CN114954489A

CN114954489A - Method and device for identifying driving behaviors and styles of automobile

Info

Publication number: CN114954489A
Application number: CN202210449573.7A
Authority: CN
Inventors: 吉天成; 万振华; 张海春
Original assignee: Open Source Network Security Internet Of Things Technology Wuhan Co ltd; Huazhong University of Science and Technology
Current assignee: Open Source Network Security Internet Of Things Technology Wuhan Co ltd; Huazhong University of Science and Technology
Priority date: 2022-04-26
Filing date: 2022-04-26
Publication date: 2022-08-30

Abstract

The invention discloses a method and a device for identifying driving behaviors and styles of automobiles, wherein the identification method comprises the following steps: acquiring running data of a vehicle in a running process; extracting first judgment data in the driving data at the vehicle end, and judging the first judgment data according to a discrimination rule to determine driving behaviors corresponding to the first judgment data; and transmitting second determination data in the traveling data to the server side to determine the second determination data using a driving style recognition model of the server side to determine a driving style corresponding to the second determination data. The invention can reduce the occupation of the calculation resources of the server end, can identify some driving behaviors with higher requirements on timeliness at the vehicle end, and is beneficial to the vehicle to judge the corresponding driving behaviors in time.

Description

Method and device for identifying driving behaviors and styles of automobile

Technical Field

The invention relates to the technical field of auxiliary driving, in particular to a method and a device for identifying driving behaviors and styles of an automobile.

Background

In recent years, the technology of internet of things is rapidly developed and applied to various industries, in the field of automobiles, more and more automobile access networks form an internet of vehicles, and the automobile access networks can provide navigation, remote upgrading and other information services for automobile owners, but still face many traditional driving safety problems. In the driving process, a driver of the automobile may have various dangerous driving behaviors, and the safety threats to the driver and other road users in various aspects such as life, property and the like can be caused.

Therefore, in order to find a danger in advance and avoid an accident, it is necessary to analyze and identify the driving behavior and driving style of the driver. At present, the method for identifying the driving behavior and the driving style of the automobile mainly uploads the related driving data to a cloud server, and the driving data is analyzed and processed by the cloud server, so that the corresponding driving behavior and style are judged.

Therefore, there is a need to provide a new method and apparatus for identifying driving behaviors and styles of automobiles to solve the above problems.

Disclosure of Invention

The invention aims to provide a method and a device for identifying driving behaviors and styles of automobiles, electronic equipment and a computer readable storage medium, which can reduce the occupation of computing resources of a server end, can identify some driving behaviors with higher requirements on timeliness at a vehicle end and are beneficial to timely judgment of corresponding driving behaviors by the vehicle.

In order to achieve the above object, the present invention provides a method for identifying driving behaviors and styles of an automobile, comprising:

acquiring running data of a vehicle in a running process;

extracting first judgment data in the driving data at a vehicle end, and judging the first judgment data according to a discrimination rule to determine driving behaviors corresponding to the first judgment data; and

transmitting second determination data in the traveling data to a server to determine the second determination data using a driving style recognition model of the server to determine a driving style corresponding to the second determination data.

Optionally, the driving data comprises:

linear velocity, linear acceleration, angular velocity, angular acceleration, and ACC state of vehicle travel.

Optionally, the discrimination rule includes:

judging that the driving behavior meets the characteristics of rapid acceleration or rapid deceleration according to the linear acceleration;

judging that the driving behavior conforms to the characteristic of sharp turning according to the angular acceleration;

judging that the driving behavior accords with the characteristics of overspeed driving according to the linear speed;

judging that the driving behavior conforms to the characteristics of flameout and sliding according to the linear speed and the ACC state;

and judging that the driving behavior accords with the characteristics of idle preheating or ultra-long idling according to the linear speed, the ACC state and the duration.

