A vehicle driving behavior portrait analysis system based on a K-means algorithm

Abstract

A vehicle driving behavior portrait analysis system based on a K-means algorithm, characterized by comprising a UAV collection and monitoring module, a collection and transmission module, a driving trajectory analysis module, a driving behavior portrait analysis module, a driving behavior portrait classification module, a driving behavior warning module, and a road risk control module; the vehicle driving behavior portrait analysis system based on the K-means algorithm proposed in the application collects and monitors driving trajectory data of vehicles on a road through a UAV, proposes a driving trajectory analysis algorithm, predicts the trajectory of a driving vehicle, performs driving behavior portrait clustering analysis based on predicted data of the trajectory of the driving vehicle by using a K-means portrait clustering algorithm, classifies driving behavior portraits according to the result of the portrait clustering analysis, and takes different road management measures based on the difference in the number of driving vehicles of various driving behavior portraits.

Description

Technical Field

[0001] This invention relates to the field of intelligent highways, and in particular to a vehicle driving behavior profiling and analysis system based on the K-means algorithm. Background Technology

[0002] With the rapid development of the automotive industry and transportation infrastructure, global road traffic volume is constantly increasing. This has brought convenience but also inevitably led to a large number of traffic safety problems. As vehicle ownership rises, the frequency and severity of traffic accidents have also increased significantly. Traditional traffic management methods and limited resources are struggling to cope with increasingly complex traffic conditions, and corresponding road management measures are also insufficient to meet road maintenance needs. Therefore, seeking effective technological means to improve driving safety, optimize traffic flow, and reduce accident rates has become an urgent issue. Against this backdrop, intelligent transportation systems and advanced data analysis technologies have gradually received widespread attention.

[0003] Traditional driving behavior monitoring methods typically rely on vehicle-mounted monitoring equipment, such as dashcams, in-vehicle GPS devices, and various sensors. These devices record the vehicle's operating status, including basic information such as speed, location, and speeding warnings. However, these methods usually only monitor the behavior of a single vehicle and struggle to provide macroscopic traffic flow information. Furthermore, for accident-prone road sections, the lack of comprehensive monitoring hinders the identification of potential traffic hazards and fails to effectively reduce the risk of accidents.

[0004] In recent years, drone technology, as a flexible, rapid, and low-cost means of aerial data collection, has attracted widespread attention and application across various fields. In the transportation sector, drones, through high-altitude camera and positioning technology, can monitor traffic flow over large areas in real time, rapidly becoming an ideal tool for traffic monitoring and data collection. Drones not only break through the fixed-view limitations of traditional cameras, providing a more dynamic and comprehensive perspective, but also support the analysis of traffic environmental conditions through their sensors for temperature, humidity, and fog / haze.

[0005] In conclusion, combining drone technology with data analytics not only overcomes many shortcomings of traditional driver monitoring methods but also provides a novel solution for traffic safety and management. The development and application of this system will undoubtedly drive the informatization and intelligentization of the transportation sector, thereby making significant contributions to building smart cities and improving public transportation safety. Through comprehensive driver behavior monitoring and precise risk warning mechanisms, it will significantly improve road traffic safety, increase travel efficiency, and provide traffic managers with a scientific basis for decision-making. Summary of the Invention

[0006] The purpose of this invention is to provide a vehicle driving behavior profiling and analysis system based on the K-means algorithm, which aims to solve the problems mentioned in the background above.

[0007] To achieve the above objectives, this invention proposes a vehicle driving behavior profiling and analysis system based on the K-means algorithm. The system comprises a UAV data acquisition and monitoring module, a data acquisition and transmission module, a driving trajectory analysis module, a driving behavior profiling and analysis module, a driving behavior profiling and classification module, a driving behavior early warning module, and a road risk management module. In the UAV data acquisition and monitoring module, the user uses a UAV to collect driving trajectory data of vehicles traveling on the road. In the data acquisition and transmission module, a data signal transmission channel is established using a wireless communication protocol. In the driving trajectory analysis module, a driving trajectory analysis algorithm is proposed to predict the trajectory of traveling vehicles. In the driving behavior profiling and analysis module, a K-means profiling and clustering algorithm is proposed to perform driving behavior profiling and clustering analysis based on the predicted driving trajectory data. In the driving behavior profiling and classification module, driving behavior profiling and classification are performed based on the results of the clustering analysis. In the driving behavior early warning module, different levels of driving warnings are issued for vehicles with different driving behavior profilings. In the road risk management module, different levels of road risk management measures are implemented based on the number of vehicles with different driving behavior profilings.

