Intelligent Traffic Signal Control Method and System Based on Traffic Flow Prediction

Abstract

The application relates to an intelligent traffic signal control method and system based on traffic flow prediction, and belongs to the technical field of intelligent traffic. Traffic vehicle video streams and traffic flow data are collected by a high-definition camera and a traffic flow sensor group, different vehicles and their speeds, directions and motion trajectories are identified through the traffic vehicle video streams, special vehicles are judged, traffic violation behaviors and road abnormal conditions are monitored according to the traffic vehicle video streams and the traffic flow data, a traffic prediction model is established according to traffic vehicle video stream and traffic flow historical record data, each lane traffic flow, vehicle position, vehicle category, lane line information, vehicle motion trajectory, traffic violation, road abnormal condition, future time period traffic flow value and vehicle speed of each intersection are acquired in real time from a server, and different signal lamp control schemes are generated according to different vehicles. The application combines vehicle types, traffic violations, road abnormalities and the like to design a signal control scheme, and realizes accurate traffic congestion dredging.

Description

Technical Field

[0001] This invention relates to an intelligent traffic signal control method and system based on traffic flow prediction, belonging to the field of intelligent traffic management technology. Background Technology

[0002] With the acceleration of urbanization and the rapid growth of motor vehicle ownership, traffic congestion has become a major challenge for urban operations. To address this issue, an increasing number of technologies are being applied to the transportation sector. However, traditional traffic signal control systems still suffer from the following problems: fixed signal cycles struggle to respond in real-time to dynamic changes in traffic flow, leading to increased congestion during peak hours and wasted resources during off-peak hours; the reliance on single magnetic induction coil data fails to effectively utilize video and other data for traffic situation identification; and the limited handling of special vehicles hinders effective overall coordination of local road conditions. With the increasing maturity and widespread application of artificial intelligence technology, exploring an intelligent traffic signal control method and system that deeply integrates data-driven and intelligent decision-making technologies to overcome the core challenges of poor dynamic adaptability and insufficient multi-intersection coordination in traditional traffic signal control, thereby achieving efficient traffic flow management and optimized resource allocation, is urgently needed. Summary of the Invention

[0003] The purpose of this invention is to overcome the shortcomings of existing technologies and provide an intelligent traffic signal control method based on traffic flow prediction, which can effectively alleviate peak congestion.

[0004] The technical solution adopted in this invention is as follows:

[0005] The intelligent traffic signal control method based on traffic flow prediction includes the following steps:

[0006] S1. Receive traffic vehicle video streams and traffic flow data through a high-definition camera and traffic flow sensor array, and preprocess the received data;

[0007] S2. Identify different vehicles, their speed, direction, and trajectory through traffic video streams, determine whether they are special vehicles, and send the vehicle type, speed, direction, and trajectory information detected at each intersection to the server;

[0008] The Yolov8 model is used to detect targets in each frame of the video, obtaining vehicle position and category information. A U-Net model is used for lane detection. Vehicle speed is calculated by first establishing a mapping between pixel coordinates and actual distances based on the video's shooting parameters and the calibration information of the actual scene. Then, the displacement of each vehicle between consecutive frames is calculated, and combined with the video's frame rate, the displacement is finally converted into actual speed. The direction of motion is determined based on the change in the vehicle's center point position within consecutive frames. The vehicle trajectory is achieved by integrating the trained Yolov8 model with the DeepSORT algorithm to realize the detection and tracking of dynamic targets.

[0009] S3. Monitor traffic violations and road anomalies based on traffic vehicle video streams and traffic flow data. Traffic violations include running red lights and speeding, while road anomalies include sudden traffic incidents and road construction. Send the traffic violations and road anomalies to the server. The server records and saves the sudden traffic incidents and road construction information for each road, and records and sends penalty information for vehicles that violate traffic rules.

[0010] The red light violation detection first checks the traffic light status (red / green / yellow), then determines whether the vehicle entered the stop line area during the red light period based on the detected vehicle movement trajectory, and finally makes a judgment based on the red light status and whether the vehicle entered the stop line area.

[0011] Speeding detection combines the measured vehicle speed with the speed limit threshold in traffic flow data to determine whether speeding has occurred.

[0012] Detection of sudden traffic incidents includes emergency braking detection and wrong-way driving detection. Emergency braking detection is achieved by calculating the rate of change of trajectory acceleration (e.g., rate of change of velocity > 5 m / s²). 2 The judgment, reverse driving detection, is to determine whether the angle between the trajectory direction and the lane line direction is >90°;

[0013] Road construction inspection uses methods such as identifying static obstacles and assessing the characteristics of the construction area to make judgments;

[0014] S4. Build a traffic prediction model based on traffic vehicle video streams and historical traffic flow data. Use the traffic prediction model to predict traffic flow and vehicle speed for future time periods based on real-time data, and send the predictions to the server.

