CN108995653A - A kind of driver's driving style recognition methods and system - Google Patents

Abstract

The invention relates to a method and system for identifying a driver's driving style, which belongs to the technical field of automobile intelligent interaction, and solves the problems of low identification accuracy and poor practicability of the driver's driving style in the prior art. A method for identifying a driver's driving style, the method comprising the following steps: collecting driver operation information and vehicle driving information; and initially identifying the driver's driving style according to the collected driver operation information and vehicle driving information to obtain the driver's driving style Preliminary recognition results; according to the obtained preliminary recognition results of the driver's driving style, change the vehicle state; according to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data, further identify the result of the driver's driving style. The invention realizes the accurate identification of the driver's driving style and has strong practicability.

Description

A driver's driving style recognition method and system

technical field

The invention relates to the technical field of automobile intelligent interaction, in particular to a driver's driving style recognition method and system.

Background technique

With the start of car intelligence, people's demand for a good car experience makes people hope that cars will understand themselves more and more, and customize corresponding service content and assisted driving according to their own status and needs. Accurately identified driver's driving style plays an extremely important role in providing drivers with more humanized services and safer and more comfortable assisted driving.

The driver's driving style recognition method at this stage mainly identifies the driver's driving style by passively detecting and analyzing the driver's driving data; however, due to the uncertainty of the data obtained by this method, instability and Inconvenience makes the driver's driving style identified by this method have low accuracy and is easily affected by external factors, so the practicability is not strong.

Contents of the invention

In view of the above analysis, the embodiment of the present invention aims to provide a driver's driving style recognition method and system to solve the problems of low accuracy and poor practicability in driver's driving style recognition in the prior art.

On the one hand, the present invention provides a kind of driver's driving style identification method, comprising the following steps:

Collect driver operation information and vehicle driving information;

According to the collected driver's operation information and vehicle driving information, the driver's driving style is preliminarily identified, and the preliminary identification result of the driver's driving style is obtained;

Change the vehicle state according to the preliminary identification result of the driver's driving style;

According to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data, the driving style result of the driver is further identified.

The beneficial effects of the above technical solution are as follows: by collecting the driver's operation information and vehicle driving information, the driver's driving style can be preliminarily identified, and the vehicle state can be changed according to the driver's preliminary recognition result of the driver's driving style, and the driver can adapt to the vehicle state according to the operation data. And the corresponding vehicle state data, and further identify the driver's driving style result. This identification method improves the accuracy of the driver's driving style identification and has strong environmental adaptability.

Further, the driver's operation information is obtained through the vehicle CAN bus; the vehicle driving information of the driving vehicle is collected through the device sensor on the vehicle.

Further, based on the collected driver's operation information and vehicle driving information, the driver's driving style is initially identified, specifically including:

extracting operation characteristic parameters from the driver operation information, and extracting driving characteristic parameters from the driving information;

Classify in a preset classifier according to the operating characteristic parameters and driving characteristic parameters, and judge the driver’s driving style corresponding to the operating characteristic parameters and driving characteristic parameters according to the classification results in the classifier;

Among them, the operating characteristic parameters include but not limited to: steering wheel angle, steering wheel angular acceleration, brake pedal position, accelerator pedal position, clutch pedal position and transmission gear in the tactile information, and the driving characteristic parameters include but not limited to: driving vehicle driving information Vehicle speed, position, acceleration, yaw rate, speed of the vehicle relative to surrounding vehicles, distance and acceleration.

The beneficial effect of the above further scheme is that the driver's driving style can be preliminarily identified through the above scheme.

Further, the above-mentioned preset classifier is established, specifically including:

Collect the driver's training operation information and vehicle driving information within the preset time;

extracting training characteristic parameters from the operation information and vehicle driving information, the training characteristic parameters including driving characteristic parameters corresponding to the vehicle driving information and operating characteristic parameters corresponding to the operation information;

Annotate labels for different training driving characteristic parameters to indicate the corresponding driver's driving style;

Based on the preset classification algorithm, the training features under different labels are learned and trained to form a preset classifier.

