locations within the control area through a real-time updated traffic flow simulation model.

[0066] S42. Initialize the reinforcement learning environment, where the environment state includes traffic status information at each location in the control area.

[0067] S43. Define the action space and reward function. The action space is the speed limit value at different locations in the traffic flow guidance area. The reward function is related to the traffic congestion status of the ramp congestion monitoring area and the traffic pressure on the traffic flow guidance area from the upstream toll station.

[0068] S44. Train the agent using the reinforcement learning algorithm Deep Q-Network (DQN) to select the optimal speed limit location and speed limit value based on real-time traffic conditions.

[0069] S45. Apply the speed limit position and speed limit value decided by the agent to the simulation model, obtain simulation feedback, and update the agent.

[0070] S46. Repeat steps S4-S5 until the model converges, meaning the agent can effectively alleviate congestion at ramp entrances under different traffic conditions.

[0071] S47. Apply the trained agent to the actual traffic management system and adjust the speed limit strategy in real time.

[0072] The specific method for generating speed limiting policies is as follows:

[0073] 1. Environment modeling and state representation.

[0074] In this invention, the environment is defined as a dynamic transportation system comprising highway sections, vehicles, road infrastructure, weather conditions, and real-time traffic flow. Environmental state S t It provides a comprehensive picture of the current traffic situation, specifically including the following information:

[0075] Congestion status: Reflects traffic density, queue length, or average vehicle speed in congested areas.

[0076] Average vehicle speed in the traffic flow guidance area: the average speed of vehicles in the current area affected by weather conditions.

[0077] Real-time traffic input status of upstream toll stations: vehicle traffic data passing through highway toll stations.

[0078] Weather conditions: Factors such as rain and snow that may affect driving safety and speed.

[0079] State vector S t The above indicators are combined and then input into the deep Q-network model for processing.

[0080] 2. Action definition.

[0081] Speed ​​limiting policy through action a t Let A represent the action set A, which contains all possible speed limit decisions. Each action a... t Includes two dimensions:

[0082] Speed ​​limit location: Speed ​​limits are implemented on specific road sections within the traffic flow guidance area at a certain distance from the congestion identification area.

[0083] Speed ​​limit: The specific speed limit set, such as 30km / h, 50km / h, 70km / h, 90km / h, etc.

[0084] In this invention, the action space is designed as a set of discrete values, covering multiple common speed limiting strategy combinations.

[0085] 3. Reward function design.

[0086] Reward function R t This is used to quantify the effectiveness of each speed limit strategy, with the goal of maximizing traffic congestion relief through dynamic adjustments to speed limits. The reward function is designed considering the following factors:

[0087] Congestion mitigation effect: Measured by reducing the travel time of vehicles in the congestion identification area.

[0088] Traffic efficiency: quantified by increasing the average speed of traffic flow in the controlled area and reducing traffic congestion.

[0089] Safety: Consider the impact of weather conditions on safe driving and avoid the risk of traffic accidents caused by excessive speed limits.

[0090] The reward function can be expressed in the following form:

[0091] R t =-α·T congestion +β·V average -γ·P pressure

[0092] Among them, T congestion The congestion time within the congestion monitoring area is used to represent the traffic congestion status within that area. Other quantities that reflect congestion status can be used as substitutes. V average P represents the average speed of traffic flow in the traffic flow guidance area. pressure The traffic pressure generated by the toll station's traffic input in the downstream control area is represented by α, β, and γ, which are the corresponding weighting coefficients.

[0093]

[0094] Upstream delay D delay This usually refers to the extra waiting time for vehicles queuing to enter a toll station due to congestion or speed limits, measured in seconds (s) or minutes (min).

[0095] D delay =T actual -T expected

[0096] Where T actual T is the actual time required for a vehicle to travel a certain distance. expected This is the estimated travel time under conditions of no congestion.

[0097] V input The flow input status of the upstream toll station area is directly proportional to the traffic pressure. That is, a larger flow input means greater traffic pressure on the downstream control area. Therefore, it should be treated as a negative penalty term in the reward function.

[0098] 4. Construction of deep Q-network model.

[0099] This invention employs a Deep Q-Network (DQN) to construct a rate-limiting policy generation model. A Q-Network is a multi-layer neural network used to estimate state-action pairs (S...). t ,a t The Q value of ) is the value of the state S given by the given state S. t Next, execute action a tThe accumulated rewards that can be obtained in the future. Specifically, these include:

[0100] Input layer: Receives state vector S t As input.

