CN116894225A

CN116894225A - Driving behavior abnormality analysis method, device, equipment and medium thereof

Info

Publication number: CN116894225A
Application number: CN202311160279.5A
Authority: CN
Inventors: 钟薇; 杜孝平; 乌尼日其其格; 吕东昕; 陈磊
Original assignee: Guoqi Beijing Intelligent Network Association Automotive Research Institute Co ltd
Current assignee: Guoqi Beijing Intelligent Network Association Automotive Research Institute Co ltd
Priority date: 2023-09-08
Filing date: 2023-09-08
Publication date: 2023-10-17
Anticipated expiration: 2043-09-08
Also published as: CN116894225B

Abstract

The application discloses a driving behavior abnormality analysis method, a device, equipment and a medium thereof, comprising the following steps: acquiring a vehicle running state data set; identifying abnormal driving behavior data in the vehicle running state data set, and matching corresponding abnormal labels for the abnormal driving behavior data; taking the abnormal driving behavior data as input data, taking an abnormal label corresponding to the abnormal driving behavior data as output data, and training the target network model to obtain a trained target network model; the method comprises the steps of inputting real-time running state data of a vehicle to a trained target network model, and analyzing driving behaviors of the vehicle to obtain an analysis result of the vehicle; and screening the affected vehicles based on the current position of the vehicle and the abnormal label and pushing the early warning information to the affected vehicles under the condition that the analysis result is the abnormal label. According to the embodiment of the application, the driving behavior of the vehicle can be effectively analyzed and predicted, and the abnormal driving behavior can be effectively early-warned.

Description

Driving behavior abnormality analysis method, device, equipment and medium thereof

Technical Field

The application belongs to the technical field of vehicle monitoring, and particularly relates to a driving behavior abnormality analysis method, a device, equipment and a medium thereof.

Background

With the increase of the quantity of the automobile, the road traffic environment is increasingly complicated, and for the intelligent network-connected automobile, the driving behavior of the driver, especially the abnormal driving behavior, is difficult to understand and predict, so that the intelligent network-connected automobile needs to take over by a safety officer in actual driving. On the one hand, the abnormal behavior analysis of the driver can remind the driver of paying attention to the driving behavior of the driver, and on the other hand, the active safety early warning and the corresponding capability of the intelligent network-connected automobile can be improved through the analysis and the prediction of the typical abnormal behavior.

Therefore, a scheme capable of effectively analyzing and predicting the driving behavior of the vehicle is needed in the related art, so that the abnormal driving behavior of the internet-connected vehicle can be effectively early warned, and the driving safety of the intelligent internet-connected vehicle is improved.

Disclosure of Invention

The embodiment of the application provides a driving behavior abnormality analysis method, a device, equipment and a medium thereof, which can effectively analyze and predict the driving behavior of a vehicle and effectively early warn the abnormal driving behavior.

In a first aspect, an embodiment of the present application provides a driving behavior abnormality analysis method, including: acquiring a vehicle running state data set; identifying abnormal driving behavior data in the vehicle running state data set, and matching corresponding abnormal labels for the abnormal driving behavior data, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors; taking the abnormal driving behavior data as input data, taking an abnormal label corresponding to the abnormal driving behavior data as output data, and training the target network model to obtain a trained target network model; the method comprises the steps of inputting real-time running state data of a vehicle to a trained target network model, and analyzing driving behaviors of the vehicle to obtain an analysis result of the vehicle; and screening the affected vehicles based on the current position of the vehicle and the abnormal label and pushing the early warning information to the affected vehicles under the condition that the analysis result is the abnormal label.

In some implementations of the first aspect, the target network model is a residual network model based on a self-attention mechanism, the target network model including 16 residual modules and 12 attention modules.

In some implementations of the first aspect, before inputting the real-time operating state data of the vehicle to the trained target network model, the method further includes: under the condition that the real-time running state data of the vehicle is obtained, carrying out normalization processing and denoising processing on the real-time running state data.

In some implementations of the first aspect, the vehicle operation state data includes vehicle operation state data reported at a plurality of reporting times when N vehicles pass through the same road segment, the vehicle operation state data includes a speed value and an acceleration, and identifying abnormal driving behavior data in the vehicle operation state data set includes: dividing the multiple reporting time scores of each vehicle into different speed intervals based on multiple speed values reported by each vehicle in the N vehicles at multiple reporting times; based on the acceleration corresponding to all reporting moments in each speed interval, determining an acceleration mean value and an acceleration standard deviation corresponding to each speed interval; determining abnormal driving judgment conditions matched with each speed interval based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval; screening target reporting moments meeting abnormal driving judgment conditions matched with the speed interval from all reporting moments falling into the speed interval to obtain abnormal driving time points, wherein the abnormal driving time points are reporting moments corresponding to target vehicles in N vehicles; selecting a first reporting time when acceleration starts to change and a second reporting time when acceleration ends to change from all reporting times in preset time before and after an abnormal driving time point of a target vehicle; and intercepting the vehicle running state data of the target vehicle from the first reporting moment to the second reporting moment to obtain abnormal driving behavior data.

In some implementations of the first aspect, the acceleration includes a longitudinal acceleration and/or a lateral acceleration, and the abnormal driving determination condition includes any one of: the first sub-condition is that the longitudinal acceleration is greater than a first longitudinal acceleration threshold; the second sub-condition is that the longitudinal acceleration is less than a second longitudinal acceleration threshold; the third sub-condition is that the lateral acceleration is greater than the first lateral acceleration threshold; the fourth sub-condition is that the lateral acceleration is less than a second lateral acceleration threshold; the first longitudinal acceleration threshold and the second longitudinal acceleration threshold are determined based on a longitudinal acceleration mean value and a longitudinal acceleration standard deviation corresponding to each speed interval, and the first transverse acceleration threshold and the second transverse acceleration threshold are determined based on a transverse acceleration mean value and a transverse acceleration standard deviation corresponding to each speed interval.