Further, according to the preliminary recognition result of the driver's driving style, the vehicle state is changed, including:

According to the driver's operation information and vehicle driving information, judge the driver's operating habits;

Match the active detection model corresponding to the driver's operating habits in the pre-established active detection model;

The active detection model changes the vehicle state according to the preliminary identification result of the driver's driving style and the current vehicle state.

The beneficial effect of the above further solution is that the vehicle state can be changed according to the obtained preliminary identification result of the driver's driving style through the above solution.

Further, according to the driver's operation information and vehicle driving information, the driver's operating habits are judged, specifically including:

Collecting driver operation information and vehicle running information within a certain historical time period, and extracting driving characteristic parameters from the information, the driving characteristic parameters including α, yaw rate ω, longitudinal vehicle speed υ, longitudinal acceleration a data set;

Calculate the entropy values H(α), H(ω), H(υ), H(a) corresponding to the above-mentioned driving characteristic parameters;

by As a judgment standard, the driver's operating habit is obtained; where, when γ=1, the driver's operating habit is the longitudinal control type; when γ=2, the driver's operating habit is the lateral control type.

Furthermore, the establishment of the above-mentioned active detection model specifically includes: setting the driver’s operating habits and the corresponding way to change the vehicle state, and collecting the operation information and vehicle state of drivers with different driving styles within a preset time period under different vehicle states. driving information;

Extract the driving characteristic parameter training set from the information, use Gaussian kernel density estimation to establish the Gaussian kernel density estimation model of driving characteristic parameters under different vehicle states, so as to establish an active detection model with the goal of maximizing the Gaussian kernel density difference of the characteristic parameters .

Further, according to the operation data of the driver adapting to the vehicle state and the corresponding vehicle state data, the result of the driver's driving style is further identified, including:

Extract the driving style characteristic parameters from the driver's operation data adapted to the vehicle state and the corresponding vehicle state data;

Classifying the driving style characteristic parameters in a preset classifier;

A driver's driving style corresponding to the driving style characteristic parameter is identified by the classifier.

The beneficial effect of the above further solution is that the driver's driving style can be further recognized through the above solution.

Further, the above-mentioned preset classifiers are established, including:

Collect training information of drivers with different driving styles in different vehicle states within a preset time period, the training information includes driver operation information and vehicle driving information;

Extracting driving characteristic parameters from the training information as a training set;

Label the acquired different training driving feature parameters to indicate the corresponding driver's driving style;

Establish a Gaussian kernel density estimation model of driving characteristic parameters under different driving styles;

According to the Bayesian theorem and the Gaussian kernel density estimation model, a conditional probability model of driving characteristic parameters under different driving styles is established;

According to the conditional probability model, a driving style classifier with the maximum conditional probability as the criterion is established.

In another aspect, the present invention provides a driver's driving style recognition system, the system includes a driving style information collection module, a driving style preliminary recognition module, a vehicle active detection module and a driving style judgment module;

The driving style information collection module is used to collect driver operation information and vehicle driving information;

The driving style preliminary identification module is used to perform preliminary identification of the driver's driving style according to the information collected by the driving style information collection module, and obtain the preliminary identification result of the driver's driving style;

The vehicle active detection module is used to change the state of the vehicle according to the driving style preliminary recognition result identified by the driving style preliminary recognition module;

The driving style judgment module is used to further obtain the result of the driver's driving style according to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data.

The beneficial effect of the above technical solution is that the identification of the driver's driving style is realized through the above system, and the accuracy and environmental adaptability of the identification of the driver's driving style are improved.

In the present invention, the above technical solutions can also be combined with each other to realize more preferred combination solutions. Additional features and advantages of the invention will be set forth in the description which follows, and some of the advantages will be apparent from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by what is particularly pointed out in the written description, claims as well as the appended drawings.

Description of drawings

The drawings are for the purpose of illustrating specific embodiments only and are not to be considered as limitations of the invention, and like reference numerals refer to like parts throughout the drawings.

Fig. 1 is a schematic flow chart of the method described in Embodiment 1 of the present invention;

Fig. 2 is a schematic diagram of the system described in Embodiment 2 of the present invention.