[0015] The vehicle movement trajectories in S2 and the daily traffic flow data in step S1 are processed to complete the timestamp alignment to form sequence data. The sequence data is used as input, and the timestamp, weekday, and holiday are used as input signals. The traffic flow and vehicle speed are used as outputs. Then, a long short-term memory network is used to build a model, and the network is trained through a loss function to generate a traffic prediction model.

[0016] S5. The system retrieves real-time traffic flow, vehicle location, vehicle type, lane information, vehicle trajectory, traffic violations, road anomalies, and future traffic flow and speed data for each lane at each intersection from the server. Based on the different vehicles, it generates different traffic light control schemes and sends instructions to each intersection for dynamic traffic light control.

[0017] Based on the vehicle category determination in step S2, if there are special vehicles, the following procedure will be implemented:

[0018] (1) Calculate the weight of each road segment based on the traffic state parameters of each road segment in the road network, and then establish an optimal path selection model on the road network connectivity graph abstracted by graph theory. Discrete particle swarm optimization algorithm is used to discretize the particle optimization process, and the optimal path is solved with the goal of minimizing the time for vehicles to reach their destination. Finally, the preset driving route with the shortest travel time is planned.

[0019] (2) Based on the determined planned route, associate all the traffic light nodes that are passed through, calculate the switching time of the traffic lights according to the route travel time, and provide feedback to the traffic lights according to the real-time location of the special vehicles, so as to achieve green lights all the way for the special vehicles.

[0020] (3) Receive new obstacle events and traffic flow information from the server in real time, update the traffic status parameters of each road segment in real time, dynamically adjust the driving route, and update the traffic light control scheme in real time according to the new route.

[0021] Based on the vehicle category determination in step S2, if no special vehicles exist, proceed with the following plan:

[0022] The optimal green light time and phase switching timing are calculated using a multi-objective optimization algorithm. The multi-objectives include minimizing the average waiting time, maximizing road capacity, and reducing environmental pollution. Furthermore, the optimization parameters are dynamically adjusted based on real-time traffic flow and abnormal road conditions.

[0023] In the above method, the data preprocessing in step S1 involves frame processing and information labeling of traffic vehicle video stream data, noise removal using a mean filtering algorithm for traffic flow data, and outlier identification and removal using a median absolute deviation algorithm.

[0024] The special vehicles mentioned in step S2 include ambulances, police cars, fire trucks, and rescue vehicles. The specific process for calculating speed in step S2 is as follows:

[0025] (1) First, the traffic video stream is converted into images frame by frame, the Yolov8 model is trained, and the category and bounding box coordinates of each vehicle are output. The trained Yolov8 model is integrated with the DeepSORT algorithm, and a unique ID is given to each vehicle in the video stream. The U-Net model is used for lane line detection.

[0026] (2) Based on the video shooting parameters and the ratio of pixels to actual distance calculated by shooting objects of known size, the pixel coordinates are converted into three-dimensional world coordinates using the camera intrinsic parameter matrix, thereby establishing the mapping relationship between pixel coordinates and actual distance.

[0027] (3) For consecutive frames with the same vehicle ID, calculate the pixel displacement of the center point of the bounding box of each consecutive frame. Use the mapping relationship in (2) to convert the pixel displacement into the actual displacement. Combine the frame rate to calculate the speed. Speed ​​= actual displacement × frame rate.

[0028] Step S3 introduces a multi-target tracking algorithm to ensure accurate differentiation of each vehicle's behavior even when multiple vehicles appear in the frame simultaneously. The identified behaviors and traffic anomalies are then sent to the server, which makes corresponding decisions based on the information. For detecting sudden traffic events, the velocity change rate (acceleration) a(t) between consecutive frames is calculated for each vehicle's trajectory.

[0029]

[0030] If |a(t)|>5m / s 2 Output emergency braking event records, including vehicle ID, time, and acceleration value;

[0031] Extract the direction vector of each vehicle in step S2 and lane direction vector Calculate the angle θ between the trajectory direction and the lane line direction:

[0032]

[0033] If θ > 90°, output the reverse driving event record, including vehicle ID, time, and angle value.

