CN113460061B

CN113460061B - Driving behavior analysis method and system

Info

Publication number: CN113460061B
Application number: CN202110776974.9A
Authority: CN
Inventors: 程德心; 谢赤天; 郝江波; 周风明
Original assignee: Wuhan Kotei Informatics Co Ltd
Current assignee: Wuhan Kotei Informatics Co Ltd
Priority date: 2021-07-08
Filing date: 2021-07-08
Publication date: 2022-05-31
Anticipated expiration: 2041-07-08
Also published as: CN113460061A

Abstract

The invention provides a driving behavior analysis method and a system, wherein the method comprises the following steps: acquiring driving data of a vehicle sensor and a vehicle CAN (controller area network), and fusing the driving data based on a timestamp of a reference sensor; dividing a driving scene, analyzing the incidence relation between the driving scene and driving data, and extracting data segments from the fusion data; extracting target characteristic values from the data fragments, solving a Weber distribution parameter of each characteristic value, equally dividing the characteristic values into boxes, solving branch points in each box, connecting the same branch point in each box, and obtaining a Weber distribution curve corresponding to the characteristic values through linear regression fitting; and drawing a scatter diagram of each driving scene, and taking the characteristic data distributed in the preset interval range as a behavior intensive area under the corresponding scene to obtain the driving behavior characteristics under different driving scenes. Therefore, the driving behavior analysis process can be simplified, the accuracy of the analysis result is guaranteed, huge workload in the traditional analysis model training process is avoided, and the analysis efficiency is guaranteed.

Description

Driving behavior analysis method and system

Technical Field

The invention belongs to the field of automatic driving, and particularly relates to a driving behavior analysis method and system.

Background

Along with the rapid development of intelligent automobiles, the diversified functions of the automobiles can meet the daily use requirements of people, and automatic driving is gradually realized. Although the automatic driving automobile can meet the driving requirements based on users, for users in different countries and regions, the driving habits of the users have certain differences due to different traffic regulations, traffic environments and geographic environments. For the uniformly developed automatic driving system, it is difficult to adjust the functional parameters of the driving habits of users in different areas in a targeted manner.

At present, a common method is to construct a driving behavior analysis model based on a deep learning algorithm, and input collected driving data and a marked scene into the analysis model, so as to determine a corresponding relationship between a behavior parameter and the scene, so as to adjust an automatic driving function parameter. Although the method has high analysis efficiency, the sample amount required to be collected and labeled at the early stage is very large, so that the model training workload is large.

Disclosure of Invention

In view of this, embodiments of the present invention provide a driving behavior analysis method and system, which are used to solve the problem of a large workload in the existing driving behavior analysis method.

In a first aspect of embodiments of the present invention, there is provided a driving behavior analysis method, including:

acquiring driving data acquired by a vehicle sensor and a vehicle CAN bus, and fusing the driving data based on a timestamp of a reference sensor;

dividing a driving scene, analyzing the incidence relation between the driving scene and the driving data, and extracting a data segment corresponding to the driving scene from the fusion data;

extracting target characteristic values from the data fragments, solving a weber distribution parameter of each characteristic value, equally dividing the characteristic values into boxes, solving a quantile point in each box body, connecting the same quantile point in each box body, and obtaining a weber distribution curve corresponding to the characteristic values through linear regression fitting;

and drawing a scatter diagram of each driving scene, and taking the characteristic data distributed in the preset interval range as a behavior intensive area under the corresponding driving scene to obtain the driving behavior characteristics under different driving scenes.

In a second aspect of an embodiment of the present invention, there is provided a driving behavior analysis system including:

the data fusion module is used for acquiring driving data acquired by a vehicle sensor and a vehicle CAN bus and fusing the driving data based on a timestamp of a reference sensor;

the segment extraction module is used for dividing a driving scene, analyzing the incidence relation between the driving scene and the driving data and extracting a data segment corresponding to the driving scene from the fusion data;

the characteristic extraction module is used for extracting target characteristic values from the data fragments, solving the Weber distribution parameter of each characteristic value, equally dividing the characteristic values into boxes, solving the branch points in each box, connecting the same branch point in each box, and then obtaining a Weber distribution curve corresponding to the characteristic values through linear regression fitting;

and the behavior analysis module is used for drawing a scatter diagram of each driving scene, and taking the characteristic data distributed in the preset interval range as a behavior intensive area under the corresponding driving scene to obtain the driving behavior characteristics under different driving scenes.