Optionally, the driving style identification model is a k-means clustering model, and a plurality of corresponding sampling data points are generated according to second determination data;

the "determination of the second determination data by the driving style recognition model" includes:

a. setting a corresponding number of clustering centers according to the preset type of the driving style;

b. dividing a plurality of sampling data points corresponding to the second judgment data into clusters represented by the cluster centers closest to each other;

c. calculating a mean value for the sampled data points within each of the divided clusters to re-determine the cluster center;

d. and repeating the steps b and c until an iteration termination condition is met, and determining the type of the driving style represented by each cluster.

Optionally, the driving style recognition model comprises a plurality of classification models constructed based on a random forest algorithm, and each classification model is trained according to different types of sample data;

inputting the second determination data to the corresponding classification model according to the data type of the second determination data;

the classification model classifies the input second determination data according to the type of the driving style to output a classification result;

and performing weighted fusion on each classification result according to preset weight, and comparing the weighted results of different classification results to judge the type of the driving style.

Optionally, the "training each classification model according to sample data of different classes respectively" includes:

acquiring different types of the sample data;

preprocessing the sample data to extract feature data;

and inputting the different types of feature data into the corresponding classification models for training.

Optionally, the sample data types include angular acceleration, angular velocity and linear acceleration;

the data type of the second judgment data is consistent with the type of the sample data;

and the classification model outputs classification results according to the angular acceleration data, the angular velocity data and the linear acceleration data in the second judgment data respectively.

In order to achieve the above object, the present invention further provides a device for recognizing driving behaviors and styles of a vehicle, comprising:

the acquisition module is used for acquiring running data of the vehicle in the running process;

the behavior identification module is used for extracting first judgment data in the driving data at a vehicle end and judging the first judgment data according to a discrimination rule to determine driving behaviors corresponding to the first judgment data;

the transmission module is used for transmitting second judgment data in the driving data to a server side;

and the style recognition module is used for judging the second judgment data on the basis of a driving style recognition model at the server so as to determine the driving style corresponding to the second judgment data.

In order to achieve the above object, the present invention also provides an electronic device, including:

a processor;

a memory having stored therein executable instructions of the processor;

wherein the processor is configured to perform the method of identifying driving behavior and style of a car as described above via execution of the executable instructions.

To achieve the above object, the present invention also provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, implements the method for recognizing the driving behavior and style of a vehicle as described above.

The invention also provides a computer program product or computer program comprising computer instructions stored in a computer readable storage medium. The processor of the electronic device reads the computer instructions from the computer readable storage medium, and the processor executes the computer instructions, so that the electronic device executes the method for identifying the driving behavior and style of the automobile.

The invention processes the first judgment data in the running data at the vehicle end and processes the second judgment data in the running data at the server end, wherein the first judgment data in the running data is extracted at the vehicle end and judged according to the distinguishing rule to determine the corresponding driving behavior, and the second judgment data is judged on the server end based on the driving style identification model to determine the corresponding driving style, thereby avoiding the condition that all identification processes are processed at the server end, reducing the occupation of the computing resources at the server end, and identifying some driving behaviors with higher requirements on timeliness at the vehicle end so as to make timely judgment on the corresponding driving behavior by the vehicle.

Drawings

Fig. 1 is a data flow diagram of a method for identifying driving behaviors and styles of a vehicle according to an embodiment of the present invention.

FIG. 2 is a flowchart of a method for identifying driving style by using a k-means clustering model according to an embodiment of the present invention.

FIG. 3 is a flowchart of a method for identifying driving style by using a classification model constructed by a random forest algorithm according to an embodiment of the present invention.

Fig. 4 is a schematic block diagram of a device for recognizing driving behaviors and styles of a car according to an embodiment of the present invention.

FIG. 5 is a schematic block diagram of an electronic device according to an embodiment of the invention.

Detailed Description

In order to explain the technical contents, structural features, objects and effects of the present invention in detail, the following description is made in conjunction with the embodiments and the accompanying drawings.

Referring to fig. 1, an embodiment of the present invention discloses a method for identifying driving behaviors and styles of an automobile, including:

acquiring running data of a vehicle in a running process;

extracting first judgment data in the driving data at the vehicle end, and judging the first judgment data according to a discrimination rule to determine driving behaviors corresponding to the first judgment data; and

the second determination data in the travel data is transmitted to the server side to determine the second determination data using the driving style recognition model of the server side to determine the driving style corresponding to the second determination data.