[0008] Furthermore, the drone data collection and monitoring module utilizes drones to automatically and continuously monitor the driving trajectories of vehicles on highways through pre-set drone flight paths, and collects vehicle trajectory data.

[0009] Furthermore, the data acquisition and transmission module, based on the LoRaWAN wireless communication protocol, establishes a wireless communication data transmission channel between the ground station and the UAV, enabling two-way communication between the ground station and the UAV, and transmitting UAV flight control commands and UAV monitoring data.

[0010] Furthermore, the driving trajectory analysis algorithm calculates the distribution coefficient of the center position of the driving vehicle based on the center position coordinates of the driving vehicle, constructs the distribution feature function of the center position of the driving vehicle, obtains the trajectory analysis function of the driving vehicle, and then predicts the trajectory of the driving vehicle through the trajectory distribution probability function of the driving vehicle.

[0011] Furthermore, the detailed process of predicting the trajectory of moving vehicles is as follows:

[0012] For moving vehicles, the driving path data collected by drones is used to analyze the vehicle's trajectory, and the coordinates of the vehicle's center position are represented as follows: ,in, Let x and y represent the horizontal and vertical coordinates of the vehicle's center position, respectively. The distribution of the vehicle's center position at the next moment differs depending on the driver's driving behavior. A characteristic function for the vehicle's center position distribution is defined. First, the vehicle's center position distribution coefficient is calculated using the following formula:

[0013]

[0014] in, Indicates the standard deviation of parts. Indicates the global standard deviation. Let represent the distribution coefficient. Based on the distribution coefficient, a characteristic function for the distribution of the center position of the moving vehicle is constructed, and the function is expressed as follows:

[0015]

[0016] in, express Characteristic function of the distribution of the center position of vehicles at any given time. The function representing the distribution of the x-coordinate of the center position of the moving vehicle. The function representing the distribution of the ordinate of the center position of the moving vehicle. , They represent The horizontal coordinate of the center position of the vehicle at all times The horizontal coordinate of the center position of the vehicle at any given time. They represent The vertical coordinate of the center position of the vehicle at all times The vertical coordinate of the center position of the vehicle at any given time. The derivative is represented by the characteristic function of the vehicle's center position distribution, which is used to construct the vehicle trajectory analysis function, as follows:

[0017]

[0018] in, This represents a function for analyzing the trajectory of a moving vehicle. express Parameters for analyzing the trajectory of vehicles at all times. express Characteristic function of the distribution of the center position of vehicles at any given time. Let represent the total time. Based on the vehicle trajectory analysis function, we obtain the analysis parameters. Based on these parameters, we calculate the vehicle trajectory probability distribution function using the following formula:

[0019]

[0020] in, express The coordinates of the center position of the vehicle at all times. express The coordinates of the center position of the vehicle at all times. express The proposed driving trajectory analysis algorithm analyzes vehicle trajectory parameters at any given time. Based on the vehicle trajectory distribution probability function, it analyzes and predicts the probability that the vehicle's center position will appear in different locations at subsequent times, thus predicting the vehicle's trajectory. The algorithm calculates the vehicle's center position distribution coefficient based on the vehicle's center position coordinates, constructs a vehicle center position distribution feature function based on this coefficient, obtains analysis parameters through the trajectory analysis function, and then uses the vehicle trajectory distribution probability function to obtain predicted vehicle trajectory data. This algorithm improves the accuracy of vehicle position prediction by predicting subsequent vehicle positions based on the distribution of the vehicle's trajectory, reducing the time required for vehicle monitoring.

[0021] Furthermore, the K-means profiling clustering algorithm uses a neural network to match the predicted vehicle trajectory data with neurons in the neural network, optimizes the neuron weights based on the matching results, iterates to obtain the initial number of clusters, and performs driving behavior profiling clustering analysis on the predicted vehicle trajectory data.