In a third aspect of the embodiments of the present invention, there is provided an apparatus, including a memory, a processor, and a computer program stored in the memory and executable by the processor, where the processor executes the computer program to implement the steps of the method according to the first aspect of the embodiments of the present invention.

In a fourth aspect of the embodiments of the present invention, a computer-readable storage medium is provided, which stores a computer program, which when executed by a processor implements the steps of the method provided by the first aspect of the embodiments of the present invention.

In the embodiment of the invention, based on the driving data collected by multiple sensors, the data of the multiple sensors are fused through a time synchronization technology, multiple characteristic value combinations are designed aiming at different driving scenes, natural human driving behavior data under various driving scenes are analyzed by utilizing Weber distribution based on a statistical thought, a driving behavior data distribution model is obtained, reliable data support is provided for automatic driving function development and physical signs, the driving behavior analysis process can be simplified, the accuracy of an analysis result is ensured, and huge workload in the traditional training process based on a deep learning model is avoided. The reliability of the analysis result can be ensured and the analysis efficiency can be improved under different scenes.

Drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments or the prior art descriptions will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and those skilled in the art can also obtain other drawings according to the drawings without creative efforts.

Fig. 1 is a schematic flow chart of a driving behavior analysis method according to an embodiment of the present invention;

fig. 2 is a schematic structural diagram of a driving behavior analysis system according to an embodiment of the present invention.

Detailed Description

In order to make the objects, features and advantages of the present invention more obvious and understandable, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention, and it is obvious that the embodiments described below are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.

It should be understood that the terms "comprises" and "comprising," when used in this specification or claims and in the accompanying drawings, are intended to cover a non-exclusive inclusion, such that a process, method or system, or apparatus that comprises a list of steps or elements is not limited to the listed steps or elements. In addition, "first" and "second" are used to distinguish different objects, and are not used to describe a specific order.

Referring to fig. 1, a flow chart of a driving behavior analysis method according to an embodiment of the present invention includes:

s101, acquiring driving data acquired by a vehicle sensor and a vehicle CAN bus, and fusing the driving data based on a timestamp of a reference sensor;

the vehicle sensor generally comprises a camera, a laser radar, a millimeter wave radar, an IMU (inertial measurement unit) inertial navigation system and the like, driving data collected by the vehicle sensor and signal data transmitted in a vehicle CAN (controller Area network) bus are obtained, and driving behavior characteristics of users under different scenes are analyzed based on the data.

According to sampling frequency of each sensor, taking data with the highest sampling frequency as a reference sampling sensor, traversing other sensor data frames, performing time matching on a reference sampling sensor time stamp and other sensor data frames, taking a data frame with the closest time as sensor data under a current time stamp, and fusing the sensor data with the current time stamp data of the reference sampling sensor;

s102, dividing a driving scene, analyzing the incidence relation between the driving scene and the driving data, and extracting a data segment corresponding to the driving scene from the fusion data;

according to common scene types, the high-speed and urban roads are subjected to scene division, such as turning, straight going, car following, overtaking and the like.

The method comprises the steps of analyzing and determining the incidence relation between a driving scene and fusion data, designing a scene extraction rule based on a strong-correlation signal, and extracting a scene data fragment from the fusion data according to the extraction rule.

For a specific scene, the association between the fusion data and the scene is analyzed, for example, in the overtaking scene, data collected by the sensor changes within a certain period of time, and at this time, the fusion data corresponding to the period of time can be extracted as a data fragment in the overtaking scene. The association between the scene and the driving data can be determined according to common knowledge, or can be determined based on theoretical analysis, and when the management relationship is determined to exist, the corresponding data segment can be selected, and the data features are generally extracted.