In some implementations of the first aspect, the acceleration includes a longitudinal acceleration, and determining an abnormal driving judgment condition matched with each speed interval based on an acceleration average value and an acceleration standard deviation corresponding to each speed interval includes: adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first longitudinal acceleration threshold value; the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second longitudinal acceleration threshold value; the abnormal driving judgment condition is determined as that the longitudinal acceleration is greater than the first longitudinal acceleration threshold value or that the longitudinal acceleration is less than the second longitudinal acceleration threshold value.

In some implementations of the first aspect, the acceleration includes a lateral acceleration, and determining the abnormal driving judgment condition matched with each speed interval based on the acceleration average value and the acceleration standard deviation corresponding to each speed interval includes: adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first transverse acceleration threshold; the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second transverse acceleration threshold value; the abnormal driving determination condition is determined as whether the lateral acceleration is greater than the first lateral acceleration threshold value or the lateral acceleration is less than the second lateral acceleration threshold value.

In some implementations of the first aspect, matching corresponding abnormal labels for abnormal driving behavior data includes at least one of: in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the first sub-condition, matching an abnormal label of the abnormal acceleration behavior; in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the second sub-condition, matching an abnormal label of the abnormal deceleration behavior; when the target vehicle is at the intersection of the road section at the abnormal driving time point, matching an abnormal tag of the abnormal turning behavior for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition; when the target vehicle is located in the road section at the abnormal driving time point, the abnormal tag of the abnormal lane change behavior is matched for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition.

In a second aspect, an embodiment of the present application provides a driving behavior abnormality analysis apparatus including: the acquisition module is used for acquiring a vehicle running state data set; the recognition matching module is used for recognizing abnormal driving behavior data in the vehicle running state data set and matching corresponding abnormal labels for the abnormal driving behavior data, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane change behaviors and abnormal turning behaviors; the training module is used for training the target network model by taking the abnormal driving behavior data as input data and the abnormal labels corresponding to the abnormal driving behavior data as output data to obtain a trained target network model; the analysis module is used for analyzing the driving behavior of the vehicle by inputting the real-time running state data of the vehicle into the trained target network model to obtain an analysis result of the vehicle; and the early warning module is used for screening the affected vehicles based on the current positions of the vehicles and the abnormal labels and pushing early warning information to the affected vehicles under the condition that the analysis results are the abnormal labels.

In some implementations of the second aspect, the target network model is a residual network model based on a self-attention mechanism, the target network model including 16 residual modules and 12 attention modules.

In some implementations of the second aspect, the apparatus further includes: the preprocessing module is used for carrying out normalization processing and denoising processing on the real-time running state data under the condition that the real-time running state data of the vehicle are obtained before the real-time running state data of the vehicle are input to the trained target network model.

In some implementations of the second aspect, the vehicle operation state data includes vehicle operation state data reported at a plurality of reporting times when the N vehicles pass through the same road segment, the vehicle operation state data includes a speed value and an acceleration, and the identification matching module includes: the dividing sub-module is used for dividing the multiple reporting time scores of each vehicle into different speed intervals based on multiple speed values reported by each vehicle in the N vehicles at multiple reporting time instants; the determining submodule is used for determining an acceleration mean value and an acceleration standard deviation corresponding to each speed interval based on the accelerations corresponding to all reporting moments in each speed interval; the determining submodule is further used for determining abnormal driving judgment conditions matched with each speed interval based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval; the screening sub-module is used for screening target reporting moments meeting abnormal driving judgment conditions matched with the speed interval from all the reporting moments falling into the speed interval to obtain abnormal driving time points, wherein the abnormal driving time points are reporting moments corresponding to target vehicles in N vehicles; the selecting sub-module is used for selecting a first reporting time when acceleration starts to change and a second reporting time when acceleration ends to change from all reporting times in preset time periods before and after an abnormal driving time point of the target vehicle; and the intercepting sub-module is used for intercepting the vehicle running state data of the target vehicle from the first reporting moment to the second reporting moment to obtain abnormal driving behavior data.

In some implementations of the second aspect, the acceleration includes a longitudinal acceleration and/or a lateral acceleration, and the abnormal driving determination condition includes any one of: the first sub-condition is that the longitudinal acceleration is greater than a first longitudinal acceleration threshold; the second sub-condition is that the longitudinal acceleration is less than a second longitudinal acceleration threshold; the third sub-condition is that the lateral acceleration is greater than the first lateral acceleration threshold; the fourth sub-condition is that the lateral acceleration is less than a second lateral acceleration threshold; the first longitudinal acceleration threshold and the second longitudinal acceleration threshold are determined based on a longitudinal acceleration mean value and a longitudinal acceleration standard deviation corresponding to each speed interval, and the first transverse acceleration threshold and the second transverse acceleration threshold are determined based on a transverse acceleration mean value and a transverse acceleration standard deviation corresponding to each speed interval.

In some implementations of the second aspect, the acceleration includes a longitudinal acceleration, and the determining submodule is specifically configured to: adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first longitudinal acceleration threshold value; the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second longitudinal acceleration threshold value; the abnormal driving judgment condition is determined as that the longitudinal acceleration is greater than the first longitudinal acceleration threshold value or that the longitudinal acceleration is less than the second longitudinal acceleration threshold value.

In some implementations of the second aspect, the acceleration includes a lateral acceleration, and the determining submodule is specifically configured to: adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first transverse acceleration threshold; the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second transverse acceleration threshold value; the abnormal driving determination condition is determined as whether the lateral acceleration is greater than the first lateral acceleration threshold value or the lateral acceleration is less than the second lateral acceleration threshold value.

In some implementations of the second aspect, the identification matching module is specifically configured to at least one of: in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the first sub-condition, matching an abnormal label of the abnormal acceleration behavior; in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the second sub-condition, matching an abnormal label of the abnormal deceleration behavior; when the target vehicle is at the intersection of the road section at the abnormal driving time point, matching an abnormal tag of the abnormal turning behavior for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition; when the target vehicle is located in the road section at the abnormal driving time point, the abnormal tag of the abnormal lane change behavior is matched for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition.