[0022] Furthermore, a cluster analysis of driving behavior profiles is performed on the vehicle trajectory prediction data. The detailed process is as follows:

[0023] Construct a neural network and input the vehicle trajectory prediction data. The vehicle trajectory data is represented as follows:

[0024]

[0025] in, This represents the set of vehicle trajectory prediction data. These represent the predicted vehicle trajectory data for the first segment, the predicted vehicle trajectory data for the second segment, and the predicted vehicle trajectory data for the third segment, respectively. Segment vehicle trajectory prediction data, the first The vehicle trajectory prediction data is input into the neural network. A neuron matching function is then constructed to match the vehicle trajectory prediction data with the neurons. The neuron matching function is expressed as follows:

[0026]

[0027] in, This indicates that the predicted trajectory data of the driving vehicle is compared with the first... Matching coefficients between neurons Indicates the first Segment vehicle trajectory prediction data and the first The weights between neurons are calculated, and the neuron with the highest matching coefficient is selected as the optimal neuron. This is then used in conjunction with the vehicle trajectory prediction data for further weight optimization. The neurons are peripheral neurons, and the update formula is as follows:

[0028]

[0029] in, Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. The first neuron Segment vehicle trajectory prediction data and the first The weights between neurons Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. The first neuron Segment vehicle trajectory prediction data and the first The weights between neurons Represented as the first Secondary weight optimization coefficients Indicates the first The optimal set of neuron weights during sub-weight optimization. Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. A set of weights, The 2-norm is used to iteratively match neurons and optimize weights based on the input of vehicle trajectory prediction data. After a predetermined number of iterations, the output value is used as the initial number of cluster centers for clustering. From all the vehicle trajectory prediction data sets, based on the iterative output values ​​of the neural network, the same number of cluster centers are selected for clustering operations. The clustering operation is represented as follows:

[0030]

[0031] in, Indicates the first The set of vehicle trajectory prediction data and the first Cluster reference values ​​for the initial cluster centers, Indicates the first One initial cluster center, Indicates the first The proposed K-means profiling clustering algorithm uses a neural network to match the vehicle trajectory prediction data with neurons in the neural network to obtain the optimal neurons. Based on the optimal neurons, the neuron weights are optimized. After reaching a certain number of iterations, the output result is used as the initial number of cluster centers for clustering the vehicle trajectory prediction data. This algorithm improves the accuracy of the clustering results and reduces the clustering time by using a neural network to calculate the initial number of cluster centers.

[0032] Furthermore, the driving behavior profile classification module, based on the driving behavior profile analysis results, classifies the driver's driving behavior into different driving behavior profiles. These different driving behavior profiles include three types: aggressive, stable, and conservative. The driving behavior is then clustered into these three types of driving behavior profiles.

[0033] Furthermore, the driving behavior warning module, for profiles of aggressive driving behavior, plays warning voice messages via drones and roadside warning horns to alert drivers to their driving behavior.

[0034] Furthermore, the road risk management module, based on the profiles of aggressive, stable, and conservative driving behaviors, identifies the number of drivers exhibiting aggressive driving behavior on a section of highway and assigns personnel to perform duties accordingly.

[0035] Beneficial effects

[0036] 1. The driving trajectory analysis algorithm proposed in this invention calculates the distribution coefficient of the center position of the driving vehicle based on the center position coordinates of the driving vehicle, constructs the distribution feature function of the center position of the driving vehicle based on the distribution coefficient of the center position of the driving vehicle, obtains the analysis parameters through the driving vehicle trajectory analysis function, and then obtains the predicted data of the driving vehicle trajectory using the driving vehicle trajectory distribution probability function. The driving trajectory analysis algorithm proposed in this invention improves the accuracy of the driving vehicle position prediction results and reduces the time for vehicle monitoring by predicting the subsequent vehicle position by obtaining the distribution of the driving vehicle trajectory.