The invention processes the first judgment data in the driving data at the vehicle end, and processes the second judgment data in the driving data at the server end, wherein, the first judgment data in the driving data is extracted at the vehicle end, and the first judgment data is judged according to the identification rule to determine the corresponding driving behavior, and the second judgment data is judged based on the driving style identification model at the server end to determine the corresponding driving style, thereby avoiding the situation that all identification processes are processed at the server end, reducing the occupation of the calculation resources at the server end, and recognizing some driving behaviors with higher requirements on timeliness at the vehicle end so as to make timely judgment on the corresponding driving behaviors by the vehicle.

Specifically, the travel data includes: linear velocity, linear acceleration, angular velocity, angular acceleration, and ACC state of vehicle travel. The linear acceleration, the angular velocity and the angular acceleration can be obtained by a linear acceleration sensor, an angular acceleration sensor and an angular velocity sensor which are arranged on the vehicle; the linear velocity can be obtained by a GPS module, and the GPS module can also obtain time data; the ACC state CAN be acquired by reading CAN bus data through an OBD interface at the vehicle end.

It is understood that the ACC state may be 1 or 0, with an ACC state of 1 indicating that the vehicle engine is on and an ACC state of 0 indicating that the vehicle engine is off.

Specifically, the first determination data includes a linear velocity, a linear acceleration, an angular acceleration, and an ACC state.

Further, the discrimination rules include: judging that the driving behavior accords with the characteristics of rapid acceleration or rapid deceleration according to the linear acceleration; judging that the driving behavior accords with the characteristic of sharp turning according to the angular acceleration; judging that the driving behavior accords with the characteristics of overspeed driving according to the linear speed; judging that the driving behavior accords with the characteristics of flameout and sliding according to the linear speed and the ACC state; and determining that the driving behavior conforms to the characteristics of idle preheating or ultra-long idle according to the linear speed, the ACC state and the duration.

When the linear acceleration exceeds a preset rapid acceleration threshold, the driving behavior can be determined to accord with the characteristics of the rapid acceleration; when the linear acceleration exceeds a preset rapid deceleration threshold, the driving behavior can be determined to accord with the characteristics of rapid deceleration; when the angular acceleration exceeds a preset sharp turning threshold, the driving behavior can be determined to accord with the characteristic of sharp turning; when the linear speed exceeds a preset overspeed threshold value, the driving behavior can be determined to accord with the characteristics of overspeed driving; when the linear speed exceeds a preset driving threshold and the ACC state is 0, the driving behavior can be determined to be in accordance with the flameout sliding characteristic; when the linear speed is lower than a preset action threshold, the ACC state is 1 and the duration time exceeds a preset time, the driving behavior can be determined to accord with the characteristic of ultra-long idling; when the linear speed is lower than a preset action threshold value, the ACC state is 1 and the duration time is within a preset preheating time, the driving behavior can be determined to be in accordance with the characteristic of idle preheating.

The identification of the driving behaviors can be realized by judging based on simple rules, the requirement on the computing capacity of the processor is low, the identification can be completed by the processor arranged at the vehicle end, the identification of the driving behaviors at the vehicle end is beneficial to reducing the computing pressure at the server end, and the vehicle end can timely acquire the driving behaviors with the requirements on timeliness.

Specifically, the vehicle end is provided with a network transmission module, and second determination data for driving style recognition is transmitted to the server end through the network transmission module.

In some embodiments, the driving style identification model is a k-means cluster model, and the corresponding sampling data points are generated according to the second determination data.

Referring to fig. 2, the "determination of the second determination data by the driving style recognition model" includes:

a. and setting corresponding number of clustering centers according to the preset types of the driving styles.

b. And dividing a plurality of sampling data points corresponding to the second judgment data into clusters represented by the nearest cluster centers respectively.

c. And calculating a mean value of the sampling data points in each divided cluster to determine a cluster center again.

Specifically, the second determination data may be a linear velocity or a linear acceleration acquired in units of time, or may be a linear velocity or a linear acceleration acquired in units of distance.