S103, extracting target characteristic values from the data fragments, solving a Weber distribution parameter of each characteristic value, equally dividing the characteristic values into boxes, solving branch points in each box, connecting the same branch point in each box, and obtaining a Weber distribution curve corresponding to the characteristic values through linear regression fitting;

the target characteristic value refers to characteristic data associated with the scene, and generally, the target characteristic value is multiple, and multiple characteristic values are all associated with the scene.

And deleting abnormal characteristic values in each group of characteristic values by a quartile method.

Optionally, the target feature value is extracted from the data segment corresponding to the driving scene according to the driving scene analysis feature value combination. Based on the association of the scene with the feature data (sensor acquisition data), a combination of feature values in a specific scene is determined, and these feature values are extracted as target feature values.

The weber distribution can be used for counting the continuity probability density distribution of certain characteristic data, and the incidence relation between the driving behavior and the scene can be determined based on a weber distribution curve.

Specifically, the weber distribution parameters of each characteristic value are calculated, the weber distribution parameters comprise shape parameters and proportion parameters, then the characteristic values are subjected to equidistant box separation, 5%, 50% and 95% quantile points are calculated by using weber distribution in each box body, the same percentile point of each box body is connected, linear regression is used for fitting, and the weber distribution parameters are used as an integral weber distribution curve of the characteristic value data.

And S104, drawing a scatter diagram of each driving scene, and taking the characteristic data distributed in the preset interval range as behavior intensive areas under the corresponding driving scenes to obtain driving behavior characteristics under different driving scenes.

For each driving scene, drawing a scatter diagram of the characteristic data of the driving scene, simultaneously drawing a Weber distribution curve of the characteristic data, and combining the scatter diagram and the Weber distribution curve, taking the corresponding data distribution as a behavior intensive area of natural people in the current driving scene within a specific interval range, such as a 5% -95% distribution band, so as to determine the driving behaviors in different scenes, thereby providing a theoretical basis for automatic driving development.

In the embodiment, the weber distribution is utilized to analyze the natural driving behavior data in various scenes based on the statistical thought, an accurate natural driving behavior data distribution model is obtained, effective data support is provided for automatic driving, the problem that a traditional analysis model based on deep learning needs large workload to train the model can be avoided, the analysis efficiency can be guaranteed on the basis of guaranteeing the analysis accuracy, and the workload can be reduced.

It should be understood that, the sequence numbers of the steps in the foregoing embodiments do not imply an execution sequence, and the execution sequence of each process should be determined by its function and inherent logic, and should not constitute any limitation to the implementation process of the embodiments of the present invention.

Fig. 2 is a schematic structural diagram of a driving behavior analysis system according to an embodiment of the present invention, where the system includes:

the data fusion module 210 is configured to acquire driving data acquired by a vehicle sensor and a vehicle CAN bus, and fuse the driving data based on a timestamp of a reference sensor;

wherein the fusing the driving data based on the reference sensor timestamps comprises:

taking a sensor with the highest sampling frequency as a reference sensor, traversing the data frames of the other sensors, performing time matching on the time stamp of the reference sensor and the data frames of the other sensors, and taking the data frame which is closest to the time stamp of the reference sensor in the other sensors as current sensor data; current sensor data is fused with reference sensor current timestamp data.

The segment extraction module 220 is configured to divide a driving scene, analyze an association relationship between the driving scene and the driving data, and extract a data segment corresponding to the driving scene from the fusion data;

the incidence relation between the driving scene and the fusion data is analyzed and determined, scene extraction rule design is carried out based on the strong-correlation signal, and scene data fragments are extracted from the fusion data according to the extraction rule.

The feature extraction module 230 is configured to extract a target feature value from the data fragment, calculate a weber distribution parameter of each feature value, equally divide the feature values into boxes, calculate a branch point in each box, connect the same branch point in each box, and obtain a weber distribution curve corresponding to the feature value through linear regression fitting;

Optionally, the target feature value is extracted from the data segment corresponding to the driving scene according to the driving scene analysis feature value combination.

