Internet of vehicles traffic management method based on Beidou positioning

Abstract

The invention discloses an Internet of Vehicles traffic management method based on Beidou positioning, and belongs to the field of traffic management, and the method comprises the steps: obtaining Beidou positioning data and sensor network local sensing data, and extracting global coordinate information and local traffic state information; performing time synchronization processing on the Beidou positioning data and the sampling data of the sensor network by adopting a timestamp alignment method, and eliminating the difference in time scale; whether signal interruption or drifting exists in the Beidou positioning data is judged, and if signal abnormity exists, missing positioning data is estimated through an interpolation algorithm according to local sensing data of the sensor network; according to the fused data, constructing a dynamic distribution model of vehicles in a global road network, and identifying a relationship between traffic density and road network distribution; and a multi-source data regression model based on a convolutional neural network is adopted to predict the vehicle position during signal interruption or drifting, and continuous global positioning information is generated.

Description

Technical Field

[0001] The present invention belongs to the field of traffic management, and in particular relates to a vehicle networking traffic management method based on Beidou positioning. Background Art

[0002] In the urban traffic management scenario, the vehicle networking technology based on the Beidou positioning system and the sensor network faces a key technical problem: how to ensure the seamless fusion and synchronization between the real-time Beidou positioning data and the local perception data of the sensors in a high-density traffic environment. The Beidou system provides the global position information of vehicles, but due to the existence of obstacles such as high-rise buildings and tunnels in the urban environment, the Beidou signal may be temporarily interrupted or drifted, resulting in a decrease in positioning accuracy. At the same time, the sensor network can capture the local traffic status around the vehicle in real time, such as vehicle speed, vehicle distance, road conditions, etc., but these data lack the support of the global coordinate system and are difficult to directly match with the Beidou positioning data.

[0003] When the Beidou positioning data is deviated or lost, the local data collected by the sensor network cannot be accurately mapped into the global traffic network, resulting in a fault in the decision-making basis of the traffic management system. For example, in a traffic congestion scenario, the Beidou system may not be able to accurately identify the specific position of the vehicle, while the sensor network can sense the local traffic flow density but cannot determine the distribution of these traffic flows in the global road network. In this case, it is difficult for the data processing center to effectively associate the local perception data with the global positioning data through the existing fusion algorithms, thus affecting the comprehensiveness of traffic information and the accuracy of decision-making.

[0004] In addition, there may be a difference between the update frequency of the Beidou positioning data and the sampling frequency of the sensor network. The Beidou system usually updates the position information at a second-level frequency, while the sensor network may collect local data at a millisecond-level frequency. This mismatch in time scale will cause a time synchronization problem when the data processing center performs data fusion, further exacerbating the inconsistency of traffic information. Summary of the Invention

[0005] To solve the above technical problems, the present invention provides a vehicle networking traffic management method based on Beidou positioning, including:

[0006] Obtain the Beidou positioning data and the local perception data of the sensor network, and extract the global coordinate information and the local traffic status information;

[0007] Perform time synchronization processing on the global coordinate information and the local traffic status information to obtain the Beidou global information and the traffic local information;

[0008] Determine whether there is signal interruption or drift in the Beidou global information. If there is signal anomaly, estimate the missing positioning data according to the local traffic information through an interpolation algorithm;

[0009] Based on the Kalman filtering algorithm, fuse the missing positioning data with the Beidou global information, and eliminate data noise through two steps of state prediction and measurement update to obtain a fused data set;

[0010] Construct a dynamic distribution model of vehicles in the global road network, input the fused data set into the dynamic distribution model for calculation, and obtain the relationship between traffic density and road network distribution;

[0011] Construct a multi-source data regression model based on a convolutional neural network, input the relationship between traffic density and road network distribution into the multi-source data regression model to predict the vehicle position during signal interruption or drift, and generate continuous global positioning information;

[0012] Update the decision-making basis of the traffic management system based on the continuous global positioning information, judge whether the traffic density exceeds the critical value through a preset threshold, and if it exceeds, trigger a traffic congestion warning mechanism and optimize the traffic flow distribution mechanism;

[0013] Based on the optimized traffic flow distribution mechanism, adopt a timing optimization algorithm based on traffic density to dynamically adjust the timing scheme of traffic lights and alleviate the congestion problem in a high-density traffic environment.

