Track portraying method and device for in-transit personnel on expressway and medium

Abstract

The embodiment of the invention discloses a track portraying method for in-transit personnel on a highway. The method comprises the following steps: distinguishing in-transit users from non-in-transit users according to mobile phone signaling data of highway users; using a DBSCAN clustering algorithm to cluster the track points of the in-transit users; constructing user traffic behavior feature portraits according to the track point categories included by the users; clustering the user trajectory similarity by using a K-means clustering algorithm, and constructing a user traffic behavior relationship portrait; and clustering the traffic behavior risk indexes of each user by using a K-means clustering method, and constructing a traffic behavior risk portrait of the user. According to the method, the traffic behavior characteristics of the in-transit personnel on the highway can be accurately described from multiple angles.

Description

Technical Field

[0001] The embodiments of the present invention relate to trajectory profiling research in the field of transportation, and in particular to a trajectory profiling method for people on the road on a highway. Background Art

[0003] As a new technology, trajectory profiling can provide accurate data support for traffic control by monitoring and analyzing traffic conditions, thereby effectively alleviating traffic problems. However, the current user trajectory profiling cannot fully reflect the behavioral characteristics of people on the highway, and the accuracy of the information provided needs to be improved. Summary of the invention

[0004] The present invention proposes a trajectory profiling method for people on the road on a highway to solve the above technical problems.

[0005] In a first aspect, an embodiment of the present invention provides a method for profiling trajectories of people on the highway, including:

[0006] Based on the mobile phone signaling data of highway users, distinguish between users on the way and users not on the way;

[0007] Using geometric matching method, the trajectory of the en-route user is matched to the highway network;

[0008] The DBSCAN clustering algorithm is used to cluster the matched trajectory points. Different categories correspond to different traffic behaviors. According to the trajectory point categories of each user, a user traffic behavior feature profile is constructed.

[0009] A trajectory similarity model is established according to the speed and time-space information of the trajectory; the similarity of the trajectories of different users after matching is calculated using the trajectory similarity model; and each similarity is clustered using the K-means clustering algorithm, and different categories correspond to different user relationships, thereby constructing a user traffic behavior relationship portrait;

[0010] According to the matched trajectory points, the traffic behavior risk index of each user is calculated; and the K-means clustering method is used to cluster the traffic behavior risk index of each user, and different categories correspond to different risk levels; the SVM model is trained according to the K-means classification results, and the trained model is used to identify the traffic behavior risk level, thereby constructing a user traffic behavior risk profile.

[0011] In a second aspect, an embodiment of the present invention further provides an electronic device, the electronic device comprising:

[0012] one or more processors;

[0013] a memory for storing one or more programs,

[0014] When the one or more programs are executed by the one or more processors, the one or more processors implement the method for profiling trajectories of people on the highway as described in any embodiment.

[0015] In a third aspect, an embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for profiling trajectories of people on the highway as described in any embodiment.

[0016] In summary, the embodiment of the present invention provides a trajectory profiling method for people on the highway. Based on mobile phone signaling data, the trajectory profiling method for people on the highway is jointly constructed from three perspectives: traffic behavior characteristics, traffic behavior relationships, and traffic behavior risks. It provides comprehensive and accurate data support for real-time monitoring of road conditions, analysis of vehicle behavior, detection of abnormal trajectories, and automatic driving path planning and decision-making. Specifically, in the traffic behavior characteristic profiling, the method uses the DBSCAN clustering algorithm to change the clustering parameters, and gradually identifies the user's speeding, normal driving, slow driving, and stopping traffic behaviors, which can more accurately analyze user behavior; in the traffic behavior relationship profiling, a trajectory similarity model based on the three factors of time, location, and speed is proposed, and the trajectory similarity of different users is calculated from dynamic and static perspectives, which can more accurately identify user relationships; in the traffic behavior risk profiling, reasonable risk indicators can be used to more accurately assess the user's risk status; thereby improving the information richness and accuracy of the entire user trajectory profiling. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0018] Figure 1 A flow chart of a method for profiling the trajectory of people on the highway provided by an example of the present invention;

[0019] Figure 2 A result diagram of the trajectory points of a user in the user traffic behavior feature portrait provided by the example of the present invention after the first behavior type clustering;

[0020] Figure 3 for Figure 2 The result diagram of the user's trajectory points after the second behavior type clustering;

[0021] Figure 4 for Figure 2The result diagram of the user's trajectory points after the third behavior type clustering;