[0034] Step S5 uses a multi-objective optimization algorithm to calculate the optimal green light time and phase switching timing. A phase refers to the time period during which a group of non-conflicting traffic flows (such as straight, left-turn, and right-turn) simultaneously receive the green light right-of-way. The process is as follows:

[0035] (1) Define the objective function:

[0036] Minimize average waiting time

[0037] Maximize road capacity

[0038] Reduce environmental pollution

[0039] Where i is the vehicle number, N is the total number of vehicles, and t wait,iLet t be the waiting time for vehicle i. free,i Let t be the free passage time of vehicle i. idle,i e is the idling time of vehicle i. i For vehicle emission factors;

[0040] (2) Initialize the population and randomly generate N solutions, each solution containing the optimal green light time T. g Given the phase switching timing ΔT, calculate the objective function value (f1, f2, f3) for each solution. Then, classify the solutions into different levels based on the quality of the objective values. Use selection, crossover, and mutation to generate the next generation population. Repeat the iteration until convergence and output the optimal solution set, where 30s ≤ T. g ≤120s, ΔT≥5s;

[0041] (3) Use the traffic prediction model in step S4 to predict traffic flow and predict the traffic flow in the future time period. If the traffic flow is predicted to surge, extend the green light time in advance.

[0042] (4) If the traffic flow is detected to be greater than the traffic flow threshold (the traffic flow threshold for main roads is 1200 vehicles / hour / lane, and the traffic flow threshold for other main roads is 600 vehicles / hour / lane), the green light time will be extended immediately. If no vehicles are detected to pass through during the green light period, the green light time will be shortened. If a sudden traffic incident is detected, the previous intersection will switch to red light and the current intersection will switch to green light, prohibiting vehicles from entering the incident area and prioritizing the diversion of traffic flow at surrounding intersections.

[0043] (5) Adjust and optimize T g And ΔT are sent to the traffic signal controller.

[0044] Another objective of this invention is to provide an intelligent traffic signal control system based on traffic flow prediction that implements the above-mentioned method, including monitoring equipment, traffic light groups, a data processing center, and a traffic signal control service center. The monitoring equipment includes a high-definition camera group and a traffic flow sensor group. The data processing center (corresponding cabinet) includes a preprocessing module, a vehicle video stream processing module, a traffic violation and road accident detection module, a traffic prediction model construction and prediction module, and a communication module. The traffic signal control service center includes a server and a display. The server includes a signal control scheme generation module, a control module, and a data storage and transmission module.

[0045] The monitoring equipment collects traffic vehicle video streams and traffic flow data through high-definition cameras and traffic flow sensor arrays. The data processing center's preprocessing module preprocesses the received traffic vehicle video streams and traffic flow data. The vehicle video stream processing module identifies different vehicles, their speeds, directions, and trajectories based on the traffic vehicle video streams, and determines whether they are special vehicles. The traffic violation and road accident detection module monitors traffic violations and road anomalies based on the traffic vehicle video streams and traffic flow data. Traffic violations include running red lights and speeding, while road anomalies include sudden traffic incidents and road construction. The traffic prediction model building and prediction module builds a traffic prediction model based on the traffic vehicle video streams and historical traffic flow data and makes predictions in real time. The communication module transmits the processed data to the server of the traffic signal control service center.

[0046] The signal control scheme generation module in the server calculates and generates a dynamic control scheme for the traffic lights based on the data obtained from the combined processing of different vehicles; the control module then controls the traffic light group according to the control scheme.

[0047] The monitoring equipment and signal light assembly are installed above the signal light poles. The data processing center is located in a cabinet below the signal lights.

[0048] Compared with the prior art, the beneficial effects of the present invention are:

[0049] (1) By integrating real-time vehicle flow video and traffic flow data, we can deeply explore the value of information and design signal control schemes based on vehicle type, traffic violations, road abnormalities, and other conditions to achieve precise traffic congestion relief.

[0050] (2) Based on real-time traffic flow prediction data, the switching time of traffic lights can be dynamically adjusted, which can effectively alleviate congestion during peak hours and reduce resource waste during off-peak hours;

[0051] (3) For special vehicles such as ambulances and police cars, establish an intelligent dispatch system to automatically plan the optimal route and implement dynamic signal priority control;

[0052] (4) Use advanced intelligent technology to record and analyze violations in real time, improve the effectiveness of civilized driving supervision, and build a solid line of defense for traffic safety. Attached Figure Description

[0053] Figure 1 This is a flowchart of the method of the present invention;

[0054] Figure 2 This is a flowchart illustrating the design of the traffic light control scheme of the present invention.

