A real-time intelligent traffic management method based on Internet of Vehicles

Abstract

The present invention discloses a real-time intelligent traffic management method based on the Internet of Vehicles, which relates to the technical field of traffic control and monitoring. The method comprises: after obtaining a vehicle license, reading vehicle information through a communication network and establishing time-series vehicle data, and constructing a dynamic traffic flow model. Using time-series data to analyze vehicle behavior and extract behavioral features. Based on the dynamic traffic flow, locate road congestion, determine the associated vehicles, and call their time-series data and video streams to trace the congestion. Combined with the tracing results and behavioral characteristics, formulate and issue individual vehicle response plans. Generate vehicle response fitting results, update dynamic traffic flow, and optimize traffic light configuration according to the updated traffic flow, and finally implement traffic management in combination with the response plan. Thereby achieving the technical effect of improving data real-time performance and accurately predicting and regulating trends.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic control, and in particular to a real-time intelligent traffic management method based on Internet of Vehicles. Background Art

[0002] At present, intelligent transportation mainly relies on technical means such as traffic signal control, traffic flow monitoring and navigation systems, which obtain traffic data through cameras, sensors and traffic lights installed on the road, and combine historical data for signal control and route planning. There are technical problems such as data update lag, trend prediction difficulties and inaccurate control. Summary of the invention

[0003] The present invention provides a real-time intelligent traffic management method based on the Internet of Vehicles to solve the technical problems of delayed data update, difficult trend prediction, and inaccurate regulation in the prior art, and achieve the technical effect of improving data real-time performance and accurately performing trend prediction and regulation.

[0004] The present invention provides a real-time intelligent traffic management method based on Internet of Vehicles, comprising:

[0005] After obtaining the vehicle license, the communication network is used to read the vehicle information, establish time-series vehicle data, and use the time-series vehicle data to construct the dynamic traffic flow of the road.

[0006] The time series vehicle data is used to establish vehicle time series association, and the time series vehicle data is called by the vehicle time series association to perform vehicle behavior analysis and establish vehicle behavior characteristics.

[0007] Conduct road congestion analysis based on dynamic traffic flow, establish road congestion analysis positioning, set congestion-related vehicles based on road congestion analysis positioning, call the time-series vehicle data and vehicle video stream of congestion-related vehicles, use the time-series vehicle data and vehicle video stream to conduct joint congestion tracing, use the joint congestion tracing results and vehicle behavior characteristics to establish individual response plans, and send the individual response plans to the corresponding vehicles.

[0008] Establish the response fitting results of the vehicle, and use the response fitting results to update the dynamic traffic flow, use the updated dynamic traffic flow to optimize the traffic light configuration, generate the configuration optimization results, and use the configuration optimization results and individual response plans to manage traffic.

[0009] In a feasible implementation, the time series vehicle data is used to establish a vehicle time series association, and the time series vehicle data is called by the vehicle time series association to perform vehicle behavior analysis and establish vehicle behavior characteristics, which also includes:

[0010] Any vehicle is taken as the target vehicle and the target vehicle is taken as the analysis center. The behavior of the target vehicle is segmented according to the time series vehicle data.

[0011] According to the segmentation results, the association of vehicles adjacent to the target vehicle at the same time node is established to construct the vehicle time series association.

[0012] The speed change and route selection of the associated vehicle in the time series vehicle data are called by vehicle time series association, and the speed change and route selection are used as the pre-data of the segmentation result to perform the behavior analysis of the target vehicle and establish the vehicle behavior characteristics of the target vehicle.

[0013] All vehicles are traversed as target vehicles to construct vehicle behavior characteristics of all vehicles.

[0014] In a feasible implementation, the joint congestion tracing using time-series vehicle data and vehicle video streams also includes:

[0015] Extract features from time-series vehicle data and establish behavioral features and location trajectories.

[0016] Perform data analysis on vehicle video streams to establish vehicle distance characteristics and road status characteristics.

[0017] After the behavioral characteristics, location trajectories, vehicle distance characteristics, and road status characteristics are aligned in time series, the association recognition network is used to perform association event identification, and the congestion source is located based on the association event identification results to complete the joint congestion tracing.

[0018] In a feasible implementation, before establishing the response fitting result of the vehicle, the following steps are also included:

[0019] Acquire feedback information from the vehicle, where the feedback information is feedback information from the vehicle to the individual response plan.

[0020] An individual response scheme is reconstructed using the feedback information, and a response fitting result of the vehicle is established based on the reconstructed individual response scheme.

[0021] In a feasible implementation, the method further includes:

[0022] The updated dynamic traffic flow is used to generate a predicted road condition evaluation result, the predicted road condition evaluation result is used to generate a driving mode preference, and the driving mode preference is used as the first feedback data.