[0022] Figure 5 The result graph of the K-means clustering algorithm used in the risk profile of user traffic behavior provided by the example of the present invention;

[0023] Figure 6 A schematic diagram of a confusion matrix of classification results of an SVM linear kernel function model for identifying user traffic behavior risks obtained by training based on K-means clustering results provided in an example of the present invention;

[0024] Figure 7 A schematic diagram of a confusion matrix of classification results of an SVM Gaussian kernel function model for identifying user traffic behavior risks obtained by training with K-means clustering results provided in an example of the present invention;

[0025] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be described clearly and completely below. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work belong to the scope of protection of the present invention.

[0027] Figure 1 The present invention provides a flowchart of a method for profiling the trajectory of people on the highway based on mobile phone signaling data. Figure 1 The method is executed by an electronic device and specifically comprises the following steps:

[0028] S110. Differentiate between on-the-way users and off-the-way users based on mobile phone signaling data of highway users.

[0029] This embodiment uses the mobile phone signaling data of highway users as a data source. Mobile phone signaling data relies on the mobile communication network, and its coverage covers almost all cities and major areas, and can comprehensively record the user's movement trajectory. In contrast, traditional trajectory data (such as GPS data) may be limited by factors such as device penetration rate and signal stability, and it is difficult to achieve the same coverage level. Mobile phone signaling data is a passively collected data, which is automatically recorded by the base station during the communication process with the mobile terminal, so data collection can be performed uninterruptedly around the clock. This continuous data collection method helps to capture the user's complete travel trajectory, avoiding the problems of large sampling intervals and discontinuous data that may exist in traditional trajectory data. Mobile phone signaling data not only contains the user's location information, but also contains multi-dimensional information such as time, user identification ID, and communication event type. This information provides a rich data source for the method of this application, which helps to deeply explore the user's travel behavior characteristics and social activity patterns. At the same time, the spatiotemporal resolution of mobile phone signaling data is high, which can accurately determine the specific time and location of the user's communication with the base station, providing a more sophisticated data analysis basis for this application.

[0030] In a specific implementation, after acquiring the mobile phone signaling data, this step first identifies and filters out the redundant data at the time of continuous recording. Then, the moving azimuth of the user from the previous location point to the next location point and the moving speed of the user from the previous location point to the next location point are calculated, and whether it is abnormal data is determined based on the numerical range of the moving azimuth and the moving speed; if it is abnormal data, it is deleted. Finally, the actual distance between the user's starting and ending points is calculated, and the average of all non-zero speeds is calculated, and whether the user is on the way is determined based on the actual distance between the starting and ending points and the average of non-zero speeds.

[0031] Exemplarily, if the longitude and latitude of adjacent recording moments are the same, it is determined to be redundant data, the intermediate records are deleted, and only the time records of the first and last consecutive occurrences of the same location point are retained. If the difference in the moving azimuth angle is greater than 100°, it is determined to be a turning point. If the trajectory shows continuous turning back, it is determined to be ping-pong data and is eliminated. If the moving speed exceeds 180km / h, it is determined to be drift data and is eliminated. If the actual distance between the user's starting and ending points is greater than 5km, and the average non-zero speed is greater than 20km / h, the user is judged to be an in-transit user, otherwise it is a non-in-transit user.

[0032] Among them, the azimuth angle α i The calculation formula is as follows:

[0033] α i =2arctan{sin(long i -lon i+1 )×cos(lat i ),cos(lat i)×sin(lat i+1 )-sin(lat i )×cos(lat i+1 )×cos(lon i -lon i+1 )} (1)

[0034] Moving azimuth Δα i The calculation formula is as follows:

[0035] Δα i =|α i+1 -α i | (2)

[0036] Actual distance between start and end point S max The calculation formula is as follows:

[0037]

[0038] Among them, lon i ,lat i are the latitude and longitude of the user's i-th location, lon1 and lat1 are the latitude and longitude of the user's starting point, lon N ,lat N is the latitude and longitude of the user's destination. i and α i+1 are the i-th and i+1-th positions of the user, respectively, Δα i Indicated by α i and α i+1 The calculated movement azimuth.

[0039] S120: Use a geometric matching method to match the trajectory of the en-route user to the highway network.

[0040] Since the trajectory points in the mobile phone signaling data may not completely overlap with the highway, this step matches the user trajectory in the mobile phone signaling data to the highway network to obtain a more accurate user trajectory, which is used as the processing object for subsequent operations.