In a third aspect, an embodiment of the present application provides an electronic device, including: a processor and a memory storing computer program instructions; the processor when executing the computer program instructions implements the steps of the driving behaviour abnormal analysis method as shown in any one of the embodiments of the first aspect.

In a fourth aspect, embodiments of the present application provide a computer readable storage medium having stored thereon computer program instructions which, when executed by a processor, implement the steps of the driving behaviour abnormal analysis method as shown in any one of the embodiments of the first aspect.

In a fifth aspect, embodiments of the present application provide a computer program product, instructions in which, when executed by a processor of an electronic device, cause the electronic device to perform the steps of the driving behaviour abnormal analysis method as shown in any one of the embodiments of the first aspect.

According to the driving behavior abnormality analysis method, device, equipment and medium, after the vehicle running state data set is acquired, abnormal driving behavior data in the vehicle running state data set can be identified, corresponding abnormal labels are matched for the abnormal driving behavior data, and the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors. Based on the method, the abnormal driving behavior data can be used as input data, the abnormal labels corresponding to the abnormal driving behavior data are used as output data, the target network model is trained, and the trained target network model is obtained, so that the target network model can have the capability of carrying out abnormal analysis and recognition on the driving behavior of the vehicle. On the basis, the real-time running state data of the vehicle is input into the trained target network model, so that the target network model can specifically analyze the driving behavior of the vehicle according to the real-time running state data, and analyze whether the driving behavior is abnormal driving behavior or not, thereby obtaining the analysis result of the vehicle, and effectively analyzing and predicting the driving behavior of the vehicle. Under the condition that the analysis result is an abnormal tag, the abnormal tag can represent the specific type of abnormal driving behavior of the vehicle, so that the affected vehicle can be screened out from the vicinity of the vehicle by combining the abnormal tag and the current position of the vehicle, and the vehicle with safety risk and abnormal driving behavior can be effectively pre-warned by pushing pre-warning information to the affected vehicle, so that the driving safety of the intelligent network-connected vehicle is improved.

Drawings

In order to more clearly illustrate the technical solution of the embodiments of the present application, the drawings that are needed to be used in the embodiments of the present application will be briefly described, and it is possible for a person skilled in the art to obtain other drawings according to these drawings without inventive effort.

FIG. 1 is a flow chart of a driving behavior anomaly analysis method according to an embodiment of the present application;

fig. 2 is a flow chart of a driving behavior abnormality analysis method according to another embodiment of the present application;

FIG. 3 is a flowchart illustrating a driving behavior abnormality analysis method according to still another embodiment of the present application;

FIG. 4 is an exemplary schematic diagram of a target network model structure provided by an embodiment of the present application;

FIG. 5 is an exemplary schematic diagram of a residual module structure provided by an embodiment of the present application;

fig. 6 is a schematic structural diagram of a driving behavior abnormality analysis device according to an embodiment of the present application;

fig. 7 is a schematic hardware structure of an electronic device according to an embodiment of the present application.

Detailed Description

Features and exemplary embodiments of various aspects of the present application will be described in detail below, and in order to make the objects, technical solutions and advantages of the present application more apparent, the present application will be described in further detail below with reference to the accompanying drawings and the detailed embodiments. It should be understood that the particular embodiments described herein are meant to be illustrative of the application only and not limiting. It will be apparent to one skilled in the art that the present application may be practiced without some of these specific details. The following description of the embodiments is merely intended to provide a better understanding of the application by showing examples of the application.

It is noted that relational terms such as first and second, and the like are used solely to distinguish one entity or action from another entity or action without necessarily requiring or implying any actual such relationship or order between such entities or actions. Moreover, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements does not include only those elements but may include other elements not expressly listed or inherent to such process, method, article, or apparatus. Without further limitation, an element defined by the phrase "comprising … …" does not exclude the presence of other like elements in a process, method, article or apparatus that comprises the element.

Aiming at the problems in the related art, the embodiment of the application provides a driving behavior abnormality analysis method, which can identify abnormal driving behavior data in a vehicle running state data set after the vehicle running state data set is acquired, and match corresponding abnormality labels for the abnormal driving behavior data, wherein the abnormality labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors. Based on the method, the abnormal driving behavior data can be used as input data, the abnormal labels corresponding to the abnormal driving behavior data are used as output data, the target network model is trained, and the trained target network model is obtained, so that the target network model can have the capability of carrying out abnormal analysis and recognition on the driving behavior of the vehicle. On the basis, the real-time running state data of the vehicle is input into the trained target network model, so that the target network model can specifically analyze the driving behavior of the vehicle according to the real-time running state data, and analyze whether the driving behavior is abnormal driving behavior or not, thereby obtaining the analysis result of the vehicle, and effectively analyzing and predicting the driving behavior of the vehicle. Under the condition that the analysis result is an abnormal tag, the abnormal tag can represent the specific type of abnormal driving behavior of the vehicle, so that the affected vehicle can be screened out from the vicinity of the vehicle by combining the abnormal tag and the current position of the vehicle, and the vehicle with safety risk and abnormal driving behavior can be effectively pre-warned by pushing pre-warning information to the affected vehicle, so that the driving safety of the intelligent network-connected vehicle is improved.

The driving behavior abnormality analysis method provided by the embodiment of the application is described in detail below through specific embodiments and application scenes thereof with reference to the accompanying drawings.

Fig. 1 is a schematic flow chart of a driving behavior abnormality analysis method according to an embodiment of the present application, where an execution body of the driving behavior abnormality analysis method may be an electronic device. The execution body is not limited to the present application.

The electronic device may be a device with a communication function, such as a mobile phone, a tablet computer, an integrated machine, or a device simulated by a virtual machine or a simulator, and may, of course, also include a device with a storage and calculation function, such as a cloud server or a server cluster.

The cloud server can be in communication connection with the vehicle and receives real-time running state data reported by the vehicle, so that the cloud server can analyze and predict driving behaviors of all vehicles by analyzing the real-time running state data of all vehicles and identify vehicles with abnormal driving behaviors, and therefore safety risks of the vehicles are informed to other affected vehicles, and early warning is effectively carried out.