[0023] A speed optimization suggestion is generated according to the road condition evaluation result and the vehicle behavior characteristics, and the speed optimization suggestion is used as the second feedback data.

[0024] The first feedback data and the second feedback data are synchronized to the corresponding vehicle.

[0025] In a feasible implementation, the method further includes:

[0026] A deep learning model of the vehicle is established according to the vehicle behavior characteristics, and the deep learning model is used to predict the continuous behavior of the corresponding vehicle to establish a continuous behavior prediction result.

[0027] Potential congestion items are established through continuous behavior prediction results, and joint processing plans for potential congestion items are generated.

[0028] Predictive management of potential congestion events is performed through the joint processing solution.

[0029] In a feasible implementation, the method further includes:

[0030] Based on the configuration optimization results and individual response plans, the congestion in the current area is sorted out and predicted, and a real-time congestion index is established.

[0031] Obtain the associated areas of the current area, optimize the traffic scheduling of the associated areas based on the real-time congestion index, and perform joint traffic management based on the traffic scheduling optimization results.

[0032] In a feasible implementation, the method further includes:

[0033] Determine whether the vehicle supports assisted driving.

[0034] If supported, a response fleet cluster is established based on the individual response plan to manage assisted driving using the response fleet cluster.

[0035] The present invention discloses a real-time intelligent traffic management method based on the Internet of Vehicles, comprising: after obtaining a vehicle license, reading vehicle information through a communication network, establishing a time-series vehicle data model, and using these data to construct a dynamic traffic flow model of a road; based on the time-series vehicle data, establishing a vehicle time-series association, calling the association information to perform vehicle behavior analysis, and generating a vehicle behavior feature model; using the dynamic traffic flow model to analyze the road congestion situation, and identifying congestion-related vehicles through congestion analysis positioning, calling the time-series data and vehicle video streams of these related vehicles, and combining the two to perform joint congestion tracing; according to the congestion tracing results and vehicle behavior characteristics, formulating individual response plans, and sending these plans to corresponding vehicles; by analyzing the response fitting results of the vehicles, updating the dynamic traffic flow model, and using the updated dynamic traffic flow to optimize the traffic light configuration, generating an optimized configuration result; finally, combining the configuration optimization results and the individual response plan to perform overall traffic management, and optimize the road flow control and congestion management. The present invention discloses a real-time intelligent traffic management method based on the Internet of Vehicles, which solves the technical problems of delayed data update, difficult trend prediction and inaccurate regulation, and achieves the technical effect of improving data real-time performance and accurately performing trend prediction and regulation. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1A flow chart of a real-time intelligent traffic management method based on Internet of Vehicles of the present invention;

[0037] Figure 2 The present invention is a flowchart of establishing vehicle behavior characteristics in a real-time intelligent traffic management method based on Internet of Vehicles. DETAILED DESCRIPTION

[0038] The technical solution provided in the embodiments of the present invention is to solve the technical problems of data update lag, trend prediction difficulty, and inaccurate regulation in the prior art. The overall idea adopted is as follows:

[0039] First, after obtaining the vehicle license, the communication network is used to read the vehicle information, establish the time-series vehicle data, and use it to construct the dynamic traffic flow of the road; then, the time-series vehicle data is used to establish the vehicle time-series association, and the time-series vehicle data is called based on the vehicle time-series association to perform vehicle behavior analysis, and then establish the vehicle behavior characteristics; then, based on the dynamic traffic flow, road congestion analysis is performed, road congestion analysis positioning is established, and congestion-related vehicles are set based on this, and the time-series vehicle data and vehicle video streams of these vehicles are called to perform joint congestion tracing; then, using the joint congestion tracing results and vehicle behavior characteristics, individual response plans are established, and the individual response plans are issued to the corresponding vehicles; then, the response fitting results of the vehicles are established, and the dynamic traffic flow is updated with the response fitting results; finally, the updated dynamic traffic flow is used to optimize the traffic light configuration, generate the configuration optimization results, and implement traffic management by combining the configuration optimization results and the individual response plans.

[0040] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention are shown in the drawings, rather than all of them.

[0041] Example 1

[0042] Figure 1 The present invention is a flowchart of a real-time intelligent traffic management method based on Internet of Vehicles, wherein the method comprises:

[0043] After obtaining the vehicle license, the communication network is used to read the vehicle information, establish time-series vehicle data, and use the time-series vehicle data to construct the dynamic traffic flow of the road.