[0041] In a specific implementation, first, the road network data can be downloaded from the OpenStreetMap official website, and the longitude and latitude of the road network points can be obtained through the arcGIS software layer selection, screening, and point construction operations in the editor. The road network points should be on the road network within the user distribution range, and the distance of the road network points along the road network should not exceed 0.1km.

[0042] Then, the actual straight-line distance between every two adjacent road network points is calculated, and the westernmost road network point is taken as the origin of the route. The cumulative distance from each road network point to the origin of the route is calculated, and the cumulative distance is used as the mileage of the road network point; and the road network points whose longitude differs from the trajectory point by no more than 0.01 are selected to calculate the actual straight-line distance between the two.

[0043] Finally, the road network point with the shortest actual straight-line distance to the trajectory point is selected as the matching point.

[0044] Match all the trajectory points of the same user, interpolate the matching trajectory with the road network points according to the road network point mileage of the matching points, and restore the user trajectory, that is, obtain the user trajectory matched on the road network.

[0045] S130. Cluster each matched trajectory point using the DBSCAN clustering algorithm, with different categories corresponding to different traffic behaviors; and construct a user traffic behavior feature profile based on the trajectory point category of each user.

[0046] This step analyzes the motion characteristics of the trajectory points by clustering, and constructs a characteristic profile of the user's traffic behavior based on the motion characteristics of the trajectory points.

[0047] In a specific implementation, first, the road network mileage positions of the matching points can be supplemented with a 1-minute time interval using a linear interpolation method to unify the time density.

[0048] Then, the DBSCAN clustering algorithm is used to select points one by one from the supplemented matching points in chronological order, whose time difference does not exceed 60s and whose spatial difference does not exceed 1km. If the number of points that meet the above two conditions is greater than or equal to 2, they are classified into a cluster, called slow points, and the other points are noise points, called moving points.

[0049] Next, the DBSCAN clustering algorithm is used again to select points with a time difference of no more than 60 seconds and a spatial difference of no more than 0.17 km from the slow points obtained by the first clustering in chronological order. If the number of points that meet the above two conditions is greater than or equal to 10, they are grouped into a cluster, called stay points, and the other points are noise points, called slow points.

[0050] Finally, the DBSCAN clustering algorithm is used again to select points with a time difference of no more than 60 seconds and a spatial difference of no more than 2 km from the moving points obtained by the first clustering in chronological order. If the number of points that meet the above two conditions is greater than or equal to 2, they are classified into a cluster, called normal points, and the other points are noise points, called speeding points.

[0051] After clustering is completed, the traffic behavior indicators of each user are calculated according to the distribution of stop points, slow-moving points, normal points and speeding points covered by each user in time and space, including travel duration, number of trips, speeding duration, number of speeding times, slow-moving duration, number of slow-moving times, travel distance, etc., and a frequency distribution histogram is drawn. The indicator calculation results and the histogram are used together as a portrait of the traffic behavior characteristics of the user group. Figure 2 , Figure 3 and Figure 4 The distribution diagram of each trajectory point type after the first, second, and third behavior type clustering of a certain user is shown. It can be seen that this method can vividly reflect the distribution of user behavior types.

[0052] S140: Establish a trajectory similarity model according to the speed and time-space information of the trajectory.

[0053] This step provides a new trajectory similarity calculation method for more accurately calculating the trajectory similarity of users.

[0054] In a specific implementation, first, the speed difference and the consistency of the speed direction can be used to evaluate the similarity of the dynamic characteristics of traffic behavior, that is, speed similarity; at the same time, the difference in road network mileage of matching points in each time interval can be used to evaluate the similarity of the static characteristics of traffic behavior, that is, spatial similarity; finally, the above speed similarity and spatial similarity are combined to establish a trajectory similarity model.

[0055] Optionally, the trajectory similarity model calculation formula is as follows:

[0056] DS j1,j2,i =|P j1,i -P j2,i | (4)

[0057]

[0058] K j1,j2 =KS j1,j2 ×KV j1,j2 (10)

[0059] Among them, DS j1,j2,i KS is the mileage difference between the matching points of user j1 and user j2 during the i-th time unit, j1,j2,i KS is the spatial similarity between user j1 and user j2 during the i-th time unit, j1,j2 is the spatial similarity between user j1 and user j2 during the study period, DV j1,j2,i KV is the speed difference between user j1 and user j2 during the i-th time unit, j1,j2,i is the speed similarity between user j1 and user j2 during the i-th time unit, KV j1,j2is the speed similarity between user j1 and user j2 during the study period, K j1,j2 is the comprehensive similarity between user j1 and user j2 during the study period.