[0101] Hidden layers: Several fully connected layers, using ReLU as the activation function.

[0102] Output layer: Outputs each possible action a t The Q value.

[0103] In addition, to improve the stability of training, this invention uses a target network, which is a Q-network with the same structure as the main network but with slower parameter updates, to calculate the target Q-value.

[0104] 5. Model training.

[0105] Model training is completed through the following steps:

[0106] Experience replay mechanism: Store each interaction data with the environment in the experience replay buffer, and randomly sample small batches of samples for training, thereby breaking the correlation between samples and reducing the instability of model training.

[0107] Q-value update: Updated based on the Bellman equation, i.e.:

[0108]

[0109] Where α is the learning rate, γ is the discount factor, and Q... target The Q-value is calculated for the target network.

[0110] Exploration and exploitation strategy: An ∈-greedy strategy is adopted. During the model training process, the agent selects a random action with probability ∈ (exploration) and selects the action with the highest current Q value with probability 1-∈ (exploitation), thereby balancing the exploration and exploitation strategies in the unknown environment.

[0111] 6. Rate limiting strategy generation and application.

[0112] Through training the DQN model described above, this invention can achieve real-time traffic state S t Generate the optimal speed limiting strategy a t The generated speed limit strategy includes specific speed limit locations and values, and is implemented through a dynamic traffic control system. The implementation effectiveness is monitored through a feedback mechanism and continuously used to optimize the model, thereby improving the accuracy and effectiveness of speed limit decisions.

[0113] S5. Send the speed limit policy task to the nearest drone airport. After receiving the task, the drones belonging to the airport take off and cruise to the target location to issue speed limit warnings. Specifically, this includes the following steps:

[0114] S51. The airport receives the speed limit policy and performs task allocation, airspace allocation, and path planning for idle drones.

[0115] S52. The UAV receives the mission, goes through the takeoff decision process, and ascends to the target altitude. The UAV then cruises to the target location according to the preset path plan.

[0116] S53. The drone arrives at the mission location, automatically calibrates its positioning, lands above the highway guardrail, and performs a capture and fixation operation; the drone's guidance information screen rises and displays speed limit information.

[0117] S6. After the speed limit policy is implemented, real-time feedback on the policy implementation is provided, and changes in congestion in the congestion monitoring area are monitored. If the policy implementation effect is not good after several time windows, the policy is updated, and the drone issues a warning to change the road speed limit.

[0118] In one embodiment, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the above-described UAV-based highway traffic flow guidance method.

[0119] In one embodiment, a computer program product is provided, including a computer program / instructions that, when executed by a processor, implement the aforementioned drone-based highway traffic flow guidance method.

[0120] Those skilled in the art will understand that embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product embodied on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0121] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0122] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0123] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0124] Example 2

[0125] The UAV-based highway traffic flow guidance system described in this invention sets up a data acquisition system at frequently congested road sections (usually ramp entrances). Using devices such as cameras and geomagnetic induction coils, it collects road traffic flow information and information from highway toll gantry upstream of the ramp entrance, providing effective data for traffic flow prediction in congested sections. Simultaneously, a traffic modeling system is used to model key road sections and updates the model in real time based on the collected information. When congestion is detected, the system calculates the speed limit location and information, and then uses a UAV upstream of the congested section to provide precise speed limit prompts, thereby alleviating congestion and improving traffic efficiency in the shortest possible time. Figure 3 The diagram shows the structure of the UAV-based highway traffic flow guidance system of the present invention, which includes a data acquisition and processing subsystem, a cloud-based traffic flow simulation subsystem, and a guidance subsystem.

[0126] The data acquisition and processing subsystem includes a data acquisition module and a congestion identification module. The data acquisition module comprises a video data acquisition unit, a geomagnetic coil detection auxiliary unit, and a toll station gantry information unit. The video data acquisition unit collects traffic flow information within the control area, including vehicle type, flow rate, and headway. The geomagnetic coil detection auxiliary unit performs data gap repair and anomaly correction on the video data acquisition data. The toll station gantry information unit collects traffic flow input from upstream areas into the control area. The congestion identification module uses algorithms to calculate average traffic flow speed and average headway based on the traffic flow data collected by the data acquisition module, and then uses a threshold method to determine congestion.