[0055] Figure 3 This is a diagram of the external structure of the system of the present invention;

[0056] Figure 4 This is a schematic diagram of the system structure of the present invention;

[0057] Figure 5 This is a schematic diagram of the server rack structure;

[0058] The components include: 1. High-definition camera; 2. Traffic flow sensor group; 3. Traffic light group; 4. Traffic light pole; 5. Cabinet; 6. Countdown display; 7. Server; 8. Monitor. Detailed Implementation

[0059] The present invention will be further described below with reference to specific embodiments.

[0060] Example 1: A smart traffic signal control method based on traffic flow prediction, comprising the following steps:

[0061] S1. Receive traffic vehicle video streams and traffic flow data through a high-definition camera and traffic flow sensor array, and preprocess the received data:

[0062] The traffic vehicle video stream data is processed and labeled with information, including vehicle information, lane lines, and traffic light area labels. The traffic flow data is processed to remove noise and outliers. The mean filtering algorithm is used to remove noise from the traffic flow data, and the median absolute deviation algorithm is used to identify and remove outliers from the traffic flow data.

[0063] S2. Identify different vehicles, their speed, direction, and trajectory through traffic video streams, and determine whether they are special vehicles. Send the vehicle type, speed, direction, and trajectory information detected at each intersection to the server.

[0064] (1) First, the traffic video stream is converted into images frame by frame. The images are scaled to 640×640. The U-Net model is used for lane line recognition. The Yolov8 model is trained and the category and bounding box coordinates of each vehicle are output.

[0065] (2) Based on the shooting parameters such as video focus and sensor size, and by using an object of known size (such as a calibration board) to calculate the ratio between pixels and actual distance, the pixel coordinates are converted into three-dimensional world coordinates using the camera intrinsic parameter matrix (K), thereby establishing a mapping relationship between pixel coordinates and actual distance;

[0066] (3) For consecutive frames with the same vehicle ID, calculate the pixel displacement of the center point of the bounding box of each consecutive frame. Using the mapping relationship in step (2), convert the pixel displacement into the actual displacement, and combine it with the frame rate to calculate the speed. Speed ​​= actual displacement × frame rate.

[0067] (4) Calculate and determine the direction of motion based on the change of the vehicle’s center point position in consecutive frames, project the vehicle’s center point onto the nearest lane line, and correct the vehicle’s direction in combination with the direction of the lane line, thereby further determining the driving direction.

[0068] (5) The bounding box coordinates, category and other information of each target obtained by the YOLOv8 model, as well as the recorded motion trajectory information of each target, are input into DeepSORT. In the first frame of the video, all vehicles detected by YOLOv8 are initialized as new tracking targets and assigned a unique ID. The ReID (Re-identification) model is used to extract the appearance features of each vehicle to distinguish different vehicles. Then, the Kalman filter is used to predict the motion trajectory of the target and the filter is used to track the target. If the target occlusion or target intersection occurs, DeepSORT will make corrections based on historical information to ensure the accuracy of tracking, thereby realizing the detection and tracking of dynamic targets and predicting vehicle motion trajectories.

[0069] S3. Monitor traffic violations and road anomalies based on traffic vehicle video streams and traffic flow data. Traffic violations include running red lights and speeding, while road anomalies include sudden traffic incidents and road construction. Send traffic violations and road anomalies to the server. The server records and saves sudden traffic incidents and road construction information for each road, and records and sends penalty information for vehicles violating traffic rules.

[0070] (1) The YOLOv8-tiny target detection model is used to locate the traffic light area, and then the traffic light area is segmented by color to output red / yellow / green status. The motion trajectory of each vehicle obtained by the above method is used to extract the stop line position through the U-net model. The red light violation judgment condition is set for each vehicle's trajectory.

[0071] Scenario 1: The traffic light is red at time t, the vehicle's center point is behind the stop line at time t, and the vehicle crosses the stop line at time t+1 when the traffic light is red. This is considered running a red light.

[0072] Scenario 2: If the traffic light is red at time t, the vehicle's center point is in front of the stop line at time t+1, the traffic light is red at time t+1, the vehicle continues to move past the intersection at time t+1, and there are no special vehicles present, then it is considered running a red light.

[0073] Scenario 3: The traffic light is red at time t, and the vehicle's center point is in front of the stop line at time t+1. The traffic light is red at time t+1, and the vehicle's center point is behind the stop line at time t+1. This is considered running a red light.

[0074] Finally, the red light violation event record is output, including vehicle ID, time, and location;

[0075] (2) Extract the instantaneous speed of each vehicle calculated above, extract the speed limit value of the current road segment from the traffic flow data, and for each vehicle, determine whether there is a time point t that satisfies: vehicle speed (t) > speed limit value. If there is, output the speeding event record, which includes vehicle ID, time, and speed value.