[0044] Specifically, first, obtain access permissions from multiple connected vehicles in the target area to ensure compliance of data reading; then, use connected vehicle technology to exchange data with vehicles through the communication network. Exemplary, data exchange is performed through cellular networks (such as 4G, 5G) and short-range wireless communication technologies (such as DSRC, dedicated short-range communication). The acquired time-series vehicle data provides basic data for the subsequent establishment of time-series vehicle data and road traffic models.

[0045] Specifically, time series vehicle data refers to vehicle status information recorded in a time series, reflecting the vehicle's movement trajectory and state changes over a period of time. By continuously reading vehicle information and recording it by time, the time series data set can be established to analyze the vehicle's driving dynamics. The frequency of vehicle information collection is determined according to actual needs. Generally, high-speed vehicles require higher-frequency data collection (such as once per second), while low-speed vehicles can reduce the collection frequency.

[0046] Optionally, the acquired time-series vehicle data include: location information, the latitude and longitude coordinates of the vehicle obtained through global positioning systems such as GPS and Beidou; speed and acceleration, real-time speed and acceleration data obtained through vehicle sensors (such as wheel speed sensors, acceleration sensors, etc.); driving direction, the vehicle's driving direction calculated by the vehicle's steering wheel sensor or positioning trajectory; vehicle status data, including fuel level, power level, headlight status, brake status and other information.

[0047] Furthermore, based on the time series data of vehicles, the distribution of vehicles on each road at different times is analyzed to build a dynamic traffic flow model of the road. This model is used to reflect the flow pattern of vehicles on the road and is helpful for subsequent real-time road condition prediction and control.

[0048] Optionally, the basis of the dynamic traffic flow model is a simplified digital model of the target area roads, which is used to represent the geometry and structure of the roads, and the distribution and flow of vehicles on the roads are dynamically reflected by mapping the time series data of the vehicles onto the simplified digital model.

[0049] Optionally, the dynamic traffic flow model simplifies the target area road into a one-dimensional or two-dimensional geometric structure, representing the length, width, number of lanes and lane distribution of the road. Among them, the one-dimensional model means that the road is simplified into a line segment, and the position of the vehicle is represented by coordinate points. The two-dimensional model represents the width, turns, intersections and other features of the road through a two-dimensional plane, and the position of the vehicle is represented by plane coordinates. Optionally, the model also includes road attributes, such as speed limit, number of lanes, section length, etc.

[0050] Optionally, based on the vehicle position recorded in the time series data, the vehicle state (including speed, position, direction, etc.) is mapped to the dynamic traffic model to dynamically update the spatial distribution of the vehicle. The latitude and longitude information of the vehicle is converted into coordinate values ​​in the digital road model, and the speed and direction of the vehicle represent the traffic dynamics at the current moment.

[0051] For example, the dynamic traffic model is a simplified representation of the flow of vehicles on the road. In order to improve computational efficiency and visualization, vehicles are represented in the model as "particles". Different types of vehicles (such as cars, trucks, and buses) are distinguished by the size of the particles, and the position of the particles indicates the specific position of the vehicle in the digital model of the road. In each time period, the density of particles on different road sections reflects the density of traffic. The higher the traffic density, the higher the degree of congestion on the road; the lower the density, the smoother the road. At the same time, the distribution of particle size at a certain moment on the road can characterize the composition characteristics of the traffic (i.e., the proportion and distribution of different types of vehicles in the traffic).

[0052] Optionally, the dynamic traffic flow model can be displayed in the form of graphics or animation through visualization tools to facilitate intuitive understanding of the traffic flow conditions on the road. For example, the traffic density and speed distribution at different times on a certain road can be displayed in real time on the screen of the traffic control center.

[0053] The time series vehicle data is used to establish vehicle time series association, and the time series vehicle data is called by the vehicle time series association to perform vehicle behavior analysis and establish vehicle behavior characteristics.

[0054] Specifically, vehicle temporal association refers to the association between different vehicles in the same time period due to factors such as adjacent spatial positions and similar driving speeds. By determining and establishing this association, the interaction and behavior of these vehicles in a specific time period can be analyzed, and the driving habits and behavioral characteristics of the target vehicle can be further analyzed.

[0055] In some embodiments, Figure 2 As shown, the time series vehicle data is used to establish a vehicle time series association, and the time series vehicle data is called by the vehicle time series association to perform vehicle behavior analysis and establish vehicle behavior characteristics, and also includes:

[0056] Take any vehicle as the target vehicle and the target vehicle as the analysis center, and segment the target vehicle behavior according to the time series vehicle data; establish the association of vehicles adjacent to the target vehicle at the same time node according to the segmentation result to construct the vehicle time series association;

[0057] The speed change and route selection of the associated vehicles in the time series vehicle data are called by vehicle time series association, and the speed change and route selection are used as the pre-data of the segmentation result to perform the behavior analysis of the target vehicle and establish the vehicle behavior characteristics of the target vehicle; all vehicles are traversed as target vehicles to construct the vehicle behavior characteristics of all vehicles.