Detailed ways

Preferred embodiments of the present invention will be described in detail below in conjunction with the accompanying drawings, wherein the accompanying drawings constitute a part of the application and together with the embodiments of the present invention are used to explain the principle of the present invention and are not intended to limit the scope of the present invention.

Example 1

An embodiment of the present invention provides a driver's driving style recognition method, including the following steps:

Step S101, collecting driver operation information and vehicle driving information;

Specifically, the driver's operation information is obtained through the vehicle CAN bus; the vehicle driving information of the driving vehicle is collected through the device sensors on the vehicle, and the device sensors may include at least one of the following sensors or a combination of several sensors: biaxial acceleration sensor , speed sensor, yaw rate sensor, etc.;

Step S102. Preliminarily identify the driver's driving style according to the collected information, and obtain a preliminary identification result of the driver's driving style;

Specifically, the operation characteristic parameters are extracted from the driver operation information, the driving characteristic parameters are extracted from the driving information, and the classification is performed in a preset classifier according to the operation characteristic parameters and the driving characteristic parameters, and according to the According to the classification results in the classifier described above, the driving style of the driver corresponding to the driving characteristic parameters can be identified; the driving style includes but is not limited to: aggressive type Steady

Specifically, establishing the preset classifier includes: collecting training operation information and vehicle driving information of the driver within a preset time, extracting training characteristic parameters from the operation information and vehicle driving information, and the training characteristic parameters are the driving characteristic parameters corresponding to the vehicle driving information and the operating characteristic parameters corresponding to the operating information; label the different training driving characteristic parameters to indicate the driving style of the corresponding driver; based on the preset classification algorithm, the The training features under different labels are studied and trained to form a preset classifier; wherein, the training feature parameters include but are not limited to: vehicle speed, lateral acceleration, longitudinal acceleration, yaw rate, steering wheel angle and accelerator pedal position;

Step S103, according to the obtained preliminary identification result of the driver's driving style, according to the set active detection model, appropriately change the state of the vehicle;

Specifically, it includes, according to the driver's operation information and vehicle driving information, judging the driver's operating habits; matching the active detection model corresponding to the driver's operating habits in the pre-established active detection model; the active detection model, according to the obtained The preliminary results of the driver's driving style change the vehicle state appropriately.

The changing of the vehicle state specifically includes: changing the vehicle control parameters related to the driver's operation, the vehicle control parameters include but not limited to: steering torque, steering coefficient, accelerator pedal opening and engine throttle opening function relationship.

The said judging the driver's operating habits based on the operating information and vehicle driving information specifically includes collecting the driving information and the driver's operating information during the vehicle running within a specific period of time, and extracting the driving characteristic parameters from the information. The driving characteristic parameters include data sets of steering wheel angular velocity α, yaw angular velocity ω, longitudinal vehicle speed υ, and longitudinal acceleration a; according to the entropy value theory, entropy values H(α), H(ω), and H(υ), H(a), so that As a judging standard, the driver's operating habits are obtained; the operating habits include: longitudinal control type (γ=1), lateral control type (γ=2).

The process of establishing the active detection model in advance specifically includes: setting the driver's operating habit as longitudinal control type, and the corresponding way to change the vehicle state is to change the accelerator pedal torque output mode, and the accelerator pedal torque output mode includes: sports type (β=1), economical type (β=2), hybrid type (β=3); collect the vehicle driving information and operating information of drivers with different driving styles in different accelerator pedal torque output modes within a preset time, and Extract driving feature parameters from the information, as a training set, the driving feature parameters include steering wheel angular velocity Yaw rate speed lateral acceleration Using Gaussian kernel density estimation to establish a Gaussian kernel density estimation model for each driving characteristic parameter under different accelerator pedal torque output modes Thus, it is established to maximize the characteristic parameter Gaussian kernel density difference ( x = α, ω, υ, a) is the longitudinal control type active detection model of the target.

The process of establishing the horizontal control type active detection model is similar to the above-mentioned establishment of the vertical control type active detection model.