[0060] S150, using the trajectory similarity model, calculating the similarities of the trajectories of different users after matching; and clustering each similarity using the K-means clustering algorithm, where different categories correspond to different user relationships, thereby constructing a user traffic behavior relationship portrait.

[0061] This step analyzes the relationship between users by clustering the trajectory similarities, and builds a relational portrait of user traffic behavior based on the user relationship.

[0062] In a specific implementation, first, a time-position sequence and a time-speed sequence can be created for each user in units of 5 minutes; then, the spatial difference and speed difference between all users in each time unit are obtained through matrix operations, and the spatial similarity and speed similarity are calculated based on the differences; finally, the comprehensive similarity is obtained based on the spatial similarity and speed similarity.

[0063] Then, using non-zero spatial similarity as input, the initial cluster center point is randomly selected, and the K-means clustering algorithm is used to divide these user relationships into four categories. The classification result when the silhouette coefficient is optimal is taken as the final classification result.

[0064] It should be noted that when calculating the similarity, in order to eliminate invalid similarity values ​​caused by inconsistent moving directions, the sign function Sgn() needs to be used to identify the moving direction. In addition, in order to avoid the generation of invalid similarity values ​​caused by long-term stays, when establishing the time-position sequence and time-speed sequence, the position and speed of the stay point are assigned 9999, and the filtering conditions are set accordingly when calculating the similarity.

[0065] S160. Calculate the traffic behavior risk index of each user based on the matched trajectory points; and use the K-means clustering method to cluster the traffic behavior risk index of each user, with different categories corresponding to different risk levels; train the SVM model based on the K-means grading results, and use the trained model to identify the traffic behavior risk level, thereby constructing a user traffic behavior risk profile.

[0066] In this step, a traffic behavior risk index is constructed, and the risk level of user behavior is analyzed by clustering, thereby constructing a characteristic profile of user traffic behavior.

[0067] In a specific implementation, first, a risk index of user traffic behavior is constructed. Optionally, speed stability and acceleration stability mean that the vehicle can maintain a relatively constant speed, avoid sudden acceleration or deceleration, reduce the driver's driving burden and driving fatigue, and the risk of traffic accidents such as rear-end collision and rollover caused by this will also be greatly reduced. Therefore, the user traffic behavior risk index should include speed stability and acceleration stability. In addition, this embodiment mainly evaluates the traffic behavior risk of the user during movement, so the value of zero speed should be eliminated. According to the speed stability, the evaluation indicators include the speed average, speed standard deviation, speed overspeed time ratio, and the speed difference between the 85th and 15th percentile speeds; according to the acceleration stability, the evaluation indicators include the positive acceleration average, the negative acceleration average, the positive acceleration standard deviation, the negative acceleration standard deviation, the acceleration standard deviation, the rapid acceleration time ratio, and the rapid deceleration time ratio.

[0068] Then, each risk index value of each user is calculated, and the risk index data is dimensionless processed using the Z-Score standardization method. The dimensionless risk index data is used as input, and the K-means clustering algorithm is used to divide the risks of people in transit into four risk levels. Figure 5 The clustering result of a risk level is shown as an example.

[0069] After clustering, the SVM model can be trained to identify the risk level of traffic behavior based on the K-means classification results. Optionally, the data set obtained by K-means clustering can be divided into a training set and a test set; then, a suitable kernel function and penalty coefficient are selected to build an SVM model; then, the model is trained using the training set to find the optimal hyperplane by maximizing the interval; finally, the trained model is used to predict the test set and analyze the accuracy of the model. Figure 6 and Figure 7 The SVM models of the two kernel functions and the confusion matrices of the classification results of the two models are shown respectively.

[0070] Finally, the above risk levels, or the above risk level set SVM model, are used as the user's traffic behavior risk profile.