[0076] (3) For the trajectory of each vehicle, calculate the rate of change of velocity (acceleration) a(t) between consecutive frames:

[0077]

[0078] If |a(t)|>5m / s 2 Output emergency braking event records, including vehicle ID, time, and acceleration value;

[0079] Extract the direction vector of each vehicle calculated above. and lane direction vector Calculate the angle θ between the trajectory direction and the lane line direction:

[0080]

[0081] If θ > 90°, output the reverse driving event record, including vehicle ID, time, and angle value.

[0082] (4) Use the YOLOv8 target detection model to locate obstacles on the road, and then identify whether there are orange or yellow highlighted areas, identify specific shapes such as cones and water-filled barriers. If the obstacles are densely distributed and form a closed area, it is determined to be a construction area. Then, combine the obstacle type, color and spatial distribution to generate the construction area confidence score. If the confidence score exceeds the threshold (such as 0.8), output the construction event record, which includes the location of the construction area and the confidence score.

[0083] (5) In order to ensure that the behavior can be accurately distinguished when multiple vehicles appear at the same time, each vehicle is assigned a unique ID and cross-frame association is performed. The vehicle's appearance, such as color and model, and motion features such as speed and direction are combined to improve tracking stability. The trajectory and behavior features of each vehicle are classified as: running red lights, speeding, emergency braking, driving against traffic, and normal driving. Then the detected behaviors and anomalies are sent to the server via the HTTP protocol.

[0084] S4. Build a traffic prediction model based on traffic vehicle video streams and historical traffic flow data. Use the traffic prediction model to predict traffic flow and vehicle speed for future time periods based on real-time data, and send the predictions to the server.

[0085] (1) Data processing: Align the calculated vehicle trajectory data and daily traffic flow data according to timestamps so that data at the same point in time can correspond;

[0086] (2) Model building: The vehicle movement trajectory in S2 and the daily traffic flow data in step S1 are processed to form sequence data by aligning the timestamps. The sequence data is used as input, and the timestamp, weekday, and holidays are used as input signals. The traffic flow and speed are used as outputs. An LSTM network is selected for model building. Multi-layer LSTM units capture the long-term dependencies in the input time series. The LSTM output is connected to a fully connected layer for feature integration and prediction. The output is traffic flow and speed.

[0087] (3) Model training: The data is divided into training set, validation set and test set in a ratio of 8:1:1. The Adam optimizer is used and the learning rate is set to 0.001. The training set is used to train the model and the model parameters are updated through the backpropagation algorithm. During the training process, the validation set is used to monitor the model loss value and accuracy. The test set is used to calculate the error index between the predicted value and the true value. The error index is the mean squared error and the mean absolute value.

[0088] (4) Real-time prediction and deployment: Input the real-time traffic video stream and historical data into the trained model, output the predicted traffic flow value and vehicle speed, which are used as input for subsequent dynamic adjustment of traffic lights, and store the prediction results in the server. The server judges the predicted vehicle speed and the speed limit of the lane, and issues a warning when the vehicle is about to exceed the speed limit.

[0089] (5) Model optimization: Based on the performance and feedback of real-time prediction, adjust the model structure, add features, adjust hyperparameters, and update the model regularly, using new data for training to adapt to changes in traffic conditions.

[0090] S5. The system retrieves real-time traffic flow, vehicle location, vehicle type, lane information, vehicle trajectory, traffic violations, road anomalies, and future traffic flow and speed data for each lane at each intersection from the server. Based on the different vehicles, it generates different traffic light control schemes and sends instructions to each intersection for dynamic traffic light control.

[0091] Determine if there is any special vehicle information based on the vehicle category. If so, continue with the following procedure:

[0092] a. Calculate the weight of each road segment based on the traffic state parameters of each road segment in the road network, and then establish the optimal path selection model on the road network connectivity graph abstracted by graph theory. Discrete particle swarm optimization algorithm is used to discretize the particle optimization process, and the optimal path is solved with the goal of minimizing the time for vehicles to reach their destination. Finally, the preset driving route with the shortest travel time is planned ([1] Li Xue. Optimization and Implementation of Green Channel for Special Authorized Vehicles [D]. Nanjing University of Information Science and Technology, 2021);

[0093] b. Based on the determined planned route, associate all traffic light nodes along the route, calculate the switching time of the traffic lights according to the route travel time, and provide feedback to the traffic lights based on the real-time location of special vehicles, thereby enabling special vehicles to have green lights all the way.