And the behavior analysis module 240 is configured to draw a scatter diagram of each driving scenario, and use the feature data distributed in the predetermined interval range as a behavior intensive area in the corresponding driving scenario to obtain driving behavior features in different driving scenarios.

It is understood that in one embodiment, the electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program performing the steps S101 to S104 as in the first embodiment, and the processor implements the driving behavior analysis when executing the computer program.

It will be understood by those skilled in the art that all or part of the steps in the method for implementing the above embodiments may be implemented by instructing the relevant hardware through a program, where the program may be stored in a computer-readable storage medium, and when the program is executed, the program includes steps S101 to S104, and the storage medium includes, for example, ROM/RAM.

It is clear to those skilled in the art that, for convenience and brevity of description, the specific working processes of the above-described systems, apparatuses and units may refer to the corresponding processes in the foregoing method embodiments, and are not described herein again.

In the above embodiments, the descriptions of the respective embodiments have respective emphasis, and reference may be made to the related descriptions of other embodiments for parts that are not described or illustrated in a certain embodiment.

The above-mentioned embodiments are only used for illustrating the technical solutions of the present invention, and not for limiting the same; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some technical features may be equivalently replaced; and such modifications or substitutions do not depart from the spirit and scope of the corresponding technical solutions of the embodiments of the present invention.

Claims

1. A driving behavior analysis method is characterized in that,

extracting target characteristic values from the data fragments, solving a Weber distribution parameter of each characteristic value, equally dividing the characteristic values into boxes, solving branch points in each box, connecting the same branch point in each box, and obtaining a Weber distribution curve corresponding to the characteristic values through linear regression fitting;

and drawing a scatter diagram of each driving scene, and taking the characteristic data distributed in the preset interval range as a behavior intensive area under the corresponding driving scene to obtain driving behavior characteristics under different driving scenes.

2. The method of claim 1, wherein the fusing the driving data based on the reference sensor's timestamp comprises:

taking a sensor with the highest sampling frequency as a reference sensor, traversing the data frames of the other sensors, performing time matching on the time stamp of the reference sensor and the data frames of the other sensors, and taking the data frame which is closest to the time stamp of the reference sensor in the other sensors as current sensor data;

current sensor data is fused with reference sensor current timestamp data.

3. The method of claim 1, wherein the analyzing the association relationship between the driving scene and the driving data, and the extracting the data segment corresponding to the driving scene from the fusion data comprises:

analyzing and determining the incidence relation between the driving scene and the fusion data;

designing a scene extraction rule based on the strongly correlated signal;

and extracting scene data segments from the fusion data according to the extraction rule.

4. The method of claim 1, wherein extracting the target feature value from the data segment further comprises:

and deleting the abnormal characteristic values in each group of characteristic values by a quartile method.

5. The method of claim 1, wherein extracting the target feature value from the data segment further comprises:

and analyzing the characteristic value combination according to the driving scene, and extracting a target characteristic value from the data segment corresponding to the driving scene.

6. A driving behavior analysis system, comprising:

7. The system of claim 6, wherein the fusing the driving data based on the reference sensor time stamps comprises:

current sensor data is fused with reference sensor current timestamp data.

8. The system of claim 6, wherein the analyzing the association relationship between the driving scene and the driving data, and the extracting the data segment corresponding to the driving scene from the fusion data comprises:

and analyzing and determining the incidence relation between the driving scene and the fusion data, designing a scene extraction rule based on a strong-correlation signal, and extracting a scene data segment from the fusion data according to the extraction rule.

9. A terminal device comprising a memory, a processor and a computer program stored in said memory and executable on said processor, characterized in that said processor, when executing said computer program, carries out the steps of a driving behavior analysis method according to any one of claims 1 to 5.

10. A computer-readable storage medium, in which a computer program is stored which, when executed, carries out the steps of a driving behavior analysis method according to any one of claims 1 to 5.