[0071] In summary, this embodiment provides a trajectory profiling method for people on the highway. Based on mobile phone signaling data, it jointly constructs a trajectory profiling method for people on the highway from three perspectives: traffic behavior characteristics, traffic behavior relationships, and traffic behavior risks, and provides comprehensive and accurate data support for real-time monitoring of road conditions, analysis of vehicle behavior, detection of abnormal trajectories, and automatic driving path planning and decision-making. Specifically, in the traffic behavior characteristic profiling, this method uses the DBSCAN clustering algorithm to change the clustering parameters, and gradually identifies the user's speeding, normal driving, slow driving, and stopping traffic behaviors, which can more accurately analyze user behavior; in the traffic behavior relationship profiling, a trajectory similarity model based on time, location, and speed is proposed, and the trajectory similarity of different users is calculated from dynamic and static perspectives, which can more accurately identify user relationships; in the traffic behavior risk profiling, reasonable risk indicators can be used to more accurately assess the user's risk status; thereby improving the information richness and accuracy of the entire user trajectory profiling.

[0072] Figure 8 A schematic diagram of the structure of an electronic device provided by an embodiment of the present invention is shown in FIG. Figure 8 As shown, the device includes a processor 60, a memory 61, an input device 62 and an output device 63; the number of processors 60 in the device can be one or more. Figure 8 A processor 60 is taken as an example; the processor 60, the memory 61, the input device 62 and the output device 63 in the device can be connected by a bus or other means. Figure 8 The example of connecting through bus is taken in the following.

[0073] The memory 61 is a computer-readable storage medium that can be used to store software programs, computer executable programs and modules, such as program instructions / modules corresponding to the trajectory profiling method for people on the highway in the embodiment of the present invention. The processor 60 executes various functional applications and data processing of the device by running the software programs, instructions and modules stored in the memory 61, that is, realizing the above-mentioned trajectory profiling method for people on the highway.

[0074] The memory 61 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 61 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 61 may further include a memory remotely arranged relative to the processor 60, and these remote memories may be connected to the device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0075] The input device 62 may be used to receive input digital or character information and generate key signal input related to user settings and function control of the device. The output device 63 may include a display device such as a display screen.

[0076] An embodiment of the present invention further provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the method for profiling trajectories of people on the highway of any embodiment.

[0077] The computer storage medium of the embodiment of the present invention can adopt any combination of one or more computer-readable media. The computer-readable medium can be a computer-readable signal medium or a computer-readable storage medium. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of electricity, magnetism, light, electromagnetic, infrared, or semiconductor, or any combination of the above. More specific examples (non-exhaustive list) of computer-readable storage media include: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In this document, a computer-readable storage medium can be any tangible medium containing or storing a program, which can be used by an instruction execution system, a device or a device or used in combination with it.

[0078] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, which carry computer-readable program code. Such propagated data signals may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer-readable signal media may also be any computer-readable medium other than a computer-readable storage medium, which may send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0079] The program code embodied on the computer readable medium may be transmitted using any appropriate medium, including but not limited to wireless, wireline, optical fiber cable, RF, etc., or any suitable combination of the foregoing.

[0080] Computer program code for performing the operations of the present invention may be written in one or more programming languages ​​or a combination thereof, including object-oriented programming languages ​​such as Java, Smalltalk, C++, and conventional procedural programming languages ​​such as C or similar programming languages. The program code may be executed entirely on the user's computer, partially on the user's computer, as a separate software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In the case of a remote computer, the remote computer may be connected to the user's computer through any type of network, including a local area network (LAN) or a wide area network (WAN), or may be connected to an external computer (e.g., through the Internet using an Internet service provider).

[0081] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention.

Claims

1. A trajectory profiling method for people on the highway, characterized in that: include: Based on the mobile phone signaling data of highway users, distinguish between users on the way and users not on the way; Using geometric matching method, the trajectory of the en-route user is matched to the highway network; The DBSCAN clustering algorithm is used to cluster the matched trajectory points. Different categories correspond to different traffic behaviors. According to the trajectory point categories of each user, a user traffic behavior feature profile is constructed. Establishing a trajectory similarity model according to the speed and time-space information of the trajectory; and calculating the similarity of the trajectories of different users after matching using the trajectory similarity model; The K-means clustering algorithm is used to cluster the similarities. Different categories correspond to different user relationships, thus constructing a user traffic behavior relationship portrait. According to the matched trajectory points, the traffic behavior risk index of each user is calculated; The K-means clustering method is used to cluster the traffic behavior risk indicators of each user, and different categories correspond to different risk levels; The SVM model is trained according to the K-means classification results. The trained model is used to identify the traffic behavior risk level, thereby constructing a user traffic behavior risk profile.