Further, an initial clustering center is selected according to the distribution condition of the sampling data points, and in order to obtain a better clustering effect, the initial clustering centers can be dispersedly arranged in the distribution interval of the sampling data points, for example, when the sampling data points of the linear velocity are distributed between 20km/h and 90km/h, and the preset types of the driving styles are three, the initial clustering centers can be arranged to be 30km/h, 50km/h and 80 km/h.

It is understood that the distance referred to in step b is a euclidean distance; the iteration termination condition is that the clustering center is converged or the iteration frequency reaches the preset frequency.

Two examples are given below:

when the second judgment data is the linear velocity collected every second, the number of the clustering centers can be set to be 3, the clustering centers respectively correspond to three driving styles of robust type, aggressive type and fatigue type, after a clustering iteration termination condition is met, a plurality of sampling data points form three clusters, each cluster respectively corresponds to three driving styles of robust type, aggressive type and fatigue type, wherein the cluster with higher linear velocity is divided into the aggressive driving style, the cluster with lower linear velocity is divided into the fatigue driving style, and the cluster with moderate linear velocity is divided into the robust driving style.

When the second judgment data is linear acceleration collected every second, the number of the clustering centers can be set to be 3, the clustering centers respectively correspond to three driving styles of a steady type driving style, an acceleration aggressive driving style and a deceleration aggressive driving style, after a clustering iteration termination condition is met, a plurality of sampling data points form three clusters, each cluster respectively corresponds to the three driving styles of the steady type driving style, the acceleration aggressive driving style and the deceleration aggressive driving style, wherein the cluster with the linear acceleration near 0 is divided into the driving style of the steady type, the cluster with the linear acceleration which is a positive value and is higher is divided into the driving style of the acceleration aggressive driving style, and the cluster with moderate linear acceleration is divided into the driving style of the deceleration aggressive driving style.

In some embodiments, the driving style recognition model comprises a plurality of classification models constructed based on a random forest algorithm, and each classification model is trained according to different types of sample data.

Specifically, the "training of each classification model according to sample data of different classes respectively" includes:

acquiring different types of sample data;

preprocessing the sample data to extract feature data;

and inputting different types of feature data into corresponding classification models for training.

The types of the sample data can comprise angular acceleration, angular velocity and linear acceleration, corresponding classification models can be trained according to different types of sample data, and decision accuracy can be improved by using a plurality of classification models to make driving style decisions.

Further, the sample data is preprocessed through an MMA (Multiscale multi-fractional analysis) algorithm to generate a Hurst surface map, so that characteristic data is extracted, and the characteristic data reflects the fluctuation trend of the sample data.

Referring to fig. 3, the "determination of the second determination data by the driving style recognition model" includes:

s41, the second determination data is input to the corresponding classification model according to the data type of the second determination data.

S42, the classification model classifies the input second determination data according to the driving style to output a classification result.

And S43, performing weighted fusion on each classification result according to preset weight, and comparing the weighted results of different classification results to judge the type of the driving style.

Specifically, the data type of the second determination data is identical to the type of the sample data, and if the data types are angular acceleration, angular velocity, and linear acceleration, components of the data on the x axis and the y axis on the driving plane are extracted to avoid interference from the z axis component.

Further, the preset weight may be set according to the accuracy of the classification model, or may be set according to the emphasis of the data type, for example, the weight of the linear acceleration may be set to be higher than the weight of the angular velocity and the angular acceleration.

An example is given below:

when the second determination data is linear acceleration, angular velocity and angular acceleration acquired every second and the types of the driving styles are classified into robust type, aggressive type and fatigue type, the linear acceleration, the angular velocity and the angular acceleration are respectively input into the corresponding classification models, the classification models can respectively output classification results according to the input data, and then each classification result is subjected to weighted fusion according to preset weights, if the classification result output by the linear acceleration is robust type, the classification result output by the angular velocity is also robust type, and the classification result output by the angular acceleration is aggressive type, if the classification result is subjected to fusion determination according to an equal weight method, the driving style can be determined to be robust type.