Step S104, according to the operation data of the driver's adaptation to the vehicle state and the corresponding vehicle state data, further identify the driver's driving style result;

Specifically, the relevant driving style characteristic parameters are extracted from the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data, and the driving style characteristic parameters are classified in a preset classifier. a classifier identifying a driver's driving style corresponding to said relevant driving characteristic;

The above-mentioned preset classifier establishment process specifically includes: collecting vehicle driving information and driver operation information of drivers with different driving styles in different accelerator pedal torque output modes within a preset time, and extracting driving characteristic parameters from the training information , as a training set, the driving characteristic parameters include steering wheel angular velocity Yaw rate speed lateral acceleration Label the acquired different training driving feature parameters To mark the corresponding driver's driving style; use Gaussian kernel density estimation to establish a Gaussian kernel density estimation model for each characteristic parameter under different driving styles According to Bayesian theorem and the Gaussian kernel density estimation model, the conditional probability model of each characteristic parameter under different driving styles is established Thereby, according to the conditional probability model, set up the judging criterion with the maximum conditional probability ( x = α, ω, υ, a) driving style classifier.

The process of identifying the driver's driving style corresponding to the relevant driving characteristics according to the classifier specifically includes: collecting the operation data and vehicle data of the driver adapting to the new accelerator pedal torque output mode, and extracting the driving characteristic parameter as the steering wheel angular velocity α _β , yaw rate ω _β , vehicle speed υ _β , and lateral acceleration a _β are input into the classifier to obtain the driver's driving style

An embodiment of the present invention provides a driver's driving style recognition method, which can further identify the driver's driving style by initially judging the driver's driving style, and then using the active detection model based on the driver's response to the vehicle's active detection action The driver's driving style; the method improves the accuracy and environmental adaptability of the driver's driving style identification, making the identified driver's driving style more in line with the actual situation.

Example 2

An embodiment of the present invention provides a driver's driving style identification system, the system includes a driving style information collection module, a driving style preliminary identification module, a vehicle active detection module and a driving style determination module;

The driving style information collection module is used to collect driver operation information and vehicle driving information;

Specifically, the driving style information collection module obtains the driver's operation information through the vehicle CAN bus; collects the vehicle driving information of the driving vehicle through the device sensors on the vehicle, and the device sensors may include at least one or more of the following sensors Combination: dual-axis acceleration sensor, speed sensor, yaw rate sensor, etc.;

The driving style preliminary identification module is used to perform preliminary identification of the driver's driving style according to the information collected by the driving style information collection module, and obtain the preliminary identification result of the driver's driving style;

Specifically, the driving style preliminary recognition module extracts relevant operating characteristic parameters from the relevant driver operating information, extracts relevant driving characteristic parameters from the driving information, and according to the operating characteristic parameters and driving characteristics The parameters are classified in a preset classifier, and the driver’s driving style corresponding to the driving characteristic parameters is judged according to the classification result in the classifier; the driving style includes but is not limited to: aggressive type Steady

Specifically, establishing the preset classifier includes: collecting the driver's training operation information and vehicle running information within a preset time, extracting training feature parameters from the vehicle running information and operating information, and the training feature The parameters are the driving characteristic parameters corresponding to the vehicle driving information and the operating characteristic parameters corresponding to the operating information. The training characteristic parameters include but are not limited to: vehicle speed, lateral acceleration, longitudinal acceleration, yaw rate, steering wheel angle and throttle Pedal position; mark labels for different training driving feature parameters to indicate the corresponding driver's driving style; learn and train the training features under different labels based on the preset classification algorithm to form a preset classifier.

The vehicle active detection module is used to appropriately change the state of the vehicle according to the set active detection model according to the driving style preliminary identification result identified by the driving style preliminary identification module;

Specifically, the vehicle active detection module matches the active detection model corresponding to the driver's operating habits in the pre-established active detection model according to the obtained preliminary identification results of the driving style; operating habits; the active detection model appropriately changes the state of the vehicle according to the obtained preliminary identification result of the driving style.