[0094] c. Receive new obstacle events and traffic flow information from the server in real time, update the traffic status parameters of each road segment in real time, dynamically adjust the driving route, and update the traffic light control scheme in real time according to the new route.

[0095] Determine if there is any special vehicle information based on the vehicle category. If not, proceed with the following steps:

[0096] (1) Define the objective function:

[0097] Minimize average waiting time

[0098] Maximize road capacity

[0099] Reduce environmental pollution

[0100] Where i is the vehicle number, N is the total number of vehicles, and t wait,i Let t be the waiting time for vehicle i. free,i Let t be the free passage time of vehicle i. idle,i e is the idling time of vehicle i. i For vehicle emission factor; (This objective function is calculated for each intersection. Vehicle numbers can be obtained directly, and the total number of vehicles can also be calculated based on the numbers. Waiting time can be calculated based on the time spent at the intersection. Free passage time is calculated under conditions of no traffic jams and normal traffic lights. Idle time is calculated based on red light waiting time and stopping time during congestion. Vehicle emission factor is a specific parameter that is queried from the database.)

[0101] (2) Initialize the population and randomly generate N solutions, each solution containing the optimal green light time T. g Given the phase switching timing ΔT, calculate the objective function value (f1, f2, f3) for each solution. Then, classify the solutions into different levels based on the quality of the objective values. Use selection, crossover, and mutation to generate the next generation population. Repeat the iteration until convergence and output the optimal solution set, where 30s ≤ T. g ≤120s, ΔT≥5s;

[0102] (3) Use the traffic prediction model in step S4 to predict traffic flow and predict the traffic flow in the future time period. If the traffic flow is predicted to surge, extend the green light time in advance.

[0103] (4) If the traffic flow is detected to be greater than the traffic flow threshold (the traffic flow threshold for main roads is 1200 vehicles / hour / lane, and the traffic flow threshold for other main roads is 600 vehicles / hour / lane), the green light time will be extended immediately. If no vehicles are detected to pass through during the green light period, the green light time will be shortened. If a sudden traffic incident is detected, the previous intersection will switch to red light and the current intersection will switch to green light, prohibiting vehicles from entering the incident area and prioritizing the diversion of traffic flow at surrounding intersections.

[0104] (5) The optimized T g The ΔT signal is sent to the traffic signal controller, such as adjusting the green light time from 60 seconds to 75 seconds for direction A and from 40 seconds to 30 seconds for direction B.

[0105] Real-time monitoring of average waiting time, traffic capacity, and emissions: a 5% reduction. If performance does not meet the standards, parameter fine-tuning is triggered. Data acquisition, optimization, and control are re-executed every minute to accumulate historical data and regularly update the noise filtering model and optimization algorithm parameters.

[0106] Example 2: Intelligent Traffic Signal Control System Based on Traffic Flow Prediction (see...) Figures 3-5 The system includes monitoring equipment, traffic light group 3, data processing center and traffic signal control service center. The monitoring equipment includes high-definition camera group and traffic flow sensor group. The data processing center (corresponding cabinet) includes preprocessing module, vehicle video stream processing module, traffic violation and road accident detection module, traffic prediction model construction and prediction module, and communication module. The traffic signal control service center includes server 7 and display 8. The server includes signal control scheme generation module, control module, data storage and transmission module.

[0107] The monitoring equipment collects traffic vehicle video streams and traffic flow data through a high-definition camera 1 and a traffic flow sensor group 2. The data processing center's preprocessing module preprocesses the received traffic vehicle video streams and traffic flow data. The vehicle video stream processing module identifies different vehicles, their speeds, directions, and trajectories based on the traffic vehicle video streams and determines whether they are special vehicles. The traffic violation and road accident detection module monitors traffic violations and road anomalies based on the traffic vehicle video streams and traffic flow data. Traffic violations include running red lights and speeding, while road anomalies include sudden traffic incidents and road construction. The traffic prediction model construction and prediction module builds a traffic prediction model based on the traffic vehicle video streams and historical traffic flow data and makes predictions in real time. The communication module transmits the processed data to the server of the traffic signal control service center.

[0108] The signal control scheme generation module in the server calculates and generates a dynamic control scheme for the traffic lights based on the data obtained from the combined processing of different vehicles; the control module then controls the traffic light group according to the control scheme.

[0109] The monitoring equipment and signal light assembly are installed above the signal light pole 4. The data processing center is located in the cabinet 5 below the signal light. A countdown display 6 is mounted on the signal light pole 4.