[0037] 2. The K-means profiling clustering algorithm proposed in this invention utilizes a neural network to match vehicle trajectory prediction data with neurons in the neural network, obtaining the optimal neurons. Based on the optimal neurons, the neuron weights are optimized. After reaching the required number of iterations, the output result is used as the initial number of cluster centers for clustering the vehicle trajectory prediction data. The K-means profiling clustering algorithm proposed in this invention uses a neural network to calculate the initial number of cluster centers, improving the accuracy of the clustering results, reducing the clustering time, and increasing the speed of driving behavior profile clustering analysis. Attached Figure Description

[0038] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present invention, and those skilled in the art can obtain other drawings based on these drawings without creative effort.

[0039] Figure 1 This is a system schematic diagram of the present invention; Detailed Implementation

[0040] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative effort are within the scope of protection of the present invention.

[0041] To achieve the above objectives, this invention proposes a vehicle driving behavior profiling and analysis system based on the K-means algorithm. The system comprises a UAV data acquisition and monitoring module, a data acquisition and transmission module, a driving trajectory analysis module, a driving behavior profiling and analysis module, a driving behavior profiling and classification module, a driving behavior early warning module, and a road risk management module. In the UAV data acquisition and monitoring module, the user uses a UAV to collect driving trajectory data of vehicles traveling on the road. In the data acquisition and transmission module, a data signal transmission channel is established using a wireless communication protocol. In the driving trajectory analysis module, a driving trajectory analysis algorithm is proposed to predict the trajectory of traveling vehicles. In the driving behavior profiling and analysis module, a K-means profiling and clustering algorithm is proposed to perform driving behavior profiling and clustering analysis based on the predicted driving trajectory data. In the driving behavior profiling and classification module, driving behavior profiling and classification are performed based on the results of the clustering analysis. In the driving behavior early warning module, different levels of driving warnings are issued for vehicles with different driving behavior profilings. In the road risk management module, different levels of road risk management measures are implemented based on the number of vehicles with different driving behavior profilings.

[0042] Specifically, the drone data collection and monitoring module uses drones to automatically monitor the driving trajectories of vehicles on highways in real time through pre-set drone flight paths, and collects vehicle trajectory data.

[0043] Specifically, the data acquisition and transmission module, based on the LoRaWAN wireless communication protocol, establishes a wireless communication data transmission channel between the ground station and the drone, enabling two-way communication between the ground station and the drone, and transmitting drone flight control commands and drone monitoring data.

[0044] Specifically, the driving trajectory analysis algorithm calculates the vehicle center position distribution coefficient based on the vehicle's center position coordinates, constructs the vehicle center position distribution feature function, obtains the vehicle trajectory analysis function, and then predicts the vehicle trajectory using the vehicle trajectory distribution probability function. The detailed process is as follows:

[0045] For moving vehicles, the driving path data collected by drones is used to analyze the vehicle's trajectory, and the coordinates of the vehicle's center position are represented as follows: ,in, Let x and y represent the horizontal and vertical coordinates of the vehicle's center position, respectively. The distribution of the vehicle's center position at the next moment differs depending on the driver's driving behavior. A characteristic function for the vehicle's center position distribution is defined. First, the vehicle's center position distribution coefficient is calculated using the following formula:

[0046]

[0047] in, Indicates the standard deviation of parts. Indicates the global standard deviation. Let represent the distribution coefficient. Based on the distribution coefficient, a characteristic function for the distribution of the center position of the moving vehicle is constructed, and the function is expressed as follows:

[0048]

[0049] in, express Characteristic function of the distribution of the center position of vehicles at any given time. The function representing the distribution of the x-coordinate of the center position of the moving vehicle. The function representing the ordinate distribution of the center position of the moving vehicle. , They represent The horizontal coordinate of the center position of the vehicle at all times The horizontal coordinate of the center position of the vehicle at any given time. They represent The vertical coordinate of the center position of the vehicle at all times The vertical coordinate of the center position of the vehicle at any given time. The derivative is represented by the characteristic function of the vehicle's center position distribution, which is used to construct the vehicle trajectory analysis function, as follows:

[0050]

[0051] in, This represents a function for analyzing the trajectory of a moving vehicle. express Parameters for analyzing the trajectory of vehicles at all times. express Characteristic function of the distribution of the center position of vehicles at any given time. Let represent the total time. Based on the vehicle trajectory analysis function, we obtain the analysis parameters. Based on these parameters, we calculate the vehicle trajectory probability distribution function using the following formula:

[0052]

[0053] in, express The coordinates of the center position of the vehicle at all times. express The coordinates of the center position of the vehicle at all times. express The vehicle trajectory analysis parameters at each time point are used to analyze and predict the probability of the vehicle's center position appearing in different locations at subsequent times, based on the vehicle trajectory trajectory distribution probability function, thereby predicting the vehicle's trajectory.