As shown in fig. 1, the driving behavior abnormality analysis method provided by the embodiment of the present application may include steps 110 to 150.

Step 110, acquiring a vehicle running state data set;

step 120, identifying abnormal driving behavior data in the vehicle running state data set, and matching corresponding abnormal labels for the abnormal driving behavior data, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors;

step 130, training the target network model by taking the abnormal driving behavior data as input data and the abnormal labels corresponding to the abnormal driving behavior data as output data to obtain a trained target network model;

step 140, analyzing the driving behavior of the vehicle by inputting real-time running state data of the vehicle into the trained target network model to obtain an analysis result of the vehicle;

and step 150, screening the affected vehicles based on the current positions of the vehicles and the abnormal labels and pushing the early warning information to the affected vehicles under the condition that the analysis result is the abnormal labels.

According to the driving behavior anomaly analysis method, after the vehicle running state data set is obtained, the abnormal driving behavior data in the vehicle running state data set can be identified, and the abnormal driving behavior data is matched with the corresponding abnormal labels, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors. Based on the method, the abnormal driving behavior data can be used as input data, the abnormal labels corresponding to the abnormal driving behavior data are used as output data, the target network model is trained, and the trained target network model is obtained, so that the target network model can have the capability of carrying out abnormal analysis and recognition on the driving behavior of the vehicle. On the basis, the real-time running state data of the vehicle is input into the trained target network model, so that the target network model can specifically analyze the driving behavior of the vehicle according to the real-time running state data, and analyze whether the driving behavior is abnormal driving behavior or not, thereby obtaining the analysis result of the vehicle, and effectively analyzing and predicting the driving behavior of the vehicle. Under the condition that the analysis result is an abnormal tag, the abnormal tag can represent the specific type of abnormal driving behavior of the vehicle, so that the affected vehicle can be screened out from the vicinity of the vehicle by combining the abnormal tag and the current position of the vehicle, and the vehicle with safety risk and abnormal driving behavior can be effectively pre-warned by pushing pre-warning information to the affected vehicle, so that the driving safety of the intelligent network-connected vehicle is improved.

The specific implementation of the above steps will be described in detail below with reference to specific embodiments.

Referring to step 110, a vehicle operating state data set is acquired.

Specifically, the vehicle running state data set may include vehicle running state data reported at a plurality of reporting moments when N vehicles pass through the same road section, where N is a positive integer, the same road section may be a preset experimental road section, and N vehicles may include all vehicles passing through the experimental road section within a preset time period, where the preset time period may be set according to specific requirements, for example, set to 8:00-9:00 am.

The vehicle operating state data may include, but is not limited to: speed value, acceleration, vehicle number, longitude, latitude, heading angle (i.e., the angle between the direction of vehicle motion and the north direction), accelerator opening, brake pedal opening, steering wheel angle, yaw rate.

In some embodiments, the electronic device may be a cloud server, and the N vehicles may report vehicle running state data to the cloud server based on a preset reporting frequency, where the preset reporting frequency may be set according to specific requirements, for example, set to 3 s/time, 5 s/time, or other frequencies. The length of the different vehicles passing through the same road section is different based on the different running speeds of the different vehicles, so that the reporting frequency is different when the different vehicles pass through the same road section, although the reporting frequency is the same.

Illustratively, the vehicle operating state data set may include, but is not limited to: the safe driving model deploys (Safety Pilot Model Deployment, SPMD) datasets, (Next Generation Simulation, NGSIM) datasets.

Step 120 is involved in identifying abnormal driving behavior data in the vehicle running state data set and matching corresponding abnormal tags for the abnormal driving behavior data.

Specifically, the abnormal acceleration behavior is a behavior of acceleration with a larger longitudinal acceleration in a shorter time; the abnormal deceleration behavior is a behavior of decelerating by adopting small longitudinal acceleration in a short time; the abnormal lane change behavior is the lane change behavior by adopting larger transverse acceleration in a shorter time; the abnormal turning behavior is a behavior of driving at an intersection with a large lateral acceleration to a road different from the current direction in a short time.

In some embodiments of the present application, the vehicle running state data may include vehicle running state data reported at a plurality of reporting times when N vehicles pass through the same road section, the vehicle running state data includes a speed value and an acceleration, and in order to accurately identify abnormal driving behavior data from a large amount of vehicle running state data reported by N vehicles, fig. 2 is a flowchart of a driving behavior abnormality analysis method provided in another embodiment of the present application, and step 120 may include steps 210 to 260 shown in fig. 2.

Step 210, dividing the multiple reporting time scores of each vehicle into different speed intervals based on the multiple speed values reported by each vehicle in the N vehicles at the multiple reporting times;

step 220, determining an acceleration mean value and an acceleration standard deviation corresponding to each speed interval based on the accelerations corresponding to all reporting moments in each speed interval;

step 230, determining abnormal driving judgment conditions matched with each speed interval based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval;

step 240, screening target reporting time points meeting abnormal driving judgment conditions matched with the speed interval from all reporting time points falling into the speed interval to obtain abnormal driving time points, wherein the abnormal driving time points are reporting time points corresponding to target vehicles in N vehicles;

step 250, selecting a first reporting time when acceleration starts to change and a second reporting time when acceleration ends to change from all reporting times within a preset time before and after an abnormal driving time point of a target vehicle;

and 260, intercepting the vehicle running state data of the target vehicle from the first reporting time to the second reporting time to obtain abnormal driving behavior data.