[0014] Preferably, the process of extracting global coordinate information and local traffic state information includes:

[0015] Use the Beidou positioning system to obtain the positioning data of the target vehicle, and collect local perception data in combination with a preset sensor network;

[0016] Process the positioning data and perception data through a data fusion algorithm to extract the global coordinate information of the target vehicle;

[0017] Match the preset road grid model according to the global coordinate information to determine the road area where the target vehicle is located;

[0018] Analyze the traffic flow, vehicle speed and vehicle distance parameters of the current road area through local perception data, and extract local traffic state information.

[0019] Preferably, the process of obtaining Beidou global information and local traffic information includes:

[0020] Analyze the traffic flow, vehicle speed and vehicle distance parameters of the current road area through local perception data, and extract local traffic state information;

[0021] The Beidou positioning data and the local traffic state information are time-synchronized by using a timestamp alignment method to obtain the Beidou global information and the traffic local information.

[0022] Preferably, the process of determining whether there is a signal interruption or drift in the Beidou global information, and if there is a signal anomaly, estimating the missing positioning data according to the traffic local information includes:

[0023] Obtain the time series of Beidou positioning data and determine whether there is a signal interruption or drift;

[0024] If a signal anomaly is detected, extract the time series of local perception data from the sensor network;

[0025] Adopt an interpolation algorithm to repair the missing positioning data and estimate the complete time series;

[0026] Perform time alignment processing on the repaired positioning data and the local perception data of the sensor network, and construct a trajectory model of the target vehicle according to the aligned data;

[0027] Adopt a clustering algorithm to perform anomaly detection on the repaired positioning data and exclude abnormal points;

[0028] Perform smoothing processing on the repaired trajectory model through a regression algorithm to obtain the final positioning data.

[0029] Preferably, the process of obtaining the fusion data set includes:

[0030] Obtain the time series of Beidou positioning data and determine whether there is a signal anomaly. If there is a signal anomaly, extract the time series of local perception data from the sensor network;

[0031] Adopt the Kalman filtering algorithm and the time series of the local perception data to perform data fusion on the Beidou positioning data and the local perception data of the sensor, and eliminate data noise through two steps of state prediction and measurement update to generate the fusion data set.

[0032] Preferably, the process of obtaining the relationship between traffic density and road network distribution includes:

[0033] Obtain the global road network map and vehicle distribution point data, and count the traffic volume for each road network segment;

[0034] If the number of vehicles in a road network segment exceeds a preset threshold, mark it as a high-density road segment, otherwise mark it as a low-density road segment;

[0035] Adopt a clustering algorithm to group the vehicle distribution points, identify and remove abnormal points;

[0036] Calculate the density value of each road network segment based on the processed distribution point data;

[0037] Construct a dynamic distribution model, input the road network segment, time period, and density value, and output a relationship graph of traffic density and road network distribution;

[0038] If there are abnormal areas in the relationship graph, recalculate the density values of the relevant road network segments;

[0039] Adjust the parameters of the dynamic distribution model according to the updated density values;

[0040] Optimize the model using a regression algorithm to improve the model accuracy;

[0041] Obtain real-time vehicle distribution point data, input it into the optimized dynamic distribution model, and obtain a relationship graph of the current traffic density and road network distribution.

[0042] Preferably, the process of generating continuous global positioning information includes:

[0043] Construct a multi-source data regression model based on a convolutional neural network, input the relationship between the traffic density and road network distribution into the multi-source data regression model for calculation, and obtain a prediction result;

[0044] Calculate the position value of the vehicle in the global road network based on the prediction result. Process the position value using an interpolation algorithm to generate continuous global positioning information. If the positioning information deviates from the historical data by more than a preset threshold, recalculate the position value, and update the vehicle distribution map in the global road network according to the final positioning information.

[0045] Preferably, the process of optimizing the traffic flow distribution mechanism includes:

[0046] Update the decision-making basis of the traffic management system according to the continuous global positioning information, use a preset threshold to determine whether the traffic density exceeds the critical value. If the traffic density exceeds the critical value, trigger a traffic congestion warning mechanism. According to the congestion warning information, use an optimization algorithm to reallocate the traffic flow, generate a new traffic flow distribution plan, update the traffic state information, and feedback the updated traffic state information to the traffic management system to optimize the traffic flow distribution mechanism.

[0047] On the other hand, the present invention also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor. When the processor executes the computing program, the method is implemented.

[0048] On the other hand, the present invention also provides a computer-readable storage medium storing a computer program, and when the computer program is executed by a processor, the method is implemented.