[0110] The above is a further description of the present invention in conjunction with embodiments, and the scope of protection of the present invention is not limited thereto.

Claims

1. A method for intelligent traffic signal control based on traffic flow prediction, characterized in that, The steps include the following: S1. Receive traffic vehicle video stream and traffic flow data through a high-definition camera and a traffic flow sensor group, and pre-process the received data; S2. Identify different vehicles and their speed, direction and motion trajectory through the traffic vehicle video stream, and determine whether they are special vehicles, including ambulances, police cars, fire trucks and rescue vehicles. Send the vehicle type, speed, direction and motion trajectory information detected at each intersection to the server; S3. Monitor traffic violations and road abnormalities based on traffic vehicle video stream and traffic flow data, including red light running, speeding, sudden traffic events and road construction. Send traffic violations and road abnormalities to the server, which records and saves sudden traffic events and road construction information, and sends record and penalty information of traffic violation vehicles; S4. Establish a traffic prediction model based on traffic vehicle video stream and traffic flow historical record data, and predict future traffic flow values and vehicle speeds based on real-time data using the traffic prediction model, and send them to the server: S5. Real-time access to each lane traffic, vehicle location, vehicle category, lane line information, vehicle motion trajectory, traffic violations, road abnormalities, and future traffic flow values and vehicle speeds from the server, and generate different signal light control schemes for different vehicles. The server issues instructions to each intersection for dynamic signal light control: According to the vehicle category judgment in step S2, if there are special vehicles, the following solutions are executed: (1) Calculate the weight of each road segment based on the traffic state parameters of the road network, and then establish an optimal path selection model on the abstracted road network connected graph using graph theory. Discretize the particle optimization process using discrete particle swarm optimization algorithm, and solve the optimal path with the goal of minimizing vehicle arrival time at destination. Finally, the shortest preset driving route is planned; (2) According to the determined planning path, associate all the passing signal light nodes, calculate the signal light switching time according to the path travel time, and feedback the signal light according to the real-time position of the special vehicle, so as to realize the one-way green light for the special vehicle; (3) Real-time receive new obstacle events and traffic flow information from the server, real-time update the traffic state parameters of each road segment, dynamically adjust the driving path, and real-time update the signal light control scheme according to the new path; According to the vehicle category judgment in step S2, if there are no special vehicles, the following solutions are executed: Use multi-objective optimization algorithm to calculate the optimal green light time and phase switching time, which includes minimizing average waiting time, maximizing road capacity and reducing environmental pollution. And dynamically adjust the optimization parameters according to real-time traffic flow and road abnormalities.

2. The intelligent traffic signal control method based on traffic flow prediction according to claim 1, characterized in that, In step S2, the Yolov8 model is used to detect the target in each frame of the video, the vehicle position and category information are obtained, and the U-Net model is used for lane line detection; the vehicle speed is calculated first according to the shooting parameters of the video and the calibration information of the actual scene, the mapping relationship between pixel coordinates and actual distance is established, then the displacement between each vehicle in continuous frames is calculated, combined with the frame rate of the video, and finally the displacement is converted into actual speed; the motion direction is calculated and judged according to the change of the center point position of the vehicle in continuous frames; the vehicle trajectory is integrated with the trained Yolov8 model and DeepSORT algorithm to realize the detection and tracking of dynamic targets.

3. The intelligent traffic signal control method based on traffic flow prediction according to claim 2, characterized in that, The specific process of speed calculation in step S2 is as follows: (1) First, convert the traffic video stream into images frame by frame, train the Yolov8 model, output the category and boundary box coordinates of each vehicle, integrate the trained Yolov8 model with the DeepSORT algorithm, give each vehicle in the video stream a unique ID, and use the U-Net model for lane line detection; (2) According to the shooting parameters of the video and the calculation of the ratio of pixels to actual distance of the known size object, the pixel coordinates are converted into three-dimensional world coordinates by using the camera intrinsic matrix, so as to establish the mapping relationship between pixel coordinates and actual distance; (3) For the same vehicle ID in continuous frames, calculate the pixel displacement of the center point of the boundary box of each vehicle in continuous frames, use the mapping relationship in (2) to convert the pixel displacement into actual displacement, and combine with the frame rate to calculate the speed, speed = actual displacement x frame rate.