2. The method according to claim 1, characterized in that The method of distinguishing between on-the-way users and off-the-way users based on the mobile phone signaling data of the highway users specifically includes: From the mobile phone signaling data of highway users, redundant data is filtered according to the user location and recording time; Calculate the user movement azimuth between adjacent trajectory points of the same user, and remove ping-pong data based on the user movement azimuth and the change of the movement azimuth; Calculate the moving speed of adjacent trajectory points of the same user and remove drift data based on the user's moving speed; The moving distance and the average non-zero speed of each user are calculated, and the users on the way and the users not on the way are distinguished according to the moving distance and the average non-zero speed.

3. The method according to claim 1, characterized in that: The geometric matching method is used to match the trajectory of the en-route user to the highway network, specifically including: Increase the density of road network points through arcGIS and extract the longitude and latitude data of road network points; Calculate the actual cumulative distance from each road network point to a certain end of the road network as the road network point mileage; For each user's trajectory point, calculate the actual distance from each road network point to the current trajectory point, and select the road network point with the shortest distance as the matching point of the current trajectory point; According to the road network point mileage of all matching points of the same user, the matching trajectory is interpolated with the road network points to obtain the matched user trajectory.

4. The method according to claim 1, characterized in that: The DBSCAN clustering algorithm is used to cluster the matched trajectory points, and different categories correspond to different traffic behaviors; According to the categories of trajectory points included in each user, a user traffic behavior feature profile is constructed, including: Use linear interpolation to unify the time intervals of each matching point; Taking the matching point data after a uniform time interval as input, the DBSCAN clustering algorithm is used to distinguish slow points and moving points; Taking the slow-moving point data as input, the DBSCAN clustering algorithm is used again to distinguish the stop points and slow-moving points; Taking the motion point data as input, the DBSCAN clustering algorithm is used again to distinguish normal points and speeding points; Determine the temporal and spatial distribution of each user's stop points, slow-moving points, normal points and speeding points, and calculate the traffic behavior indicators of each user to construct a user traffic behavior feature portrait.

5. The method according to claim 1, characterized in that The establishment of a trajectory similarity model based on the speed and temporal and spatial information of the trajectory specifically includes: The similarity of dynamic characteristics of traffic behavior is calculated by using the speed difference and the consistency of speed direction; The similarity of static characteristics of traffic behavior is calculated by using the difference in road network mileage of matching points in each time interval; A trajectory similarity model is established by combining the similarity of the dynamic features and the similarity of the static features.

6. The method according to claim 1, characterized in that The K-means clustering algorithm is used to cluster the similarities, and different categories correspond to different user relationships, so as to construct a user traffic behavior relationship portrait, which specifically includes: Using the K-means clustering algorithm, the similarities between different users are divided into four categories; The classification result with the best silhouette coefficient is selected to represent the relationships between various users, thereby constructing a traffic behavior relationship portrait.

7. The method according to claim 1, characterized in that The traffic behavior risk indicators of each user include: Speed ​​stability evaluation indicators, including speed average, speed standard deviation, speeding time ratio, and speed difference between the 85th and 15th percentile speeds; The acceleration stability evaluation indicators include the average positive acceleration, the average negative acceleration, the standard deviation of positive acceleration, the standard deviation of negative acceleration, the standard deviation of acceleration, the proportion of rapid acceleration time, and the proportion of rapid deceleration time.

8. The method according to claim 1, characterized in that The K-means clustering method is used to cluster the traffic behavior risk indicators of each user, and different categories correspond to different risk levels, specifically including: dimensionless processing of the risk indicator data of each user; using the dimensionless risk indicator data as input, and clustering the en-route users into four risk levels using the K-means clustering algorithm; Correspondingly, the SVM model is trained according to the K-means grading results, and the trained model is used to identify the risk level of traffic behavior, specifically including: according to the K-means grading results, statistics are collected on the distribution of traffic behavior risks of people on the way; the K-means grading results are used as label inputs, and the risk indicator data are used as feature inputs to train the SVM model to identify the risk level of people on the way.

9. An electronic device, characterized in that: include: one or more processors; a memory for storing one or more programs, When the one or more programs are executed by the one or more processors, the one or more processors implement the method for profiling trajectories of people on the highway as described in any one of claims 1-8.

10. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method for profiling the trajectory of people on the highway as described in any one of claims 1-8 is implemented.

Images