[0049] Compared with the prior art, the present invention has the following advantages and technical effects:

[0050] The present invention discloses a vehicle networking traffic management method based on Beidou positioning. This method obtains Beidou positioning data and local perception data of the sensor network, and extracts global coordinate information and local traffic state information. The time stamp alignment method is used to achieve data synchronization, and the Kalman filter algorithm is used to fuse multi-source data to improve the positioning accuracy. When the Beidou signal is interrupted or drifts, the present invention uses the sensor network data and convolutional neural network to predict the vehicle position to ensure the continuity of the positioning information. Based on the fused data, a vehicle dynamic distribution model is constructed to identify the relationship between traffic density and road network distribution. The present invention dynamically updates the basis for traffic management decisions according to the prediction results and local perception data, and adjusts the traffic lights using a timing optimization algorithm based on traffic density, effectively alleviating the congestion problem in a high-density traffic environment and improving the intelligent level and efficiency of traffic management. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] The drawings constituting a part of this application are used to provide a further understanding of this application. The schematic embodiments of this application and their descriptions are used to explain this application and do not constitute an improper limitation of this application. In the drawings:

[0052] Figure 1 It is a schematic flowchart of the method according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0053] It should be noted that, without conflict, the embodiments in this application and the features in the embodiments can be combined with each other. The following will refer to the drawings and combine the embodiments to detail this application.

[0054] It should be noted that the steps shown in the flowchart of the drawings can be executed in a computer system such as a set of computer-executable instructions, and although the logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in a different order than here.

[0055] Embodiment 1

[0056] As Figure 1 shown, this embodiment provides a vehicle networking traffic management method based on Beidou positioning, including:

[0057] Step S101, obtain Beidou positioning data and local perception data of the sensor network, and extract global coordinate information and local traffic state information.

[0058] The Beidou positioning system is used to obtain the positioning data of the target vehicle, and the local perception data is collected in combination with the preset sensor network. The positioning data and the perception data are processed by a data fusion algorithm to extract the global coordinate information of the target vehicle.

[0059] Specifically, the Beidou positioning system can accurately determine the position of the target vehicle through multi-satellite positioning. For example, a bus running on the main urban road can obtain its longitude and latitude coordinates, speed, heading and other data per second through the Beidou receiving terminal. The roadside sensor network includes devices such as cameras and radars. For example, the video surveillance deployed at intersections can capture local traffic parameters such as traffic flow and vehicle speed. The data fusion algorithm combines the Beidou positioning and sensor perception data. For example, the Beidou positioning of a bus shows that it is located at 116 degrees east longitude, 40 degrees north latitude, and the speed is 40 kilometers per hour. At the same time, the roadside camera recognizes that the vehicle is passing through a certain intersection. After fusion, the specific position of the vehicle in the road grid can be accurately determined. The road grid model realizes refined management by dividing the urban road into several grid units. For example, Chang'an Avenue is divided into grid units of 50 meters × 50 meters, and each grid contains attributes such as road geometric features and traffic capacity.

[0060] Step S102, using the timestamp alignment method, perform time synchronization processing on the Beidou positioning data and the sampling data of the sensor network to eliminate the differences in time scales.

[0061] Using the timestamp alignment method to perform time synchronization processing on the Beidou positioning data and the sensor network sampling data, a data set with a consistent time scale is obtained. According to the data set after time synchronization, the global coordinate information of the target vehicle is extracted, and the preset road grid model is matched to determine the road area where the target vehicle is located. Through the local perception data, analyze the traffic flow, vehicle speed and vehicle distance parameters in the current road area, and extract the local traffic state information.

[0062] Specifically, timestamp alignment is the basis for realizing the data fusion of Beidou positioning data and sensor network data. For example, the sampling rate of Beidou positioning data is once per second, while the sampling rate of the roadside sensor network data is five times per second. Through timestamp alignment, the two types of data can be unified to the same time scale. The linear interpolation method is used to process the low-sampling-rate data to achieve time synchronization of the data. When determining the global coordinates of the target vehicle, the Beidou positioning system provides longitude and latitude information, and the road grid model divides the road into several grid units. The size of each grid unit is ten meters by ten meters, and the specific grid position where the vehicle is located can be determined through coordinate mapping. For example, a vehicle is located at 120 degrees 05 minutes east longitude and 30 degrees 10 minutes north latitude, corresponding to grid coordinates of grid No. 400 in the area. The local traffic state information includes multiple key parameters. The traffic flow is calculated by the number of vehicles passing through the cross-section per hour. For example, 2,000 vehicles per hour in one direction is medium traffic flow.

[0063] Step S103: Determine whether there is a signal interruption or drift in the Beidou positioning data. If there is signal anomaly, estimate the missing positioning data according to the local perception data of the sensor network through an interpolation algorithm.