Said judging the driver's operating habits according to the operating information and vehicle driving information specifically includes collecting the driving information and driver's operating information during the driving of the car within a specific period of time, and extracting the driving characteristic parameters from the information as the steering wheel The data set of corner angular velocity α, yaw angular velocity ω, vehicle speed υ, and lateral acceleration a; according to the entropy value theory, the entropy values H(α), H(ω), H(υ), H(a ), so that with As a judging standard, the driver's operating habits are obtained; the operating habits include: longitudinal control type (γ=1), lateral control type (γ=2).

The process of establishing the active detection model in advance specifically includes: setting the driver's operating habit as longitudinal control type, and the corresponding way to change the vehicle state is to change the accelerator pedal torque output mode. The accelerator pedal torque output mode includes: sports type (β=1), economical type (β=2), hybrid type (β=3); collect vehicle driving information and operating information of drivers with different driving styles in different accelerator pedal torque output modes within a preset time, and from The driving feature parameter extracted from the training information is the steering wheel angular velocity Yaw rate speed lateral acceleration training set; using Gaussian kernel density estimation to establish a Gaussian kernel density estimation model for each characteristic parameter under different accelerator pedal torque output modes Thus, it is established to maximize the characteristic parameter Gaussian kernel density difference ( x = α, ω, υ, a) is the longitudinal control type active detection model of the target.

The driving style determination module is used to further determine the driving style of the driver according to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data;

Specifically, the driving style determination module extracts relevant driving style characteristic parameters from the driver's operation data of adapting to the new state of the vehicle and the corresponding vehicle driving data, and classifies the driving style characteristic parameters in a preset classifier , using the classifier to identify the driver's driving style corresponding to the relevant driving characteristics;

The above-mentioned preset classifier establishment process specifically includes: collecting training vehicle driving information and driver operation information during vehicle driving under different accelerator pedal torque output modes of drivers with different driving styles within a preset time period, and collecting information from the training The driving characteristic parameters are extracted from the information as the steering wheel angular velocity Yaw rate speed lateral acceleration The training set; the label of the different training driving feature parameters obtained by labeling To mark the corresponding driver's driving style; use Gaussian kernel density estimation to establish a Gaussian kernel density estimation model for each characteristic parameter under different driving styles According to Bayesian theorem and the Gaussian kernel density estimation model, the conditional probability model of each characteristic parameter under different driving styles is established Thereby, according to the conditional probability model, set up the judging criterion with the maximum conditional probability ( x = α, ω, υ, a) driving style classifier.

The process of identifying the driver's driving style corresponding to the relevant driving characteristics according to the classifier specifically includes: collecting the operation data and vehicle data of the driver adapting to the new accelerator pedal torque output mode, and extracting the driving characteristic parameter as the steering wheel angular velocity α _β , yaw rate ω _β , vehicle speed υ _β , and lateral acceleration a _β are input into the classifier to obtain the driver's driving style

Those skilled in the art can understand that all or part of the processes of the methods in the above embodiments can be implemented by instructing related hardware through computer programs, and the programs can be stored in a computer-readable storage medium. Wherein, the computer-readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory, and the like.

The above is only a preferred embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present invention can easily think of changes or Replacement should be covered within the protection scope of the present invention.

Claims (10)