[0058] Specifically, first, any vehicle is designated as a target vehicle, and a behavioral analysis is performed based on the time series data of the vehicle, including taking the target vehicle as the analysis center and segmenting the time period according to its driving status within a certain period of time (such as speed changes, route selection, direction changes, etc. of the target vehicle). A single trip of the target vehicle may be divided into multiple time periods, each of which represents a specific driving state, such as constant speed cruising, lane changing, waiting to turn, choosing a fork in the road, etc.

[0059] Specifically, in each segment time of the segment division result, the target vehicle is associated with other adjacent vehicles in the same time period. The vehicle time series association is constructed based on factors such as the relative position, speed, and driving direction between two or more vehicles, and is used to reflect the interaction relationship between vehicles at a specific time point.

[0060] Optionally, the relative distance between the target vehicle and other vehicles is determined through the time series data of the vehicle (such as position coordinates). If a vehicle is within the discrimination distance of the target vehicle (such as within 100 meters), it is identified as an "adjacent vehicle". The relative distance can be adaptively adjusted based on the speed of the target vehicle. The higher the speed of the target vehicle, the greater the discrimination distance of the adjacent vehicle.

[0061] Optionally, in addition to the relative positions, the vehicle time series association is constructed by the speed and direction of the adjacent vehicles. For example, if the speed of the adjacent vehicles is similar to that of the target vehicle and they are traveling in the same direction, it can be considered that these vehicles have a strong association.

[0062] Optionally, the vehicle timing association also includes behavior association. For example, if a certain vehicle overtakes a target vehicle at a certain time point, it constitutes a behavior association.

[0063] Furthermore, through the vehicle time series association, the speed data of adjacent vehicles is used as a reference for the target vehicle behavior analysis, and then the speed changes of the target vehicle in different time periods are analyzed. For example, if the target vehicle is a slow vehicle, a fast vehicle, or a mainstream speed vehicle in the traffic flow, for example, if the target vehicle frequently accelerates and decelerates, and the speed changes are synchronized with those of adjacent vehicles, it may indicate that the owner has a strong tendency to follow the vehicle or change lanes.

[0064] For example, if the target vehicle maintains a constant speed for most of the time and rarely interacts with adjacent vehicles in terms of speed and route, it can be defined as "robust driving". If the target vehicle frequently accelerates, changes lanes, and overtakes or blocks adjacent vehicles, it can be judged that the owner has "aggressive driving" characteristics. If the target vehicle always maintains a close distance to the vehicle in front, and the speed changes are consistent with the vehicle in front, it can be judged that the target vehicle has a strong tendency to follow the vehicle.

[0065] Furthermore, all vehicles are traversed, and each vehicle is taken as a target vehicle, and the aforementioned time series data call, segmentation, association establishment and behavior analysis are repeated to comprehensively construct the behavioral characteristics of all vehicles.

[0066] Through the above-mentioned behavior analysis based on adjacent vehicle data, the vehicle behavior characteristics of the target vehicle can be more accurately judged according to the overall traffic status of the current road. Compared with the behavior characteristic definition method based on fixed judgment rules, it helps to improve the intelligence and personalization level of traffic management.

[0067] Conduct road congestion analysis based on dynamic traffic flow, establish road congestion analysis positioning, set congestion-related vehicles based on road congestion analysis positioning, call the time-series vehicle data and vehicle video stream of congestion-related vehicles, use the time-series vehicle data and vehicle video stream to conduct joint congestion tracing, use the joint congestion tracing results and vehicle behavior characteristics to establish individual response plans, and send the individual response plans to the corresponding vehicles.

[0068] Specifically, based on the traffic density, vehicle speed, lane utilization and other information contained in the dynamic traffic flow, the area and severity of road congestion are determined. For example, when the traffic density, average traffic speed, lane utilization and other information of the dynamic traffic flow trigger the congestion control time limit, it can be considered that there is congestion. Then the congested areas are located, which are usually characterized by higher traffic density and vehicles running at low speeds.

[0069] Specifically, congestion-associated vehicles refer to multiple vehicles within a time-space window corresponding to the road congestion analysis and positioning. For example, the current position of the vehicle is mapped out based on time series data and combined with the road digital model to determine which vehicles are in the congested area.

[0070] Furthermore, the time series data (such as position, speed, acceleration, etc.) of congestion-related vehicles and the vehicle video stream of individual vehicles are combined for joint data analysis. The time series data of congestion-related vehicles provides a macro perspective of traffic flow, while the vehicle video stream provides a micro perspective of driving, which helps to analyze the causes of congestion from multiple angles.