[0064] Obtain the time series of the Beidou positioning data and determine whether there is a signal interruption or drift. If signal anomaly is detected, extract the time series of the local perception data from the sensor network. Use an interpolation algorithm to repair the missing positioning data and estimate the complete time series. Perform time alignment processing on the repaired positioning data and the local perception data of the sensor network. Construct a trajectory model of the target vehicle based on the aligned data. Use a clustering algorithm to perform anomaly detection on the repaired positioning data and exclude the anomaly points. Smooth the repaired trajectory model through a regression algorithm to obtain the final positioning data.

[0065] Specifically, the time series of Beidou positioning data is usually recorded with millisecond-level accuracy, such as sampling ten times per second. Signal interruption may be manifested as no data at consecutive multiple time points, while signal drift is reflected as the positioning coordinates suddenly deviating from the normal trajectory. For example, in urban high-rise areas, satellite signals are easily blocked, resulting in a decrease in positioning accuracy or signal loss. At this time, devices such as radar and cameras in the sensor network can provide local perception data as supplementary data sources. For data missing situations, linear interpolation algorithms can be used for data repair on short time scales. For example, if a vehicle has a signal interruption for three consecutive seconds at an intersection, by analyzing the position data at the previous and subsequent moments, the trajectory points in the missing interval are estimated. For longer-term data missing, polynomial interpolation methods can be used to reconstruct the trajectory in combination with the vehicle driving characteristics. Time alignment processing requires establishing a unified time reference. The time stamp of the Beidou positioning system can be selected as the reference. The sampling data of the sensor network is resampled according to the time stamp to ensure consistency with the sampling frequency of the positioning data. For example, if the Beidou data is sampled ten times per second, the sensor data needs to be adjusted to the same sampling rate. Trajectory model construction considers vehicle motion characteristics, including parameters such as speed and acceleration. The motion states in different driving stages are described by piecewise functions, such as uniform motion, acceleration, and deceleration. The trajectory model can be used to predict short-term future positions and provide a reference for data anomaly detection. Density-based spatial clustering of applications with noise (DBSCAN) is used for clustering analysis to identify trajectory anomaly points. The vehicle positions are grouped by time window, and the density distribution of the position points within each group is calculated. Points that significantly deviate from the main density region are marked as anomaly points. For example, if a vehicle suddenly has a large lateral deviation of the positioning point on a straight road, it may be an outlier. Kernel regression is used for trajectory smoothing processing, and the Gaussian kernel function is used to perform weighted averaging on the trajectory points. The smoothed trajectory is more in line with the actual motion law of the vehicle and reduces the influence of positioning noise. For example, in a turning section, the original trajectory may be serrated, and after smoothing, a curve closer to the actual turning trajectory can be obtained. The core of the above data processing flow is to ensure the continuity and reliability of the positioning data. Through multi-source data fusion and algorithm optimization, the accuracy of vehicle trajectory reconstruction is improved, providing high-quality data support for subsequent traffic analysis and decision-making.

[0066] Step S104, using the Kalman filter algorithm, fuse the Beidou positioning data with the local perception data of the sensor network, and eliminate data noise and improve positioning accuracy through two steps: state prediction and measurement update.

[0067] Obtain the time series of Beidou positioning data, and determine whether there is signal anomaly. If there is signal anomaly, extract the time series of local perception data from the sensor network. Use the Kalman filtering algorithm to perform data fusion on the Beidou positioning data and the sensor local perception data, and eliminate data noise through two steps of state prediction and measurement update. According to the fused data, construct the trajectory model of the target vehicle, and use the clustering algorithm to perform anomaly detection on the trajectory data to exclude abnormal points. Smooth the trajectory model through the regression algorithm to obtain the final positioning data and determine the precise position of the target vehicle. Obtain the local perception data of the sensor network, and determine whether there is deviation from the Beidou positioning data. If there is deviation, use the interpolation algorithm to repair the missing data. According to the repaired data, perform Kalman filtering again, update the state prediction and measurement results, and improve the data accuracy. Align the time of the finally processed positioning data with the local perception data of the sensor network to determine the complete trajectory of the target vehicle.