In the embodiment of the application, considering that the acceleration tends to be larger when the vehicle runs at a low speed or starts, and the corresponding acceleration tends to be smaller when the vehicle runs normally, if the acceleration is taken as a judgment basis only, for example, the vehicle running state data in the stage of larger acceleration is taken as abnormal driving behavior data, and the acquisition accuracy of the abnormal driving behavior data is affected without combining with the vehicle speed. Based on the method, the abnormal acceleration is identified in a mode of dividing the speed intervals, and the abnormal driving judgment conditions matched with each speed interval are determined based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval, so that the suitability of the speed interval and the abnormal driving judgment conditions is higher. In this way, based on the abnormal driving judgment condition with high suitability to the speed interval, the target reporting time meeting the abnormal driving judgment condition can be screened from all the reporting time falling into the speed interval, so as to obtain the time point when the vehicle has abnormal behavior, namely the abnormal driving time point. Further, in order to acquire effective abnormal driving behavior data, the method takes an abnormal driving time point as a reference, analyzes the change of the transverse/longitudinal acceleration in a period of time before and after the abnormal driving time point, selects the starting point of acceleration starting change as an abnormal driving behavior starting point and the end point of acceleration change as an abnormal driving behavior end point, so that vehicle dynamic data fragments corresponding to the abnormal driving behavior are dynamically cut out, and the fragments are taken as abnormal driving behavior data, wherein the abnormal driving behavior data can effectively reflect the speed and acceleration change of the vehicle during abnormal driving, and provides basis for subsequent model training.

The interval length of the speed interval may be set according to specific requirements, for example, the interval length is set to 5, the speed interval is respectively (1 m/s,5 m/s), (5 m/s,10 m/s), (10 m/s,15 m/s), and the like, and of course, other interval lengths may also be set, which is not particularly limited in the present application.

In some embodiments of the present application, the acceleration includes a longitudinal acceleration and/or a lateral acceleration, and the abnormal driving determination condition includes any one of:

the first sub-condition is that the longitudinal acceleration is greater than a first longitudinal acceleration threshold;

the second sub-condition is that the longitudinal acceleration is less than a second longitudinal acceleration threshold;

the third sub-condition is that the lateral acceleration is greater than the first lateral acceleration threshold;

the fourth sub-condition is that the lateral acceleration is less than a second lateral acceleration threshold;

the first longitudinal acceleration threshold and the second longitudinal acceleration threshold are determined based on a longitudinal acceleration mean value and a longitudinal acceleration standard deviation corresponding to each speed interval, and the first transverse acceleration threshold and the second transverse acceleration threshold are determined based on a transverse acceleration mean value and a transverse acceleration standard deviation corresponding to each speed interval. The abnormal driving judgment condition includes any one of a first sub-condition, a second sub-condition, a third sub-condition and a fourth sub-condition, and the reporting time can be confirmed as a target reporting time and also as an abnormal driving time point as long as the acceleration at the reporting time satisfies any one of the conditions.

In some embodiments of the present application, the acceleration includes a longitudinal acceleration, and in order to determine the abnormal driving determination condition matched with each speed interval, fig. 3 is a flowchart illustrating a driving behavior abnormality analysis method according to still another embodiment of the present application, and step 230 may include steps 310 to 330 shown in fig. 3.

Step 310, adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first longitudinal acceleration threshold;

step 320, the acceleration mean value is subjected to difference operation with the acceleration standard deviation of a preset multiple to obtain a second longitudinal acceleration threshold value;

in step 330, it is determined that the abnormal driving determination condition is that the longitudinal acceleration is greater than the first longitudinal acceleration threshold, or that the longitudinal acceleration is less than the second longitudinal acceleration threshold.

The acceleration average value is an average value of longitudinal acceleration at a plurality of reporting moments falling into a speed interval, and the acceleration standard deviation is a standard deviation of the longitudinal acceleration at the plurality of reporting moments falling into the speed interval; the preset multiple may be set according to specific requirements, for example, 2 times, 3 times, or other multiple, which is not specifically limited in the present application.

For example, the preset multiple is 2, for a plurality of reporting moments falling into the same speed interval, the accelerations corresponding to the reporting moments may form an acceleration set a, and the average value of the set a is calculated And standard deviation of set A +.>Determining a first longitudinal acceleration threshold as +.>Determining the second longitudinal acceleration threshold asThe longitudinal acceleration value is greater than +.>The reporting time is an abnormal driving time point, and the longitudinal acceleration value is smaller than +.>The reporting time is an abnormal driving time point.

In some embodiments of the present application, the acceleration may comprise a lateral acceleration, and step 230 may comprise the steps of:

adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first transverse acceleration threshold;

the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second transverse acceleration threshold value;

the abnormal driving determination condition is determined as whether the lateral acceleration is greater than the first lateral acceleration threshold value or the lateral acceleration is less than the second lateral acceleration threshold value.

The acceleration average value is an average value of lateral acceleration at a plurality of reporting moments falling into a speed interval, and the acceleration standard deviation is a standard deviation of lateral acceleration at the plurality of reporting moments falling into the speed interval.

In some embodiments of the present application, the matching of the corresponding abnormal tag for the abnormal driving behavior data in step 120 may include at least one of:

In order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the first sub-condition, matching an abnormal label of the abnormal acceleration behavior;

in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the second sub-condition, matching an abnormal label of the abnormal deceleration behavior;

when the target vehicle is at the intersection of the road section at the abnormal driving time point, matching an abnormal tag of the abnormal turning behavior for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition;

when the target vehicle is located in the road section at the abnormal driving time point, the abnormal tag of the abnormal lane change behavior is matched for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition.

In the embodiment of the application, the abnormal driving behaviors in N vehicles can be identified by combining the speed values and the accelerations reported at a plurality of reporting moments when the N vehicles pass through the same road section, and the specific abnormal driving type of the abnormal driving behavior data is obtained by analyzing the abnormal driving behavior data corresponding to the abnormal driving time points, so that the specific abnormal driving type of the abnormal driving behavior data is accurately matched with the corresponding abnormal label, and when the target network model is trained based on the abnormal driving behavior data and the abnormal label thereof, the abnormal driving behavior identification capability and the identification accuracy of the target network model are improved.

Step 130 is involved, in which abnormal driving behavior data is used as input data, an abnormal label corresponding to the abnormal driving behavior data is used as output data, and the target network model is trained to obtain a trained target network model.

In some embodiments of the application, the target network model is a residual network model based on a self-attention mechanism, the target network model comprising 16 residual modules and 12 attention modules.