[0071] In some embodiments, the combined congestion tracing using time-series vehicle data and vehicle video streams also includes:

[0072] Feature extraction is performed on time-series vehicle data to establish behavioral features and position trajectories. Data analysis is performed on vehicle video streams to establish vehicle distance features and road status features. After time-series alignment of behavioral features, position trajectories, vehicle distance features, and road status features, associated event discrimination is performed through an associated recognition network. The source of congestion is located based on the associated event discrimination results to complete joint congestion tracing.

[0073] Specifically, the behavioral features extracted from the time series vehicle data include: the frequency and duration of the vehicle's acceleration, deceleration, and constant speed driving; the number of lane changes; the change in vehicle distance; sudden acceleration and deceleration behaviors, etc. The position trajectory is calculated based on the position coordinates in the vehicle time series data to help determine the vehicle's specific position on the road, its driving route, and its relative position to other vehicles.

[0074] Specifically, the vehicle video stream is used to provide real-time information about the vehicles and road environment around the target vehicle. By parsing the video stream, important data such as vehicle distance characteristics and road status characteristics can be extracted to supplement the deficiencies of time series data and provide more detailed traffic conditions. For example, through computer vision technology (such as target detection algorithm), surrounding vehicles are identified from the video stream, and the distance between the target vehicle and surrounding vehicles is calculated as the vehicle distance feature; by parsing the video stream and identifying information such as road markings, traffic lights, speed limit signs, etc., the actual road conditions can be understood.

[0075] Specifically, the association recognition network uses behavioral features, location trajectories, vehicle distance features, and road state features as inputs to identify associated events between vehicles. The association recognition network is a network based on deep learning (such as RNN, LSTM, etc.). It not only analyzes the behavior of a single vehicle, but also identifies the behavioral associations between multiple vehicles through time-series alignment features, especially the interaction between the front and rear vehicles. Then, based on the multi-vehicle behavioral associations and vehicle distance changes, it identifies the key events that cause congestion. For example, a sudden braking of the vehicle in front may cause intensive braking of the vehicles behind, triggering a chain reaction. The analysis of associated events based on the association recognition network can accurately locate the source vehicles or events that cause congestion. For example, a vehicle's slow movement or improper lane change may trigger a congestion response from the vehicle behind, and it will be located as the source of congestion.

[0076] Optionally, the association recognition network is trained with a large number of congestion cases. The training data comes from historical road congestion events, combining time-series vehicle data and vehicle video streams, and annotating the associated events that cause congestion (such as braking, lane changing, etc.). Through these historical data, the association recognition network can learn the characteristic patterns in different congestion scenarios, thereby tracing the source of congestion.

[0077] Furthermore, by identifying the source of congestion and related events, combined with the behavioral characteristics of the vehicle and the congestion chain, a personalized response plan is generated for each congestion-related vehicle to help alleviate congestion. For example, for vehicles at the source of congestion, measures such as reducing unnecessary sudden braking or lane changes, increasing slow vehicle speeds, etc. can be proposed to prevent further aggravation of congestion. For the rear vehicles in the congestion chain, adjustment suggestions can be generated, such as maintaining vehicle distance, adjusting vehicle speed, maintaining a constant speed, reducing frequent acceleration and deceleration, etc., to avoid congestion caused by fluctuation effects and relieve traffic pressure.

[0078] Furthermore, after the personalized response plan is generated, it will be sent to the corresponding vehicle through the Internet of Vehicles equipment and communication network. The driver can receive real-time driving suggestions through voice prompts, visual prompts, etc.

[0079] The above method steps, by identifying the source of congestion and related events, combined with vehicle behavior characteristics and congestion chain information, generate a personalized response plan for each related vehicle. Then generate corresponding suggestions such as reducing emergency braking, adjusting vehicle distance, optimizing vehicle speed, etc., to help ease traffic pressure, avoid further aggravation of congestion, and improve the stability and safety of traffic flow.

[0080] Establish the response fitting results of the vehicle, and use the response fitting results to update the dynamic traffic flow, use the updated dynamic traffic flow to optimize the traffic light configuration, generate the configuration optimization results, and use the configuration optimization results and individual response plans to manage traffic.

[0081] Specifically, the vehicle response fitting result is a model established through data fitting technology based on the actual response of individual vehicles to individual response plans. The fitting result reflects the execution effect and behavior adjustment of the vehicle to the individual response plan, which is used to provide a basis for subsequent dynamic traffic flow updates.