[0068] Specifically, the time series of Beidou positioning data mainly includes position coordinates and timestamps. For example, a vehicle samples once per second when driving on an urban road, recording data of latitude, longitude, and altitude. In areas with dense high-rise buildings, satellite signals may be blocked, resulting in signal interruption, or there may be drift caused by multipath effects. When the Beidou signal is abnormal, local perception data of the vehicle passing by can be obtained with the help of the sensor network deployed along the way, such as cameras on traffic light poles, roadside units and other devices. The Kalman filtering algorithm realizes data fusion through two steps of state prediction and measurement update. For the vehicle trajectory, state prediction is based on a uniform or uniformly accelerated motion model for position prediction, and measurement update corrects the prediction by combining Beidou positioning and sensor perception data. For example, when a vehicle passes through an intersection, the Beidou positioning shows that it is 5 meters off the center line of the road, while the intersection camera shows that the vehicle is driving normally in the lane. At this time, the filtering algorithm will correct the Beidou data. Trajectory anomaly detection uses the density clustering method to identify points deviating from the normal driving path as abnormal points. For example, when a vehicle is driving on a straight road, the trajectory points should be linearly distributed. If there are isolated points far from the main path, they are regarded as abnormal points and excluded. The smoothing process can use local weighted regression to eliminate small fluctuations while maintaining the overall trend of the trajectory. The local perception data of the sensor network includes information such as the passing time, position, and speed of the vehicle. When there is deviation from the Beidou positioning data, such as the sensor shows that the vehicle passes through the intersection at 10:30, while the Beidou data is missing, linear interpolation is used to estimate the position data of the missing period. Based on the repaired complete data sequence, perform Kalman filtering again and continuously iterate to optimize the state estimation.

[0069] Step S105, according to the fused data, construct a dynamic distribution model of the vehicle in the global road network, and identify the relationship between traffic density and road network distribution.

[0070] Obtain the global road network map and vehicle distribution point data, and count the traffic volume for each road network segment. If the number of vehicles in a road network segment exceeds the preset threshold, it is marked as a high-density segment; otherwise, it is marked as a low-density segment. Use a clustering algorithm to group the vehicle distribution points, identify and remove abnormal points. Calculate the density value of each road network segment based on the processed distribution point data. Construct a dynamic distribution model, input the road network segment, time period, and density value, and output a relationship graph of traffic density and road network distribution. If there are abnormal areas in the relationship graph, recalculate the density values of the relevant road network segments. Adjust the parameters of the dynamic distribution model according to the updated density values. Use a regression algorithm to optimize the model and improve the model accuracy. Obtain real-time vehicle distribution point data, input it into the optimized dynamic distribution model, and get a relationship graph of the current traffic density and road network distribution. If the relationship graph deviates significantly from the historical data, recalculate the density values.

[0071] Specifically, the traffic volume analysis in the road network map first requires obtaining global road network information and vehicle distribution data. For example, at the intersections of urban arterial roads and secondary arterial roads, vehicle position information is collected through intersection monitoring devices and in-vehicle positioning systems. The length of the main road section in a certain urban commercial area is two kilometers, and during the morning rush hour, the number of vehicles detected reaches 30 per 100 meters, exceeding the preset threshold of 20, so it is marked as a high-density section. The clustering analysis of vehicle distribution points uses the density clustering method to group vehicle distribution points that are close to each other. For example, in the road sections around schools, there are vehicle aggregation phenomena during school arrival and dismissal times, and abnormal points that do not conform to the normal distribution law are identified through clustering analysis. There is a sudden concentration of vehicles on a certain road section during non-peak hours, and after analysis, it is found to be caused by temporary construction. The construction of the dynamic distribution model needs to consider two dimensions: time and space. Taking the business district as an example, during the period from 10 am to 4 pm on weekdays, the traffic flow density shows a stable state, maintaining at about 15 vehicles per 100 meters on average. During weekends, the traffic flow density in this area increases significantly, reaching twice that of weekdays, and fluctuates greatly. The model parameters are optimized through regression analysis to improve the prediction accuracy. The matching degree analysis between real-time data and the model can detect abnormal situations. Due to a large-scale promotion event in a certain commercial area, the traffic volume surges, and there are significant deviations between the measured data and the predicted values of the historical model, so it is necessary to recalculate the road section density values.

[0072] In step S106, a multi-source data regression model based on a convolutional neural network is used to predict the vehicle positions during signal interruption or drift, and generate continuous global positioning information.

[0073] Obtain multi-source data, including vehicle sensor data, road network information, and signal status data. Use a convolutional neural network to extract features from the multi-source data and construct a regression model. If the signal status data is interrupted or drifts, activate the regression model for prediction. According to the prediction results, calculate the position value of the vehicle in the global road network. Use an interpolation algorithm to process the position value and generate continuous global positioning information. If the positioning information deviates from the historical data by more than a preset threshold, recalculate the position value. Update the vehicle distribution map in the global road network according to the final positioning information.