[0054] Specifically, the K-means profiling clustering algorithm uses a neural network to match vehicle trajectory prediction data with neurons in the neural network. Based on the matching results, it optimizes the neuron weights, iteratively obtains the initial number of clusters, and performs driving behavior profiling clustering analysis on the vehicle trajectory prediction data. The detailed process is as follows:

[0055] Construct a neural network and input the vehicle trajectory prediction data. The vehicle trajectory data is represented as follows:

[0056]

[0057] in, This represents the set of vehicle trajectory prediction data. These represent the predicted vehicle trajectory data for the first segment, the predicted vehicle trajectory data for the second segment, and the predicted vehicle trajectory data for the third segment, respectively. Segment vehicle trajectory prediction data, the first The vehicle trajectory prediction data is input into the neural network. A neuron matching function is then constructed to match the vehicle trajectory prediction data with the neurons. The neuron matching function is expressed as follows:

[0058]

[0059] in, This indicates that the predicted trajectory data of the driving vehicle is compared with the first... Matching coefficients between neurons Indicates the first Segment vehicle trajectory prediction data and the first The weights between neurons are calculated, and the neuron with the highest matching coefficient is selected as the optimal neuron. This is then used in conjunction with the vehicle trajectory prediction data for further weight optimization. The neurons are peripheral neurons, and the update formula is as follows:

[0060]

[0061] in, Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. The first neuron Segment vehicle trajectory prediction data and the first The weights between neurons Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. The first neuron Segment vehicle trajectory prediction data and the first The weights between neurons Represented as the first Secondary weight optimization coefficients Indicates the first The optimal set of neuron weights during sub-weight optimization. Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. A set of weights, The 2-norm is used to iteratively match neurons and optimize weights based on the input of vehicle trajectory prediction data. After a predetermined number of iterations, the output value is used as the initial number of cluster centers for clustering. From all the vehicle trajectory prediction data sets, based on the iterative output values ​​of the neural network, the same number of cluster centers are selected for clustering operations. The clustering operation is represented as follows:

[0062]

[0063] in, Indicates the first The set of vehicle trajectory prediction data and the first Cluster reference values ​​for the initial cluster centers, Indicates the first One initial cluster center, Indicates the first For each set of vehicle trajectory prediction data, cluster the data to the initial cluster center with the smallest cluster reference value to obtain an initial cluster distribution. In each initial cluster distribution, select the center of each initial cluster distribution as a new cluster center and perform the clustering operation again to obtain a new cluster distribution. Repeat the steps until the new cluster centers converge, and the clustering operation is completed to obtain the clustering result of the vehicle trajectory prediction data set.

[0064] Specifically, the driving behavior profile classification module classifies drivers' driving behavior into different driving behavior profiles based on the analysis results. These different driving behavior profiles include three types: aggressive, stable, and conservative. The driving behavior is then clustered into these three types of driving behavior profiles.

[0065] Specifically, the driving behavior warning module uses drones and roadside horns to play warning messages to alert drivers to aggressive driving behaviors.

[0066] Specifically, the road risk management module, based on the profiles of aggressive, stable, and conservative driving behaviors, identifies the number of drivers exhibiting aggressive driving behavior on a section of highway and assigns personnel to perform duties accordingly.

[0067] In a specific embodiment, the user utilizes a drone to automatically and continuously monitor and collect vehicle trajectories on highways via a pre-set drone flight path. Based on the LoRaWAN wireless communication protocol, monitoring data and flight control commands are transmitted between the ground station and the drone. Next, a driving trajectory analysis algorithm is used to predict vehicle trajectories. Then, based on the predicted vehicle trajectory data, a K-means clustering algorithm is used, employing a neural network, to perform driving behavior profile clustering analysis on the predicted vehicle trajectory data. Following this, based on the driving behavior profile clustering analysis results, driving behavior profiles are categorized into three types: aggressive, stable, and conservative. Then, the drone and roadside warning horns are used to alert drivers to their driving behavior. Finally, the number of drivers exhibiting aggressive driving behavior in a section of highway is identified, and corresponding management measures are taken.