[0127] The cloud-based traffic flow simulation subsystem comprises a simulation module and a speed limit strategy module. The simulation module performs initial construction, model verification, and model updates of the traffic flow simulation model for the control area. The simulation system acquires geometric data and historical traffic flow data of the control area to initially construct the traffic flow simulation model, while simultaneously performing model calibration to ensure its reliability. Once calibration is successful, the system receives real-time traffic flow data for the control area from the data acquisition module and updates the model according to a set time window. Upon receiving congestion information, the speed limit strategy module, based on the simulation model, uses a special machine learning algorithm to calculate the optimal speed limit and optimal speed limit location, aiming to minimize the duration of congestion. The optimal speed limit location is located anywhere within the traffic flow guidance area, and the minimum speed limit does not exceed the real-time congestion speed in the congestion monitoring area. After the drone implements the speed limit strategy, the simulation subsystem evaluates the strategy's effectiveness using updated data after several time windows. If the effect is unsatisfactory, the speed limit strategy is updated; if the effect is satisfactory, the strategy remains in effect until the congestion is resolved.

[0128] The guidance subsystem includes a drone airport and drone modules. The airport involves planning its layout and infrastructure construction. Airport site selection aims to enable drones to complete missions across multiple control areas within a minimal cruising range, considering factors such as economy, efficiency, and safety. Drone airport facilities include a drone control system, drone energy system, air traffic management system, operations monitoring center, intelligence center, and maintenance center. The drone control system controls automated takeoff and landing, autonomous cruising, mission execution, and return-to-factory recovery. The drone energy system supplies power to the drones, ensuring their endurance. The air traffic management system primarily handles airspace allocation and drone cruising path planning. The operations monitoring center monitors the overall airport operation and ensures drone safety during operations. The intelligence center handles data transmission between the cloud-based big data model and the airport, information exchange between airport departments, and information transmission between drones and the airport. The maintenance center is responsible for drone maintenance and support, ensuring drone stability and reliability. The speed limit strategy module sends the speed limit strategy to the airport's intelligence center, which then forwards it to various airport departments to complete operational preparations. Based on the strategy information, the air traffic management department selects an appropriate altitude and plans the drone's cruise path. Upon receiving the airport information and undergoing a takeoff decision process, the drone ascends to the target altitude and autonomously takes off and cruises to the speed limit strategy target location according to the predetermined path. After precise positioning, it lands above the highway roadside guardrail, secures itself, and raises its guidance information display screen to show speed limit warnings, reminding drivers to slow down in time, thus completing the traffic flow guidance task. The drone system includes an energy system, positioning system, flight control system, data transmission system, operational tools, and guidance information display system. The navigation and flight control systems are responsible for the drone's automatic cruise along the predetermined path. The operational tools and guidance information display system are responsible for the specific guidance task implementation. After the drone reaches the designated location, the operational tools accurately identify the roadside guardrail and perform a grabbing operation, securing the drone to the guardrail. The data transmission system is responsible for receiving strategy update information.

[0129] This invention can be applied to multiple ramps and congested sections of highways. By flexibly adjusting speed limits, it can effectively alleviate traffic congestion and improve the overall traffic flow efficiency. For example, the "congestion monitoring area" at a highway ramp is 200 meters long; the section between the ramp and the upstream highway toll station is designated as a "traffic flow guidance area," which is 15 kilometers long.

[0130] like Figure 4 The diagram shows the overall workflow of the UAV-based highway traffic flow guidance system described in this invention. It includes the following processes:

[0131] Input relevant information to complete system preparation: The process involves inputting road geometry data and historical traffic flow information to build an initial simulation model, determining evaluation criteria, and calibrating the model using parameters such as road traffic flow, average speed, and vehicle parameters. Through animation checks and accuracy calibration, the model verification is successful, demonstrating its reliability. Once the model is built, the system enters operational mode. The data acquisition and processing subsystem transmits data to the cloud-based traffic flow subsystem at specific time windows for model updates. The second data acquisition module 2 collects traffic flow information from the traffic flow guidance area and calculates the average speed. The first data acquisition module 1 collects traffic flow information from the congestion monitoring area; some data is missing, so data is repaired using data collected by the geomagnetic coil detection auxiliary unit. Simultaneously, the congestion identification module performs congestion identification. The toll station gantry information unit of the third data acquisition module 3 sends past traffic flow information from the toll station to the cloud-based model for congestion prediction. The congestion identification module in the data acquisition and processing subsystem calculates... and (Where v1 and v2 are the average speeds of the traffic flow guidance area and the congestion monitoring area, respectively, calculated from the traffic flow information collected by data acquisition unit 2 and data acquisition unit 1.) The data acquisition and processing subsystem identifies congestion and uploads the congestion information to the cloud simulation model subsystem, where the speed limit module runs the algorithm.