In the embodiment of the application, the residual error network model based on the self-attention mechanism is selected to model the abnormal driving behavior, the residual error network model solves the problems of gradient disappearance, network degradation and the like in the traditional neural network model, the gradient can be better transmitted among layers in the network, and the self-attention mechanism is added on the basis of the residual error network, so that the target network model can better screen out parameters related to the abnormal driving behavior, and the accuracy of model identification is improved.

In some embodiments, as shown in fig. 4, the target network model is obtained by sequentially concatenating an initial convolutional layer (Conv-BN-ReLU), a max pooling layer (Maxpool), 3 residual modules, 2 attention modules, 4 residual modules, 3 attention modules, 6 residual modules, 5 attention modules, 3 residual modules, 2 attention modules, an average pooling layer, a full concatenation layer, and a Softmax layer.

The initial convolution layer can be composed of a convolution layer, batch normalization and an activation function, the convolution kernel of the convolution layer can be 7 multiplied by 7, and the step length is 2; the core size in the maximum pooling layer can be selected to be 2×2, and the step size is 2.

In the embodiment of the application, the feature extraction module adds an attention mechanism to further compress unimportant features, and intermediate feature mapping can be thinned into target information which can more represent input through integration of the attention mechanism. And finally, the prediction speed of the model is improved through a global average pooling layer.

As a specific example, the structure of the residual module may be shown in fig. 5, where the batch normalization layer (Batch normalization, BN) may greatly improve the model training speed and improve the network generalization performance, x is an input variable, the residual module obtains a residual block calculation result through a residual block on the left side (including steps of BN layer, convolutional layer Conv, relu activation function, etc.), obtains an identity mapping layer result through an identity mapping layer on the right side (including one BN and 1×1 convolutional layer Conv), and obtains a final output result through a relu activation function after the residual block calculation result is added to the identity mapping layer result.

In the embodiment of the application, the batch normalization layer is added in the residual block of the residual network, so that the time for model training can be greatly shortened; by adding a 1×1 convolution layer and a batch normalization layer in the identity mapping layer, the dimension of the upper layer output data is consistent with the dimension of the residual block output data.

In some embodiments of the application, the method may further comprise, prior to training the target network model: and carrying out normalization processing and denoising processing on the abnormal driving behavior data.

In step 140, real-time running state data of the vehicle is input into the trained target network model, and driving behaviors of the vehicle are analyzed to obtain an analysis result of the vehicle.

Wherein the real-time operating state data of the vehicle may include, but is not limited to: speed value, acceleration, vehicle number, longitude, latitude, heading angle, accelerator opening, brake pedal opening, steering wheel angle, yaw rate.

In some embodiments of the present application, before inputting the real-time operating state data of the vehicle to the trained target network model in step 140, the method may further include the steps of:

under the condition that the real-time running state data of the vehicle is obtained, carrying out normalization processing and denoising processing on the real-time running state data.

Specifically, since abnormal data may occur in the real-time operation state data during the acquisition process, the abnormal data needs to be identified and processed. For example, a value range is set for each field of the real-time running state data, if the acceleration range is set to be-6 m/s2, the data with obvious errors exceeding the range is removed. Secondly, because the value ranges of the fields are different, in order to eliminate the influence of the dimension of each field on the recognition of the abnormal behavior, the data needs to be standardized/normalized, and a Z-Score standardization method can be adopted for the original dataProcessing, the calculation formula of the standardized method can be formula (1):

（1）

wherein, the liquid crystal display device comprises a liquid crystal display device,mean value of raw data>The standard deviation of the original data is +.>The sequence has a mean of 0, a variance of 1, and no dimension.

Finally, when the real-time running state data of the vehicle is collected, the data has errors due to sensor errors or communication factors, and the real-time running state data can be denoised by means of wavelet transformation and the like, so that the data quality is further improved.

And step 150, screening the affected vehicles based on the current positions of the vehicles and the abnormal labels and pushing early warning information to the affected vehicles when the analysis result is the abnormal labels.

Specifically, under the condition that an analysis result is an abnormal label, acquiring real-time running state data currently reported by the vehicle, wherein the real-time running state data comprises longitude and latitude and a course angle, matching coordinates of the vehicle in a preset high-precision map based on the longitude and latitude position and the course angle to obtain a current position of the vehicle, and converting the course angle into an included angle between the vehicle movement direction and the vehicle lane reference line direction according to the curvature of a road where the current position of the vehicle is located.

In some embodiments, the early warning information may include, but is not limited to, the following fields: the vehicle number, the timestamp, the latitude of the abnormal vehicle, the longitude of the abnormal vehicle, the lane number of the abnormal vehicle, the speed of the abnormal vehicle, the recommended lane change direction (enumeration type; 0: unchanged lane; 1: changed lane leftwards; 2: changed lane rightwards), and the recommended running speed.

In some embodiments of the present application, in the case that the analysis result is an abnormal tag, the vehicle is an abnormal vehicle, and the step 150 of screening the affected vehicle based on the current location of the vehicle and the abnormal tag may specifically include the following cases:

case one: and for the abnormal vehicle with the abnormal label being in abnormal deceleration behavior, determining the same lane vehicle behind the abnormal vehicle as the affected vehicle.

In this case, firstly, calculating the TTC value between the affected vehicle and the abnormal vehicle, and if the TTC value is smaller than a TTC1 threshold value, the distance between the affected vehicle and the front vehicle is smaller, and suggesting that the affected vehicle decelerates at the maximum deceleration; if less than TTC2 (TTC 1< TTC 2), it is recommended that the affected vehicle decelerate at a comfortable deceleration; when TTC > TTC2, the cloud server can judge whether the adjacent lanes of the affected vehicle have enough safety space, if so, the cloud server can also send corresponding lane change suggestions to the affected vehicle, and the affected vehicle can drive away from the lane where the abnormal vehicle is located.

And a second case: for an abnormal vehicle whose abnormal tag is abnormal acceleration, determining that a same lane vehicle in front of the abnormal vehicle is an affected vehicle.

In this case, on the one hand, acceleration of the affected vehicle to the highest safe vehicle speed may be recommended, and on the other hand, lane change advice may also be sent to the affected vehicle, driving off the lane in which the abnormal vehicle is located.