[0068] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0069] The various embodiments in this specification are described in a related manner. Similar or identical parts between embodiments can be referred to mutually. Each embodiment focuses on describing the differences from other embodiments. In particular, the system embodiments are basically similar to the method embodiments, so the description is relatively simple; relevant parts can be referred to the descriptions of the method embodiments.

[0070] The above description is merely a preferred embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention are included within the scope of protection of the present invention.

Claims

1. A vehicle driving behavior profiling and analysis system based on the K-means algorithm, characterized in that, It includes a drone data collection and monitoring module, a data collection and transmission module, a driving trajectory analysis module, a driving behavior profile analysis module, a driving behavior profile classification module, a driving behavior early warning module, and a road risk management module; In the drone data collection and monitoring module, users use drones to collect driving trajectory data of vehicles traveling on the road; In the data acquisition and transmission module, a data signal transmission channel is established using a wireless communication protocol; in the driving trajectory analysis module, a driving trajectory analysis algorithm is proposed to predict the trajectory of the driving vehicle. In the driving behavior profile analysis module, a K-means profile clustering algorithm is proposed to perform driving behavior profile clustering analysis based on the predicted driving trajectory data of the vehicle. In the driving behavior profile classification module, driving behavior profiles are classified based on the clustering analysis results; in the driving behavior warning module, different levels of driving warnings are issued for vehicles with different driving behavior profiles. In the road risk management module, different levels of road risk management measures are implemented based on the number of vehicles with different driving behavior profiles.

2. The vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 1, characterized in that, The drone data acquisition and monitoring module uses drones to automatically monitor the driving trajectories of vehicles on highways in real time through a pre-set drone flight path, and collects the driving trajectory data of the vehicles.

3. The vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 1, characterized in that, The acquisition and transmission module, based on the LoRaWAN wireless communication protocol, establishes a wireless communication data transmission channel between the ground station and the UAV, enabling two-way communication between the ground station and the UAV, and transmitting UAV flight control commands and UAV monitoring data.

4. The vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 1, characterized in that, The driving trajectory analysis algorithm calculates the distribution coefficient of the center position of the driving vehicle based on the center position coordinates of the driving vehicle, constructs the distribution feature function of the center position of the driving vehicle, obtains the trajectory analysis function of the driving vehicle, and then predicts the trajectory of the driving vehicle through the trajectory distribution probability function of the driving vehicle.

5. The vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 4, characterized in that, The detailed process for predicting the trajectory of the vehicle is as follows: For moving vehicles, the driving path data collected by drones is used to analyze the vehicle's trajectory, and the coordinates of the vehicle's center position are represented as follows: ,in, Let x and y represent the horizontal and vertical coordinates of the vehicle's center position, respectively. The distribution of the vehicle's center position at the next moment differs depending on the driver's driving behavior. A characteristic function for the vehicle's center position distribution is defined. First, the vehicle's center position distribution coefficient is calculated using the following formula: in, Indicates the standard deviation of parts. Indicates the global standard deviation. Let represent the distribution coefficient. Based on the distribution coefficient, a characteristic function for the distribution of the center position of the moving vehicle is constructed, and the function is expressed as follows: in, express Characteristic function of the distribution of the center position of vehicles at any given time. The function representing the distribution of the x-coordinate of the center position of the moving vehicle. The function representing the distribution of the ordinate of the center position of the moving vehicle. , They represent The horizontal coordinate of the center position of the vehicle at all times The horizontal coordinate of the center position of the vehicle at any given time. They represent The vertical coordinate of the center position of the vehicle at all times The vertical coordinate of the center position of the vehicle at any given time. The derivative is represented by the characteristic function of the vehicle's center position distribution, which is used to construct the vehicle trajectory analysis function, as follows: in, This represents a function for analyzing the trajectory of a moving vehicle. express Parameters for analyzing the trajectory of vehicles at all times. express Characteristic function of the distribution of the center position of vehicles at any given time. Let represent the total time. Based on the vehicle trajectory analysis function, we obtain the analysis parameters. Based on these parameters, we calculate the vehicle trajectory probability distribution function using the following formula: in, express The coordinates of the center position of the vehicle at all times. express The coordinates of the center position of the vehicle at all times. express The vehicle trajectory analysis parameters at each time point are used to analyze and predict the probability of the vehicle's center position appearing in different locations at subsequent times, based on the vehicle trajectory trajectory distribution probability function, thereby predicting the vehicle's trajectory.