Referring to fig. 4, the present invention further provides a device for identifying driving behaviors and driving styles of a vehicle, including:

the acquiring module 100 is used for acquiring running data of the vehicle in a running process.

The behavior identification module 200 is configured to extract first determination data from the driving data at the vehicle end, and determine the first determination data according to a recognition rule to determine a driving behavior corresponding to the first determination data.

And a transmission module 300, configured to transmit the second determination data in the driving data to the server.

A style recognition module 400 for determining the second determination data based on the driving style recognition model at the server side to determine a driving style corresponding to the second determination data.

The invention can avoid the condition that the identification processes of the driving behaviors and the driving styles are processed at the server end, reduce the occupation of the calculation resources of the server end, and can identify some driving behaviors with higher requirements on timeliness at the vehicle end so that the vehicle can judge the corresponding driving behaviors in time.

Referring to fig. 5, the present invention further provides an electronic device, including:

a processor 40;

a memory 50 having stored therein executable instructions of the processor 40;

wherein the processor 40 is configured to perform the method of identifying driving behaviour and style of the car as described above via execution of executable instructions.

The present invention also provides a computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the method for identifying driving behaviour and style of a vehicle as described above.

The embodiment of the invention also discloses a computer program product or a computer program, which comprises computer instructions, and the computer instructions are stored in a computer readable storage medium. The processor of the electronic device reads the computer instructions from the computer-readable storage medium, and the processor executes the computer instructions, so that the electronic device performs the recognition method of the driving behavior and style of the automobile as described above.

It should be understood that in the embodiments of the present invention, the Processor may be a Central Processing Unit (CPU), and the Processor may also be other general purpose processors, Digital Signal Processors (DSP), Application Specific Integrated Circuits (ASIC), Field-Programmable Gate arrays (FPGA) or other Programmable logic devices, discrete Gate or transistor logic devices, discrete hardware components, and the like. A general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware associated with computer program instructions, and the programs can be stored in a computer readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. The storage medium may be a magnetic disk, an optical disk, a Read-Only Memory (ROM), a Random Access Memory (RAM), or the like.

The above disclosure is only a preferred embodiment of the present invention, and should not be taken as limiting the scope of the invention, so that the appended claims are intended to cover all such modifications and changes as fall within the true spirit of the invention.

Claims

1. A method for recognizing the driving behavior and style of an automobile is characterized by comprising the following steps:

acquiring running data of a vehicle in a running process;

transmitting second determination data among the travel data to a server side to determine the second determination data using a driving style recognition model of the server side to determine a driving style corresponding to the second determination data.

2. The method for recognizing the driving behavior and style of a vehicle according to claim 1, wherein the driving data includes:

3. The method for recognizing the behavior and style of driving of an automobile according to claim 2, wherein the discrimination rule comprises:

and determining that the driving behavior conforms to the characteristics of idle preheating or ultra-long idle according to the linear speed, the ACC state and the duration.

4. The method for recognizing the behavior and style of driving of a vehicle according to claim 1,

the driving style identification model is a k-means clustering model, and a plurality of corresponding sampling data points are generated according to the second judgment data;

5. The method for recognizing the behavior and style of driving of a vehicle according to claim 1,

the driving style recognition model comprises a plurality of classification models constructed based on a random forest algorithm, and each classification model is trained according to different types of sample data;

6. The method according to claim 5, wherein the training of each classification model according to different types of sample data comprises:

acquiring different types of the sample data;

preprocessing the sample data to extract feature data;

7. Method for the recognition of driving behavior and style of a car according to claim 6,

the types of the sample data comprise angular acceleration, angular velocity and linear acceleration;

8. An apparatus for recognizing driving behavior and style of a vehicle, comprising:

the acquisition module is used for acquiring the driving data of the vehicle in the driving process;

9. An electronic device, comprising:

a processor;

a memory having stored therein executable instructions of the processor;

wherein the processor is configured to perform the method of identifying driving behavior and style of a car according to any one of claims 1-7 via execution of the executable instructions.

10. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out a method for identifying a driving behaviour and style of a vehicle according to any one of claims 1 to 7.