[0074] Specifically, the collection of multi-source data is the basis of the intelligent transportation system. On a main road in a certain city, in-vehicle sensors are installed to obtain key information such as the speed and position of vehicles. At the same time, the road management department has established a road network database containing information such as road length, number of lanes, and speed limits. Traffic lights upload the current signal status through a real-time communication system. These data form a complete traffic information collection network. During the feature extraction process of the convolutional neural network, taking a crossroads as an example, the first layer of the network can extract basic features of the vehicle such as position coordinates and driving speed. The second layer identifies the relationship between the vehicle and the surrounding environment, such as the distance from other vehicles and the relative position to the intersection. The last layer integrates to obtain the operating state features of the vehicle, providing a basis for subsequent prediction. When the signal light data is abnormal, the regression model plays a key role. The signal controller at a certain intersection caused data interruption due to equipment failure. At this time, the regression model predicts the current signal timing according to the historical data pattern. The model considers factors such as the conventional timing scheme and the trend of traffic flow changes to generate the optimal predicted value of the signal status. The calculation of global positioning information involves multiple steps. Taking the urban ring road as an example, first obtain the original coordinate points of the vehicle, which are discontinuous. Through the cubic spline interpolation algorithm, the discrete coordinate points are converted into a smooth driving trajectory. If it is found that a certain section of the trajectory differs greatly from the historical driving characteristics, the system will re-collect the positioning data in this area.

[0075] Step S107, update the decision-making basis of the traffic management system according to the prediction results and the local perception data of the sensor network. Judge whether the traffic density exceeds the critical value through a preset threshold. If it exceeds, trigger the traffic congestion warning mechanism and optimize the traffic flow distribution.

[0076] Obtain the local perception data of the sensor network and extract the traffic flow and vehicle speed characteristics. Update the decision-making basis of the traffic management system according to the prediction results and the local perception data. Use a preset threshold to judge whether the traffic density exceeds the critical value. If the traffic density exceeds the critical value, trigger the traffic congestion warning mechanism. According to the congestion warning information, use an optimization algorithm to re-allocate the traffic flow. Generate a new traffic flow distribution plan and update the traffic state information. Feed back the updated traffic state information to the traffic management system to complete the closed-loop processing.

[0077] Specifically, the local perception data of the sensor network includes various types. For example, a road magnetic sensor can obtain the magnetic field changes when a vehicle passes by, a coil detector can count the traffic flow, and a camera can identify the vehicle type and speed, etc. Taking the video surveillance installed on a road section as an example, vehicle features can be extracted through image processing technology, the vehicle speed can be recorded in kilometers per hour, and at the same time, the number of vehicles passing through within a unit time is counted as a traffic flow indicator. In a specific application, during the morning rush hour on the east-west main road of a certain intersection, the sensor collects data every five minutes. When it is detected that the vehicle speed drops from the original 40 kilometers per hour to 20 kilometers per hour, and the traffic flow continues to be more than 30 vehicles per minute, the system will update the basis for traffic management decisions according to these data. The setting of the critical value needs to consider factors such as road grade and the number of lanes. For an urban main road with two-way four lanes, the vehicle density critical value can be set to 100 vehicles per kilometer. When the system detects that the vehicle density on a certain road section reaches 120 vehicles per kilometer, a congestion warning will be triggered. This warning mechanism can detect potential traffic jam risks in advance. The optimization algorithm can be adopted in various ways in practical applications. For example, when there is congestion at a roundabout, the system adjusts the signal timing plan to divert some traffic flows to the surrounding roads. Specifically, the original 60-second green light time for both the east-west and north-south directions can be adjusted to 40 seconds for the east-west direction and 80 seconds for the north-south direction, guiding some vehicles to choose alternative routes. The new traffic flow distribution plan needs to consider the carrying capacity of the surrounding road network. Taking the road network around a commercial area as an example, if there is congestion on the main road, the system will analyze the real-time traffic conditions of the alternative roads. When it is found that the traffic flow on the north side branch road is small, 30% of the traffic flow can be guided to change the route through a variable message sign, and at the same time, the green light time of the branch road is appropriately extended to ensure the smooth operation of the traffic after diversion. The update of traffic state information involves multiple dimensions. For example, the system records the change trends of indicators such as road section vehicle speed, traffic flow, and vehicle density. After the optimization measures are implemented, if it is found that the vehicle speed on the main road increases to 35 kilometers per hour and the vehicle density drops to 80 vehicles per kilometer, it can be determined that the dredging measures are effective. These dynamic data will be continuously fed back to the traffic management system to evaluate the regulation effect and adjust the management strategy in a timely manner.