[0132] Upstream toll station traffic flow input: 1500 vehicles / hour; current vehicle speed in the traffic flow guidance area: 85 km / h. Based on system settings, the system determines this to be a congested state. After detecting congestion, the simulation model uses the DQN algorithm in real time to generate a speed limit strategy. The DQN state input includes the following variables: S t = [{Congestion Status}, {Upstream Traffic Input}, {Average Speed}, {Weather Conditions}]). Congestion Status: Congested. Upstream Traffic Input: 1500 vehicles / hour. Average Speed: 85 km / h. Weather Conditions: Sunny. Based on the current status, combined with historical data and traffic prediction, the DQN algorithm generates the following speed limit strategy: Speed ​​Limit Location: [Speed ​​limit value: 90 km / h, 5 km upstream of the ramp entrance].

[0133] The cloud-based traffic flow simulation subsystem selects the optimal airport and sends the generated speed limit policy to the mission execution airport. Upon receiving the mission information (including latitude, longitude, and speed limit), the airport selects an idle drone, allocates airspace, and plans the drone's flight path. The drone accepts the mission, ascends to the target airspace after takeoff decision, cruises to the mission location along the predetermined path, descends to a position above the roadside guardrail, captures a fixed value on the guardrail, and simultaneously raises its guidance information display screen to show the speed limit information. According to the speed limit policy, the drone hangs a 90 km / h speed limit warning sign 5 km upstream of the congested section, reminding drivers to gradually slow down, thereby reducing the overall speed of traffic upstream of the congested section. Feedback is received after several time windows: 5 minutes after the speed limit policy is implemented, the average vehicle speed in the congestion monitoring area returns to 70 km / h, the upstream traffic flow stabilizes at 1400 vehicles / hour, and the congestion at the ramp has been alleviated. Policy adjustment: If the system detects continued congestion, the speed limit will continue to be adjusted until the congestion is completely relieved.

[0134] This embodiment uses a precisely controlled speed limit strategy to guide the speed of traffic flow upstream of congested areas, thereby reducing the average speed of traffic flow and avoiding congestion caused by traffic flow concentrating at ramp entrances, effectively alleviating road congestion.