And a third case: for an abnormal vehicle whose abnormal tag is an abnormal lane change, determining a vehicle behind a target lane to be lane-changed as an affected vehicle.

In this case, the acceleration calculation method of the affected vehicle is identical to the first case.

In some embodiments of the present application, the highest safe vehicle speed calculation method is as follows:

When no front vehicle exists: the highest safe vehicle speed is the current road speed limit;

when a vehicle is in front, the maximum safe vehicle speed of the vehicle can be determined according to a following model consisting of a formula (2), a formula (3) and a formula (4):

（2）

wherein, the liquid crystal display device comprises a liquid crystal display device,for the speed of the front car>For the reaction time of the driver, +.>For the distance between two vehicles, < >>Is the maximum acceleration.

（3）

Wherein, the liquid crystal display device comprises a liquid crystal display device,speed of the rear vehicle>And the speed limit is realized for the road.

（4）

Wherein, the liquid crystal display device comprises a liquid crystal display device,speed of the rear vehicle>For rear vehicle acceleration->Speed of the front car>For acceleration of the front vehicle->Is the minimum distance that must be maintained when the vehicle is stopped.

When changing the lane, the distances between the affected vehicle and the front and rear vehicles on the lane of the lane change target are larger than the safety distance.

It should be noted that, in the driving behavior abnormality analysis method provided by the embodiment of the present application, the execution body may be an electronic device, or a control module for executing the driving behavior abnormality analysis method in the driving behavior abnormality analysis device. In the embodiment of the application, the driving behavior abnormality analysis device provided by the embodiment of the application is described by taking the driving behavior abnormality analysis method executed by the driving behavior abnormality analysis device as an example. The driving behavior abnormality analysis device will be described in detail below.

Fig. 6 is a schematic structural diagram of a driving behavior abnormality analysis device according to an embodiment of the present application. As shown in fig. 6, the driving behavior abnormality analysis apparatus 600 may include: the system comprises an acquisition module 610, an identification matching module 620, a training module 630, an analysis module 640 and an early warning module 650.

Wherein, the acquiring module 610 is configured to acquire a vehicle running state data set; the recognition matching module 620 is configured to recognize abnormal driving behavior data in the vehicle running state data set, and match corresponding abnormal labels for the abnormal driving behavior data, where the abnormal labels include abnormal acceleration behavior, abnormal deceleration behavior, abnormal lane change behavior, and abnormal turning behavior; the training module is used for training the target network model by taking the abnormal driving behavior data as input data and the abnormal labels corresponding to the abnormal driving behavior data as output data to obtain a trained target network model; the analysis module is used for analyzing the driving behavior of the vehicle by inputting the real-time running state data of the vehicle into the trained target network model to obtain an analysis result of the vehicle; and the early warning module is used for screening the affected vehicles based on the current positions of the vehicles and the abnormal labels and pushing early warning information to the affected vehicles under the condition that the analysis results are the abnormal labels.

In some embodiments of the application, the apparatus further comprises: the preprocessing module is used for carrying out normalization processing and denoising processing on the real-time running state data under the condition that the real-time running state data of the vehicle are obtained before the real-time running state data of the vehicle are input to the trained target network model.

In some embodiments of the present application, the vehicle operation state data includes vehicle operation state data reported at a plurality of reporting times when N vehicles pass through the same road segment, the vehicle operation state data includes a speed value and an acceleration, and the identification matching module 620 includes: the dividing sub-module is used for dividing the multiple reporting time scores of each vehicle into different speed intervals based on multiple speed values reported by each vehicle in the N vehicles at multiple reporting time instants; the determining submodule is used for determining an acceleration mean value and an acceleration standard deviation corresponding to each speed interval based on the accelerations corresponding to all reporting moments in each speed interval; the determining submodule is further used for determining abnormal driving judgment conditions matched with each speed interval based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval; the screening sub-module is used for screening target reporting moments meeting abnormal driving judgment conditions matched with the speed interval from all the reporting moments falling into the speed interval to obtain abnormal driving time points, wherein the abnormal driving time points are reporting moments corresponding to target vehicles in N vehicles; the selecting sub-module is used for selecting a first reporting time when acceleration starts to change and a second reporting time when acceleration ends to change from all reporting times in preset time periods before and after an abnormal driving time point of the target vehicle; and the intercepting sub-module is used for intercepting the vehicle running state data of the target vehicle from the first reporting moment to the second reporting moment to obtain abnormal driving behavior data.

In some embodiments of the present application, the acceleration includes a longitudinal acceleration and/or a lateral acceleration, and the abnormal driving determination condition includes any one of: the first sub-condition is that the longitudinal acceleration is greater than a first longitudinal acceleration threshold; the second sub-condition is that the longitudinal acceleration is less than a second longitudinal acceleration threshold; the third sub-condition is that the lateral acceleration is greater than the first lateral acceleration threshold; the fourth sub-condition is that the lateral acceleration is less than a second lateral acceleration threshold; the first longitudinal acceleration threshold and the second longitudinal acceleration threshold are determined based on a longitudinal acceleration mean value and a longitudinal acceleration standard deviation corresponding to each speed interval, and the first transverse acceleration threshold and the second transverse acceleration threshold are determined based on a transverse acceleration mean value and a transverse acceleration standard deviation corresponding to each speed interval.

In some embodiments of the application, the acceleration comprises a longitudinal acceleration, the determination submodule being particularly adapted to: adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first longitudinal acceleration threshold value; the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second longitudinal acceleration threshold value; the abnormal driving judgment condition is determined as that the longitudinal acceleration is greater than the first longitudinal acceleration threshold value or that the longitudinal acceleration is less than the second longitudinal acceleration threshold value.

In some embodiments of the application, the acceleration comprises a lateral acceleration, the determination submodule being particularly adapted to: adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first transverse acceleration threshold; the acceleration average value is subjected to difference operation with an acceleration standard deviation of a preset multiple to obtain a second transverse acceleration threshold value; the abnormal driving determination condition is determined as whether the lateral acceleration is greater than the first lateral acceleration threshold value or the lateral acceleration is less than the second lateral acceleration threshold value.