[0078] Step S108, according to the global road network distribution and local traffic conditions, adopt a timing optimization algorithm based on traffic density to dynamically adjust the signal timing plan of traffic lights, and alleviate the congestion problem in a high-density traffic environment.

[0079] Obtain the topological structure of the global road network and the traffic state data of local road segments. Extract traffic flow and vehicle speed characteristic information from the traffic state data. Use a preset threshold to determine whether the traffic density reaches the standard of a high-density traffic environment. If the traffic density reaches the high-density standard, trigger the timing optimization algorithm to adjust the timing plan of traffic lights. Generate a new traffic light timing plan according to the calculation results of the timing optimization algorithm. Deploy the new traffic light timing plan to the traffic signal control system. Monitor the changes in traffic state in real time, update the global road network and local state data, and form a dynamic adjustment closed-loop.

[0080] Specifically, the topological structure of the road network can be represented by nodes and connection relationships. Taking an intersection as an example, the intersection can be represented as a node, forming connection relationships with the roads in the surrounding four directions. The traffic state data of a single road segment includes key information such as traffic flow and average driving speed. These data can be collected through sensing devices such as coils and cameras installed on the road surface or roadside. For traffic flow characteristics, on a main road with two-way four lanes during the morning rush hour, the number of vehicles passing through per hour can reach three thousand, while the average vehicle speed may drop to twenty kilometers per hour. By setting a reasonable traffic density threshold, for example, when the vehicle spacing is less than ten meters and the duration exceeds five minutes, it can be determined that the high-density traffic state is entered. The timing optimization in a high-density traffic environment involves multiple aspects. Assume an intersection where a main road intersects with a secondary road. During the peak period, the green light time for the straight lane on the main road can be extended from the original forty seconds to sixty seconds, while compressing the green light time of the secondary road to improve the traffic efficiency of the main road. However, it should be noted that when adjusting the signal timing, the coordinated timing of upstream and downstream intersections needs to be considered to avoid creating new congestion points. The timing optimization algorithm will calculate the optimal timing plan based on the real-time collected traffic flow data. Taking a four-phase signal timing as an example, if the traffic flow in the north-south direction of the main road is twice that in the east-west direction, the green light time in the north-south direction can be increased to 1.5 times that in the east-west direction accordingly. The algorithm will also consider the queue length of the left-turn lane and can set a separate left-turn protection phase if necessary. When deploying the new timing plan, it needs to be adjusted gradually to avoid drastic changes. A typical adjustment process is: first increase the green light time of the main road by five seconds per cycle, and if the traffic flow improves after observing for ten minutes, continue to fine-tune until the expected effect is achieved. At the same time, the traffic signal control system will continuously monitor the operating status of each intersection. The real-time monitoring system can grasp the overall traffic situation through various sensors distributed in the road network. For example, in an area composed of ten signal intersections, when congestion occurs at a certain intersection, the system will automatically adjust the signal timing of the upstream intersection to reduce the inflow of vehicles into the congested section. This dynamic adjustment mechanism enables the traffic system to operate adaptively according to the road network load and improves the overall traffic efficiency.

[0081] On the other hand, this embodiment also provides an electronic device, including a memory, a processor, and a computing program stored in the memory and executable on the processor, where the processor implements the method when executing the computing program.

[0082] On the other hand, this embodiment also provides a computer-readable storage medium storing a computer program, where the computer program implements the method when executed by a processor.

[0083] The above are only the preferred specific embodiments of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed by the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A vehicle networking traffic management method based on Beidou positioning, characterized in that, Including: Obtain Beidou positioning data and local perception data of the sensor network, and extract global coordinate information and local traffic state information; Perform time synchronization processing on the global coordinate information and local traffic state information to obtain Beidou global information and traffic local information; Judge whether there is signal interruption or drift in the Beidou global information. If there is signal abnormality, estimate the missing positioning data according to the local traffic information through an interpolation algorithm; Fuse the missing positioning data with the Beidou global information based on the Kalman filter algorithm, and eliminate data noise through two steps of state prediction and measurement update to obtain a fused data set; Construct a dynamic distribution model of vehicles in the global road network, input the fused data set into the dynamic distribution model for calculation, and obtain the relationship between traffic density and road network distribution; Construct a multi-source data regression model based on a convolutional neural network, input the relationship between traffic density and road network distribution into the multi-source data regression model to predict the vehicle position during signal interruption or drift, and generate continuous global positioning information; Update the decision-making basis of the traffic management system based on the continuous global positioning information, judge whether the traffic density exceeds the critical value through a preset threshold, and if it exceeds, trigger a traffic congestion warning mechanism and optimize the traffic flow distribution mechanism; Based on the optimized traffic flow distribution mechanism, adopt a timing optimization algorithm based on traffic flow density to dynamically adjust the timing plan of traffic lights and relieve the congestion problem in a high-density traffic environment.