[0082] Specifically, based on the response fitting results of the vehicles, the overall dynamic traffic flow is updated. The actual response behavior of the vehicles will affect the density, speed distribution and fluidity of the traffic flow. The updated dynamic traffic flow can more accurately reflect the current road conditions.

[0083] Specifically, after the dynamic traffic flow is updated, the configuration of the traffic lights is optimized. This includes adjusting the duration and phase settings of the traffic lights to improve the efficiency of road traffic. The optimization is based on the existing traffic light management system or optimization model. For example, a traffic signal timing model (such as the Webster model) is used to calculate the optimal signal timing based on traffic flow data. This includes extending the green light time and shortening the red light time on branch roads on main roads with dense traffic to ensure smooth traffic on the main roads; optimizing the phase configuration of traffic lights according to traffic conditions in different directions. For example, according to the proportion of traffic flow choosing a direction, adjust the signal phase and adjust the time for turning left or going straight to improve the efficiency of vehicle traffic at the intersection.

[0084] In some embodiments, before establishing the response fitting result of the vehicle, the method further includes:

[0085] Acquire feedback information of the vehicle, wherein the feedback information is feedback information of the vehicle to the individual response scheme; reconstruct the individual response scheme with the feedback information, and establish a response fitting result of the vehicle based on the reconstructed individual response scheme.

[0086] Specifically, the vehicle's feedback information refers to the vehicle's feedback on the execution of the individual response plan issued, that is, the vehicle's compliance with the adjustment suggestions during actual driving, such as whether the vehicle speed is adjusted, whether the vehicle distance is maintained, whether the lane change is reduced, etc. By obtaining feedback information, the execution effect of the individual response plan can be determined.

[0087] Optionally, in addition to the behavioral feedback of the vehicle itself, the driver's subjective reaction to the individual response plan can also be obtained, such as "confirm" or "ignore" instructions input through the in-vehicle interaction terminal or the driver's response inferred through behavioral inference.

[0088] Specifically, the received feedback information is parsed, including the comparison between the actual behavior of the vehicle and the recommended solutions, identifying which suggestions are followed and which are not implemented, and then reconstructing the initial individual response solutions to better match the actual reactions of the vehicle and the driver. The reconstructed individual response solutions are consistent with the actual reactions of the vehicle and the driver, and are used as the response fitting results of the vehicle.

[0089] Furthermore, the method further comprises:

[0090] The updated dynamic traffic flow is used to generate predicted road condition evaluation results, and the predicted road condition evaluation results are used to generate driving mode preferences, which are used as first feedback data. Speed ​​optimization suggestions are generated based on the road condition evaluation results and vehicle behavior characteristics, and the speed optimization suggestions are used as second feedback data. The first feedback data and the second feedback data are synchronized to the corresponding vehicles.

[0091] Specifically, the updated dynamic traffic flow is used as the predicted road condition, and based on the same method and principle as the aforementioned road congestion analysis, a predicted road condition evaluation is performed to generate a predicted road condition evaluation result. Optionally, the predicted road condition evaluation result includes information such as road congestion level, congestion location, congestion length, and travel speed.

[0092] Optionally, through time series analysis (such as ARIMA model, LSTM, etc.), combined with the updated dynamic traffic flow, the future traffic density, speed distribution, road congestion and other traffic flow change trends can be predicted to identify possible traffic fluctuations, congested or smooth sections in the future.

[0093] Specifically, based on the predicted road condition evaluation results, the vehicle's driving mode preference is generated, which may include energy-saving mode and performance mode. Among them, the energy-saving mode aims to help the vehicle save fuel or electricity by reducing sudden acceleration, sudden braking, and maintaining a constant speed. The performance mode focuses on improving the driving experience and is suitable for drivers in smooth road conditions or who have a demand for acceleration and speed.

[0094] For example, in cases of light or moderate congestion, the driver is advised to adopt energy-saving mode to avoid acceleration and frequent braking; while in smooth road conditions, performance mode is recommended to improve traffic efficiency.

[0095] Specifically, speed optimization suggestions are generated based on the road condition evaluation results and vehicle behavior characteristics, and used as the second feedback data to help drivers adjust their speed according to actual road conditions and their own driving habits, thereby improving driving efficiency, reducing congestion, and ensuring driving safety. For example, for drivers who are used to driving fast, a deceleration reminder is issued before the driver's usual acceleration behavior in the predicted congested sections or accident-prone areas in the future to avoid sudden braking.

[0096] Furthermore, the method further comprises:

[0097] A deep learning model of the vehicle is established based on the vehicle behavior characteristics, and the deep learning model is used to predict the continuous behavior of the corresponding vehicle to establish a continuous behavior prediction result; potential congestion items are established through the continuous behavior prediction results, and a joint processing plan for the potential congestion items is generated; and the prediction and management of potential congestion items are performed through the joint processing plan.