1. A driver's driving style recognition method, is characterized in that, comprises the following steps: Collect driver operation information and vehicle driving information; According to the collected driver's operation information and vehicle driving information, the driver's driving style is preliminarily identified, and the preliminary identification result of the driver's driving style is obtained; Change the vehicle state according to the preliminary identification result of the driver's driving style; According to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data, the driving style result of the driver is further identified. 2. The method according to claim 1, characterized in that, the driver's operation information is acquired through the vehicle CAN bus; the vehicle driving information of the driving vehicle is collected through the device sensor on the vehicle. 3. The method according to claim 1, wherein the driver's driving style is initially identified according to the collected driver's operation information and vehicle driving information, specifically comprising: extracting operation characteristic parameters from the driver operation information, and extracting driving characteristic parameters from the driving information; Classify in a preset classifier according to the operating characteristic parameters and driving characteristic parameters, and judge the driver’s driving style corresponding to the operating characteristic parameters and driving characteristic parameters according to the classification results in the classifier; Wherein, the operating characteristic parameters include but not limited to: steering wheel angle, steering wheel angular acceleration, brake pedal position, accelerator pedal position, clutch pedal position and transmission gear in the tactile information, and the driving characteristic parameters include but not limited to: Vehicle speed, position, acceleration, yaw rate, speed of the vehicle relative to surrounding vehicles, distance and acceleration in driving vehicle driving information. 4. according to the described method of claim 3, it is characterized in that, setting up described preset classifier specifically comprises: Collect the driver's training operation information and vehicle driving information within the preset time; extracting training characteristic parameters from the operation information and vehicle driving information, the training characteristic parameters including driving characteristic parameters corresponding to the vehicle driving information and operating characteristic parameters corresponding to the operation information; Annotate labels for different training feature parameters to indicate the corresponding driver's driving style; Based on the preset classification algorithm, the training features under different labels are learned and trained to form a preset classifier. 5. The method according to claim 1, characterized in that changing the vehicle state according to the obtained preliminary identification result of the driver's driving style includes: According to the driver's operation information and vehicle driving information, judge the driver's operating habits; Match the active detection model corresponding to the driver's operating habits in the pre-established active detection model; The active detection model changes the state of the vehicle according to the preliminary recognition result of the driver's driving style and the current state of the vehicle. 6. The method according to claim 5, characterized in that judging the driver's operating habits according to the driver's operating information and vehicle driving information, specifically comprising: Collect driver operation information and vehicle running information within a certain historical period, and extract driving characteristic parameters from the information. The driving characteristic parameters include steering wheel angular velocity α, yaw angular velocity ω, longitudinal vehicle speed υ, and longitudinal acceleration a data set; Calculate the entropy values H(α), H(ω), H(υ), H(a) corresponding to the above-mentioned driving characteristic parameters; by As a judgment standard, the driver's operating habit is obtained; where, when γ=1, the driver's operating habit is the longitudinal control type; when γ=2, the driver's operating habit is the lateral control type. 7. The method according to claim 6, wherein establishing the active detection model specifically comprises: Set the driver's operating habits and the corresponding way to change the vehicle state, collect the operation information and vehicle driving information of drivers with different driving styles in different vehicle states within a preset time; Extract the driving characteristic parameter training set from the information, use Gaussian kernel density estimation to establish the Gaussian kernel density estimation model of driving characteristic parameters under different vehicle states, so as to establish an active detection model with the goal of maximizing the Gaussian kernel density difference of the characteristic parameters . 8. The method according to claim 1, according to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data, further identifying the driving style result of the driver, including: Extract driving style characteristic parameters from the driver's operation data of adapting to the new state of the vehicle and the corresponding vehicle driving data; Classifying the driving style characteristic parameters in a preset classifier; A driver's driving style corresponding to the driving style characteristic parameter is identified by the classifier. 9. The method according to claim 8, wherein establishing the preset classifier comprises: Collect training information of drivers with different driving styles in different vehicle states within a preset time period, the training information includes operation information and vehicle driving information; Extracting driving characteristic parameters from the training information as a training set; Label the acquired different training driving feature parameters to indicate the corresponding driver's driving style; Establish a Gaussian kernel density estimation model of driving characteristic parameters under different driving styles; According to the Bayesian theorem and the Gaussian kernel density estimation model, a conditional probability model of driving characteristic parameters under different driving styles is established; According to the conditional probability model, a driving style classifier with the maximum conditional probability as the criterion is established. 10. A driver's driving style recognition system, characterized in that the system includes a driving style information collection module, a driving style preliminary recognition module, a vehicle active detection module and a driving style judgment module; The driving style information collection module is used to collect driver operation information and vehicle driving information; The driving style preliminary identification module is used to perform preliminary identification of the driver's driving style according to the information collected by the driving style information collection module, and obtain the preliminary identification result of the driver's driving style; The vehicle active detection module is used to change the state of the vehicle according to the driving style preliminary recognition result identified by the driving style preliminary recognition module; The driving style judgment module is used to further obtain the result of the driver's driving style according to the operation data of the driver adapting to the new state of the vehicle and the corresponding vehicle driving data.