Claims

1. A method for inducing traffic flow on a highway based on a UAV, characterized in that, It comprises the following steps: S1, the ramp or service area to the upstream highway main toll station is divided into a control area, and the control area is sequentially divided into a congestion monitoring area, a traffic flow induction area and a toll station area three control sections; S2, according to the road geometric data and historical traffic flow data of the control section, a traffic flow simulation model of the control area is constructed; the current real-time traffic flow information of the control section in the time window t is obtained and uploaded to the traffic flow simulation model for real-time updating of the traffic flow simulation model; the traffic flow simulation model outputs the traffic state information of each position in the control area; S3, according to the traffic state information of each position in the control area, the real-time congestion condition of the congestion monitoring area is judged; according to the traffic state information of each position in the control area and the real-time traffic flow data of the toll station area, the traffic state information of the congestion monitoring area is predicted through the traffic flow simulation model, and then the congestion condition of the congestion monitoring area in a future period of time is predicted according to the predicted traffic state information of the congestion monitoring area; S4, when the congestion condition of the congestion monitoring area is detected or predicted, the speed limit strategy of the traffic flow induction area is calculated through a machine learning algorithm; The speed limit strategy of the traffic flow induction area calculated by the machine learning algorithm adopts a reinforcement learning algorithm, the environmental state quantity of the reinforcement learning algorithm includes the traffic state information of each position in the control area, the action space is the speed limit value of each position in the traffic flow induction area, the reward function is negatively related to the traffic pressure on the traffic flow induction area caused by the traffic congestion condition of the congestion monitoring area and the traffic flow of the upstream toll station, the more serious the congestion condition is, the larger the traffic flow is, the larger the traffic pressure is, and the lower the reward value is; The reinforcement learning algorithm is adopted to calculate the speed limit strategy of the traffic flow induction area, which comprises the following steps: S41, the traffic state information of each position in the control area is obtained through the real-time updated traffic flow simulation model; S42, the reinforcement learning environment, the action space and the reward function are initialized, wherein the environmental state includes the traffic state information of each position in the control area obtained through the real-time updated traffic flow simulation model; S43, an intelligent agent is trained by using a deep Q network DQN, the action and the corresponding reward function value are calculated according to the real-time traffic state, and the optimal speed limit position and speed limit value are selected according to the reward function value; S44, the speed limit position and speed limit value decided by the intelligent agent are input into the traffic flow simulation model, the feedback of the traffic flow simulation model is obtained, and the intelligent agent is updated; S45, steps S43-S44 are repeated until the model converges, and the speed limit strategy corresponding to the current intelligent agent is output; The specific form of the reward function is: ； wherein, is the congestion time of the congestion monitoring area, is the average speed of the traffic flow of the traffic flow induction area, represents the traffic pressure on the traffic flow induction area generated by the traffic flow of the upstream toll station, and α, β, and γ are the corresponding weight coefficients. ； represents the upstream delay, which takes into account the extra waiting time of vehicles in the stretch of road queuing to enter the toll station due to congestion or speed limits; is the traffic volume for the upstream toll station area; ； wherein is the actual time taken by the vehicle to pass through a certain stretch of road R near the toll station, is the estimated transit time through the stretch of road R in the absence of congestion. The Q value updating formula in the training of the intelligent agent by using the deep Q network DQN is: ； Where Q represents the current state. Next action Q value, and These represent the current state and the next state, respectively. and These represent the currently executing action and the next action, respectively. This indicates that at time step t, the agent is in state t. Execute action The immediate reward obtained afterward, where α0 is the learning rate and γ0 is the discount factor. The Q-value calculated for the target network; S5, the speed limit strategy task is sent to the unmanned aerial vehicle airport control center, and the unmanned aerial vehicles belonging to the airport take off and cruise to the task target location to perform speed limit warning. 2.The highway traffic flow induction method based on the UAV according to claim 1, characterized in that: The judgment of the congestion condition in step S3 comprises: the traffic state information of each position in the control area, the average traffic flow speed and the average headway of the traffic flow induction area are calculated; and the threshold method is adopted to judge the congestion condition according to the average traffic flow speed and the average headway. 3.The highway traffic flow induction method based on the UAV according to claim 1, characterized in that: The airport-owned unmanned aerial vehicle receives a task, takes off, cruises to a task target location, and performs speed limit warning, including the following steps: S51, the airport receives a speed limit strategy, and assigns a task, an airspace, and a path to an idle unmanned aerial vehicle; S52, the unmanned aerial vehicle receives the task, goes through a take-off decision process, and rises to a target height; the unmanned aerial vehicle cruises to a target location according to a preset path planning; S53, the unmanned aerial vehicle reaches the task location, automatically calibrates positioning, lands above a roadside guardrail of the expressway, and implements grabbing and fixing; an induction information screen of the unmanned aerial vehicle rises and displays speed limit information.

4. A highway traffic flow induction system based on unmanned aerial vehicle, comprising a data acquisition and processing subsystem, a cloud traffic flow simulation subsystem memory and an induction subsystem, the cloud traffic flow simulation subsystem comprising a processor and a computer program stored on the memory and executable on the processor, characterized in that, The data acquisition and processing subsystem includes three data acquisition modules, a first data acquisition module is arranged near a ramp or a service area in a congestion monitoring area, a second data acquisition module is arranged in a road section in a traffic flow induction area, and a third data acquisition module is arranged near a main toll station; the processor in the cloud traffic flow simulation subsystem implements the unmanned aerial vehicle-based expressway traffic flow induction method according to any one of claims 1 to 3 when the computer program is executed.

5. A computer-readable storage medium having stored thereon a computer program, characterized in that, The computer program is executed by the processor to implement the unmanned aerial vehicle-based expressway traffic flow induction method according to any one of claims 1 to 3.

6. A computer program product comprising computer programs and / or instructions, characterized in that, The computer program and / or instructions are executed by the processor to implement the unmanned aerial vehicle-based expressway traffic flow induction method according to any one of claims 1 to 3.

Images