[0098] Specifically, a deep learning model is used to learn the vehicle behavior patterns of multiple target vehicles and drivers through vehicle behavior characteristics, and continuous behavior prediction of the corresponding vehicles is performed. The obtained continuous behavior prediction results reflect the possible vehicle status information of the corresponding target vehicles and drivers at preset future time nodes, that is, the time series vehicle data at preset future moments.

[0099] Specifically, potential congestion items refer to traffic congestion problems that may occur in the future. Exemplarily, the dynamic traffic flow of the road is updated according to the continuous behavior prediction results to determine potential congestion items. For example, if the continuous behavior prediction results show that multiple vehicles will frequently change lanes or brake suddenly on a certain road section, the road section will be marked as a potential congestion point. Among them, the continuous length of the road section and the predicted traffic flow included can be mapped to the predicted congestion parameters of potential congestion items (congestion location, congestion length, travel speed, etc.).

[0100] Furthermore, based on the same method principle as the above-mentioned individual response plan, a corresponding joint processing plan is generated based on the potential congestion items and the continuous behavior prediction results. The joint processing plan is a forward-looking management plan based on the prediction. For example, if there is a potential congestion item caused by multiple vehicles changing lanes or slowing down on a certain road section, a plan of diversion, lane change restriction or vehicle speed adjustment can be formulated to prevent possible congestion.

[0101] Furthermore, the method further comprises:

[0102] According to the configuration optimization results and individual response plans, the congestion in the current area is sorted out and predicted, and a real-time congestion index is established. The associated areas of the current area are obtained, and the traffic scheduling of the associated areas is optimized based on the real-time congestion index, and joint traffic management is carried out based on the traffic scheduling optimization results.

[0103] Specifically, a prediction model based on dynamic traffic flow (such as traffic state estimation model, time series model) is used, combined with real-time traffic density, speed distribution and traffic light configuration, to simulate the traffic flow change trend in the current area, judge the congestion situation that may occur in the next few minutes or kilometers, and then establish a real-time congestion index based on the congestion forecast results. The real-time congestion index is used to quantify the road congestion situation in the current area and provide a decision-making basis for subsequent scheduling optimization and traffic management in related areas.

[0104] Optionally, the real-time congestion index is calculated through a traffic state estimation model or a data-driven model (such as multivariate regression analysis, deep learning, etc.), involving factors such as traffic density, vehicle speed distribution, individual response behavior, etc.

[0105] Specifically, the associated areas of the current area include the front area and the rear area. The traffic conditions in the front area directly affect the traffic flow in the current area. For the front associated area of ​​the current area, traffic scheduling can be optimized by re-planning routes, detours and other measures; the rear area refers to the downstream area connected to the current area. If congestion occurs in the rear area, vehicles cannot leave the current area in time, which may aggravate the congestion in the current area. Therefore, for the rear associated area of ​​the current area, diversion, detours or reduction of traffic input and other measures can be taken to optimize traffic scheduling and improve the outflow efficiency of the target area.

[0106] Furthermore, the method further comprises:

[0107] Determine whether the vehicle supports assisted driving; if so, establish a response fleet cluster based on the individual response plan to manage assisted driving using the response fleet cluster.

[0108] Specifically, judging the assisted driving support of the vehicle includes support at the software and hardware level and support at the user level. Among them, judging the support at the software and hardware level includes checking whether the vehicle is equipped with the necessary hardware for assisted driving (such as sensors and cameras); confirming that the software version supports the assisted driving function, etc. Judging the support at the user level includes: ensuring that the user has agreed to enable the assisted driving function, collecting user consent information, and ensuring compliance and privacy protection.

[0109] Specifically, if support at both the software and hardware levels and the user level are achieved, the corresponding vehicles will be incorporated into a response fleet cluster, which will strictly perform collaborative management of assisted driving in accordance with the corresponding individual response plans to optimize traffic flow.

[0110] In summary, the real-time intelligent traffic management method based on the Internet of Vehicles provided by the present invention has the following technical effects:

[0111] After obtaining vehicle license, vehicle information is read through the communication network, and a time-series vehicle data model is established. The dynamic traffic flow model of the road is constructed using these data. Based on the time-series vehicle data, the vehicle time-series association is established, and the association information is called for vehicle behavior analysis to generate a vehicle behavior feature model. The road congestion is analyzed using the dynamic traffic flow model, and congestion-related vehicles are identified through congestion analysis and positioning. The time-series data and vehicle video streams of these related vehicles are called, and the two are combined for joint congestion tracing. According to the congestion tracing results and vehicle behavior characteristics, individual response plans are formulated and issued to the corresponding vehicles. By analyzing the response fitting results of the vehicle, the dynamic traffic flow model is updated, and the updated dynamic traffic flow is used to optimize the traffic light configuration to generate the optimized configuration results. Finally, the configuration optimization results and individual response plans are combined for overall traffic management to optimize the road flow control and congestion management. Thus, the technical effect of improving the real-time nature of data and accurately predicting and regulating trends is achieved.