6. The vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 1, characterized in that, The K-means portrait clustering algorithm uses a neural network to match the predicted vehicle trajectory data with neurons in the neural network, optimizes the neuron weights based on the matching results, iterates to obtain the initial number of clusters, and performs driving behavior portrait clustering analysis on the predicted vehicle trajectory data.

7. A vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 6, characterized in that, The detailed process of performing driving behavior profiling and cluster analysis on the predicted vehicle trajectory data is as follows: Construct a neural network and input the vehicle trajectory prediction data. The vehicle trajectory data is represented as follows: in, This represents the set of vehicle trajectory prediction data. These represent the predicted vehicle trajectory data for the first segment, the predicted vehicle trajectory data for the second segment, and the predicted vehicle trajectory data for the third segment, respectively. Segment vehicle trajectory prediction data, the first The vehicle trajectory prediction data is input into the neural network. A neuron matching function is then constructed to match the vehicle trajectory prediction data with the neurons. The neuron matching function is expressed as follows: in, This indicates that the predicted trajectory data of the driving vehicle is compared with the first... Matching coefficients between neurons Indicates the first Segment vehicle trajectory prediction data and the first The weights between neurons are calculated, and the neuron with the highest matching coefficient is selected as the optimal neuron. This is then used in conjunction with the vehicle trajectory prediction data for further weight optimization. The neurons are peripheral neurons, and the update formula is as follows: in, Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. The first neuron Segment vehicle trajectory prediction data and the first The weights between neurons Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. The first neuron Segment vehicle trajectory prediction data and the first The weights between neurons Represented as the first Secondary weight optimization coefficients Indicates the first The optimal set of neuron weights during sub-weight optimization. Indicates the first During the second weight optimization, the optimal neuron's surrounding area is the [number]th [number]. A set of weights, The 2-norm is used to iteratively match neurons and optimize weights based on the input of vehicle trajectory prediction data. After a predetermined number of iterations, the output value is used as the initial number of cluster centers for clustering. From all the vehicle trajectory prediction data sets, based on the iterative output values ​​of the neural network, the same number of cluster centers are selected for clustering operations. The clustering operation is represented as follows: in, Indicates the first The set of vehicle trajectory prediction data and the first Cluster reference values ​​for the initial cluster centers, Indicates the first One initial cluster center, Indicates the first For each set of vehicle trajectory prediction data, cluster the data to the initial cluster center with the smallest cluster reference value to obtain an initial cluster distribution. In each initial cluster distribution, select the center of each initial cluster distribution as a new cluster center and perform the clustering operation again to obtain a new cluster distribution. Repeat the steps until the new cluster centers converge, and the clustering operation is completed to obtain the clustering result of the vehicle trajectory prediction data set.

8. The vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 1, characterized in that, The driving behavior profile classification module classifies drivers' driving behavior into different driving behavior profiles based on the driving behavior profile analysis results. The different driving behavior profiles include three types: aggressive, stable, and conservative. The driving behavior is clustered into three types of driving behavior profiles: aggressive, stable, and conservative.

9. A vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 8, characterized in that, The driving behavior warning module, for drivers with aggressive driving behavior profiles, plays warning voice messages via drones and roadside horns to alert them to their driving behavior.

10. A vehicle driving behavior profiling and analysis system based on the K-means algorithm according to claim 1, characterized in that, The road risk management module identifies the number of drivers exhibiting aggressive driving behavior in a section of highway based on profiles of aggressive, stable, and conservative driving behaviors, and then assigns personnel to perform duties accordingly.

Images