4. The intelligent traffic signal control method based on traffic flow prediction according to claim 1, characterized in that, In step S3, the red light running behavior detection first detects the signal light state, then judges whether it enters the stop line area during the red light period according to the detected vehicle trajectory, and finally judges according to the red light state and whether the vehicle trajectory enters the stop line area; The speed detection is to judge whether the vehicle speed is over the speed limit threshold in the traffic flow data; The detection of sudden traffic events includes sudden braking detection and reverse driving detection, the sudden braking detection is judged by calculating the trajectory acceleration change rate, and the reverse driving detection is to judge whether the angle between the trajectory direction and the lane line direction is > 90°; The road construction detection uses the method of recognizing static obstacles and construction area features to judge.

5. The intelligent traffic signal control method based on traffic flow prediction according to claim 4, characterized in that, For each vehicle trajectory, the speed change rate a(t) between consecutive frames is calculated: If |a(t)| > 5 m / s 2 Output hard braking event record, including vehicle ID, time, acceleration value; extracting a direction vector of each vehicle in the step S2 and a lane line direction vector calculating an included angle θ between the trajectory direction and the lane line direction If θ > 90°, output the reverse driving event record, including vehicle ID, time, and angle value.

6. The intelligent traffic signal control method based on traffic flow prediction according to claim 1, characterized in that, In step S4, the construction of the traffic prediction model is to align the time stamps of the vehicle trajectory in step S2 and the daily traffic flow data in step S1 to form sequence data, take the sequence data as input, take the time stamp, the number of weeks, and the holiday as input signals, take the traffic flow and the vehicle speed as output, then use the long short-term memory network to build the model, and train the network to generate the traffic prediction model through the loss function.

7. The intelligent traffic signal control method based on traffic flow prediction according to claim 1, characterized in that, In step S5, the optimal green light time and phase switching time are calculated by using the multi-objective optimization algorithm, the process is as follows: (1) Define the objective function: Minimizing average latency Maximizing road capacity Reducing environmental pollution where i is the vehicle index, N is the total number of vehicles, t wait,i is the waiting time for vehicle i, t free,i is the free-flow time for vehicle i, t idle,i is the idling time for vehicle i, e i is the vehicle emission factor; (2) initialize population, randomly generate N solutions, each solution contains optimal green light time T g and phase switching timing AT, calculate the objective function value (f1, f2, f3) of each solution, then divide into different levels according to the target value, select, cross, and mutate to generate the next generation population, repeat iteration until convergence to output the optimal solution set, wherein 30s≤T g ≤120s, AT≥5s; (3) using the traffic prediction model of step S4 to predict the traffic flow, and predicting the traffic flow of the future time period, if the predicted traffic flow is surging, the green light time is extended in advance; (4) if the detected traffic flow is greater than the traffic flow threshold, the green light time is immediately extended, if no vehicle is detected passing through during the green light time, the green light time is shortened, if a sudden traffic event is detected, the previous intersection is switched to red light, and the current intersection is switched to green light, the vehicle is prohibited to enter the event area, and the traffic flow of the surrounding intersections is preferentially dredged; (5) Adjusted optimized T g and ΔT are sent to the traffic signal controller.

8. An intelligent traffic signal control system based on traffic flow prediction implementing the method of claim 1, characterized by, The system comprises a monitoring device, a signal light group, a data processing center and a traffic signal control service center, the monitoring device comprises a high-definition camera group and a traffic flow sensor group, the data processing center comprises a preprocessing module, a vehicle video stream processing module, a traffic violation and road accident detection module, a traffic prediction model construction and prediction module and a communication module, and the traffic signal control service center comprises a server and a display, the server comprises a signal control scheme generation module, a control module and a data storage and transmission module; The monitoring device collects traffic vehicle video stream and traffic flow through the high-definition camera and the traffic flow sensor group; the preprocessing module of the data processing center pre-processes the received traffic vehicle video stream and traffic flow data, the vehicle video stream processing module identifies different vehicles, their speed, direction and motion trajectory according to the traffic vehicle video stream, and judges whether the vehicle is a special vehicle; the traffic violation and road accident detection module detects traffic violation behaviors and road abnormal conditions according to the traffic vehicle video stream and traffic flow data, the traffic violation behaviors include red light running and overspeed, and the road abnormal conditions include sudden traffic events and road construction; the traffic prediction model construction and prediction module establishes a traffic prediction model according to the traffic vehicle video stream and traffic flow historical record data and makes real-time prediction; and the communication module transmits the processed data to the server of the traffic signal control service center; The signal control scheme generation module of the server calculates and generates a dynamic signal light control scheme according to different vehicles and the processed data; and the control module controls the signal light group according to the control scheme.

9. The intelligent traffic signal control system based on traffic flow prediction as claimed in claim 8, wherein, The monitoring device and the signal light group are installed above the signal light pole, and the data processing center is arranged in a cabinet below the signal light.

Images