[0112] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the above-mentioned embodiments. It should be understood that those skilled in the art can still modify the technical solutions recorded in the above-mentioned embodiments, or replace some of the technical features therein by equivalents; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention.

Claims

1. A real-time intelligent traffic management method based on Internet of Vehicles, characterized in that: The method comprises: After obtaining the vehicle license, the communication network is used to read the vehicle information, establish time-series vehicle data, and construct the dynamic traffic flow of the road with the time-series vehicle data; Use time series vehicle data to establish vehicle time series association, and use vehicle time series association to call time series vehicle data to perform vehicle behavior analysis and establish vehicle behavior characteristics; Conduct road congestion analysis based on dynamic traffic flow, establish road congestion analysis positioning, set congestion-related vehicles based on road congestion analysis positioning, call the time series vehicle data and vehicle video stream of congestion-related vehicles, use the time series vehicle data and vehicle video stream to conduct joint congestion tracing, use the joint congestion tracing results and vehicle behavior characteristics to establish individual response plans, and send the individual response plans to the corresponding vehicles; Establish the response fitting results of the vehicle, and update the dynamic traffic flow with the response fitting results, optimize the traffic light configuration with the updated dynamic traffic flow, generate the configuration optimization results, and manage the traffic with the configuration optimization results and individual response plans; The joint congestion tracing using time-series vehicle data and vehicle video streams also includes: Extract features from time-series vehicle data to establish behavior features and location trajectories; Perform data analysis on vehicle video streams to establish vehicle distance characteristics and road status characteristics; After the behavioral features, location trajectories, vehicle distance features, and road status features are aligned in time series, the associated event identification is performed through the associated recognition network, and the congestion source is located based on the associated event identification results to complete the joint congestion tracing; Generate a predicted road condition evaluation result using the updated dynamic traffic flow, generate a driving mode preference using the predicted road condition evaluation result, and use the driving mode preference as the first feedback data; Generate speed optimization suggestions based on road condition evaluation results and vehicle behavior characteristics, and use the speed optimization suggestions as second feedback data; Synchronizing the first feedback data and the second feedback data to the corresponding vehicle; Establishing a deep learning model of the vehicle according to the vehicle behavior characteristics, using the deep learning model to predict the continuous behavior of the corresponding vehicle, and establishing a continuous behavior prediction result; Establish potential congestion items through continuous behavior prediction results, and generate joint processing solutions for potential congestion items; Predictive management of potential congestion events through the joint processing scheme; According to the configuration optimization results and individual response plans, the congestion in the current area is sorted out and predicted, and a real-time congestion index is established; Obtain the associated areas of the current area, optimize the traffic scheduling of the associated areas based on the real-time congestion index, and perform joint traffic management based on the traffic scheduling optimization results; Determine whether the vehicle supports assisted driving; If supported, a response fleet cluster is established based on the individual response plan to manage assisted driving using the response fleet cluster.

2. A real-time intelligent traffic management method based on Internet of Vehicles as claimed in claim 1, characterized in that: The method of establishing a vehicle time series association by using the time series vehicle data, and using the vehicle time series association to call the time series vehicle data to perform vehicle behavior analysis and establish vehicle behavior characteristics also includes: Take any vehicle as the target vehicle and the target vehicle as the analysis center, and segment the target vehicle behavior according to the time series vehicle data; According to the segmentation results, the association of vehicles adjacent to the target vehicle at the same time node is established to construct the vehicle time series association; The speed change and route selection of the associated vehicle in the time series vehicle data are called by the vehicle time series association, and the speed change and route selection are used as the pre-data of the segmentation result to perform the behavior analysis of the target vehicle and establish the vehicle behavior characteristics of the target vehicle; All vehicles are traversed as target vehicles to construct vehicle behavior characteristics of all vehicles.

3. A real-time intelligent traffic management method based on Internet of Vehicles as claimed in claim 1, characterized in that: Before establishing the response fitting result of the vehicle, the method further includes: Acquiring feedback information from the vehicle, where the feedback information is feedback information from the vehicle to the individual response plan; An individual response scheme is reconstructed using the feedback information, and a response fitting result of the vehicle is established based on the reconstructed individual response scheme.

Images