2. The method according to claim 1, wherein The process of extracting global coordinate information and local traffic state information includes: Use the Beidou positioning system to obtain the positioning data of the target vehicle, and collect local perception data in combination with a preset sensor network; Process the positioning data and perception data through a data fusion algorithm to extract the global coordinate information of the target vehicle; Match the preset road grid model according to the global coordinate information to determine the road area where the target vehicle is located; Extract local traffic state information by analyzing traffic flow, vehicle speed and vehicle distance parameters in the current road area through local perception data.

3. The method according to claim 1, characterized in that The process of obtaining Beidou global information and traffic local information includes: Analyze traffic flow, vehicle speed and vehicle distance parameters in the current road area through local perception data to extract local traffic state information; Use the timestamp alignment method to perform time synchronization processing on the Beidou positioning data and the local traffic state information to obtain the Beidou global information and traffic local information.

4. The method according to claim 1, characterized in that The process of judging whether there is signal interruption or drift in the Beidou global information. If there is signal abnormality, estimate the missing positioning data according to the local traffic information through an interpolation algorithm includes: Obtain the time series of Beidou positioning data and judge whether there is signal interruption or drift; If signal abnormality is detected, extract the time series of local perception data from the sensor network; Adopt an interpolation algorithm to repair the missing positioning data and estimate the complete time series; Perform time alignment processing on the repaired positioning data and the local perception data of the sensor network, and construct a trajectory model of the target vehicle according to the aligned data; Using a clustering algorithm, perform anomaly detection on the repaired positioning data to exclude abnormal points; Through a regression algorithm, smooth the repaired trajectory model to obtain the final positioning data.

5. The method according to claim 1, wherein The process of obtaining the fusion data set includes: Obtain the time series of Beidou positioning data, and determine whether there is signal anomaly. If there is signal anomaly, extract the time series of local perception data from the sensor network; Using the Kalman filter algorithm and the time series of the local perception data, perform data fusion on the Beidou positioning data and the sensor local perception data, and eliminate data noise through two steps of state prediction and measurement update to generate the fusion data set.

6. The method according to claim 1, characterized in that, The process of obtaining the relationship between traffic density and road network distribution includes: Obtain the global road network map and vehicle distribution point data, and count the traffic volume for each road network segment; If the number of vehicles in a road network segment exceeds the preset threshold, mark it as a high-density segment, otherwise mark it as a low-density segment; Use a clustering algorithm to group the vehicle distribution points, identify and remove abnormal points; According to the processed distribution point data, calculate the density value of each road network segment; Construct a dynamic distribution model, input the road network segment, time period and density value, and output the relationship diagram between traffic density and road network distribution; If there is an abnormal area in the relationship diagram, recalculate the density value of the relevant road network segment; According to the updated density value, adjust the parameters of the dynamic distribution model; Use a regression algorithm to optimize the model and improve the model accuracy; Obtain real-time vehicle distribution point data, input it into the optimized dynamic distribution model, and obtain the relationship diagram between the current traffic density and the road network distribution.

7. The method according to claim 1, characterized in that, The process of generating continuous global positioning information includes: Construct a multi-source data regression model based on a convolutional neural network, input the relationship between the traffic density and the road network distribution into the multi-source data regression model for calculation, and obtain a prediction result; Calculate the position value of the vehicle in the global road network based on the prediction result, use an interpolation algorithm to process the position value, generate continuous global positioning information. If the positioning information deviates from the historical data by more than the preset threshold, recalculate the position value, and update the vehicle distribution map in the global road network according to the final positioning information.

8. The method according to claim 1, wherein The process of optimizing the traffic flow allocation mechanism includes: According to the continuous global positioning information, update the decision-making basis of the traffic management system, use a preset threshold to judge whether the traffic density exceeds the critical value. If the traffic density exceeds the critical value, trigger the traffic congestion warning mechanism. According to the congestion warning information, use an optimization algorithm to re-allocate the traffic flow, generate a new traffic flow allocation plan, update the traffic state information, and feedback the updated traffic state information to the traffic management system to optimize the traffic flow allocation mechanism.

9. An electronic device, comprising a memory, a processor, and a computing program stored in the memory and executable on the processor, characterized in that, When the processor executes the calculation program, it implements the method described in any one of claims 1-8.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the method described in any one of claims 1-8.

Images