In some embodiments of the present application, the identification matching module 620 is specifically configured to at least one of: in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the first sub-condition, matching an abnormal label of the abnormal acceleration behavior; in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the second sub-condition, matching an abnormal label of the abnormal deceleration behavior; when the target vehicle is at the intersection of the road section at the abnormal driving time point, matching an abnormal tag of the abnormal turning behavior for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition; when the target vehicle is located in the road section at the abnormal driving time point, the abnormal tag of the abnormal lane change behavior is matched for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition.

According to the driving behavior abnormality analysis device provided by the embodiment of the application, after the vehicle running state data set is acquired, abnormal driving behavior data in the vehicle running state data set can be identified, and corresponding abnormal labels are matched for the abnormal driving behavior data, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors. Based on the method, the abnormal driving behavior data can be used as input data, the abnormal labels corresponding to the abnormal driving behavior data are used as output data, the target network model is trained, and the trained target network model is obtained, so that the target network model can have the capability of carrying out abnormal analysis and recognition on the driving behavior of the vehicle. On the basis, the real-time running state data of the vehicle is input into the trained target network model, so that the target network model can specifically analyze the driving behavior of the vehicle according to the real-time running state data, and analyze whether the driving behavior is abnormal driving behavior or not, thereby obtaining the analysis result of the vehicle, and effectively analyzing and predicting the driving behavior of the vehicle. Under the condition that the analysis result is an abnormal tag, the abnormal tag can represent the specific type of abnormal driving behavior of the vehicle, so that the affected vehicle can be screened out from the vicinity of the vehicle by combining the abnormal tag and the current position of the vehicle, and the vehicle with safety risk and abnormal driving behavior can be effectively pre-warned by pushing pre-warning information to the affected vehicle, so that the driving safety of the intelligent network-connected vehicle is improved.

The driving behavior abnormality analysis device in the embodiment of the application can be a device, and can also be a component, an integrated circuit or a chip in a terminal. The device may be a mobile electronic device or a non-mobile electronic device. By way of example, the mobile electronic device may be a cell phone, tablet computer, notebook computer, palm computer, vehicle mounted electronic device, wearable device, ultra-mobile personal computer (ultra-mobile personal computer, UMPC), netbook or personal digital assistant (personal digital assistant, PDA), etc., and the non-mobile electronic device may be a server, network attached storage (Network Attached Storage, NAS), personal computer (personal computer, PC), television (TV), teller machine or self-service machine, etc., and embodiments of the present application are not limited in particular.

The driving behavior abnormality analysis device in the embodiment of the application may be a device having an operating system. The operating system may be an Android operating system, an iOS operating system, or other possible operating systems, and the embodiment of the present application is not limited specifically.

As shown in fig. 7, the electronic device 700 in this embodiment may include a processor 701 and a memory 702 storing computer program instructions.

In particular, the processor 701 may comprise a Central Processing Unit (CPU), or an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), or may be configured as one or more integrated circuits implementing embodiments of the present application.

Memory 702 may include mass storage for data or instructions. By way of example, and not limitation, memory 702 may comprise a Hard Disk Drive (HDD), floppy Disk Drive, flash memory, optical Disk, magneto-optical Disk, magnetic tape, or universal serial bus (Universal Serial Bus, USB) Drive, or a combination of two or more of the foregoing. The memory 702 may include removable or non-removable (or fixed) media, where appropriate. Memory 702 may be internal or external to the integrated gateway disaster recovery device, where appropriate. In a particular embodiment, the memory 702 is a non-volatile solid state memory. The Memory may include Read-Only Memory (ROM), random-access Memory (Random Access Memory, RAM), magnetic disk storage media devices, optical storage media devices, flash Memory devices, electrical, optical, or other physical/tangible Memory storage devices. Thus, in general, the memory includes one or more tangible (non-transitory) computer-readable storage media (e.g., memory devices) encoded with software comprising computer-executable instructions and when the software is executed (e.g., by one or more processors) it is operable to perform the operations described with reference to methods in accordance with embodiments of the application.

The processor 701 implements any one of the driving behavior abnormality analysis methods of the above embodiments by reading and executing the computer program instructions stored in the memory 702.

In one example, the electronic device 700 may also include a communication interface 703 and a bus 710. As shown in fig. 7, the processor 701, the memory 702, and the communication interface 703 are connected by a bus 710 and perform communication with each other.

The communication interface 703 is mainly used for implementing communication between each module, device, unit and/or apparatus in the embodiment of the present application.

Bus 710 includes hardware, software, or both that couple the components of the online data flow billing device to each other. By way of example, and not limitation, the buses may include an Accelerated Graphics Port (AGP) or other graphics bus, an Enhanced Industry Standard Architecture (EISA) bus, a Front Side Bus (FSB), a HyperTransport (HT) interconnect, an Industry Standard Architecture (ISA) bus, an infiniband interconnect, a Low Pin Count (LPC) bus, a memory bus, a micro channel architecture (MCa) bus, a Peripheral Component Interconnect (PCI) bus, a PCI-Express (PCI-X) bus, a Serial Advanced Technology Attachment (SATA) bus, a video electronics standards association local (VLB) bus, or other suitable bus, or a combination of two or more of the above. Bus 710 may include one or more buses, where appropriate. Although embodiments of the application have been described and illustrated with respect to a particular bus, the application contemplates any suitable bus or interconnect.

The electronic device provided by the embodiment of the present application can implement each process implemented in the method embodiments of fig. 1 to 5, and can implement the same technical effects, and for avoiding repetition, a detailed description is omitted herein.

In combination with the driving behavior abnormality analysis method in the above embodiment, the embodiment of the present application may provide a driving behavior abnormality analysis apparatus including the electronic device in the above embodiment. The details of the electronic device may be referred to the related descriptions in the above embodiments, and will not be described herein.

In addition, in combination with the driving behavior abnormality analysis method in the above embodiment, the embodiment of the present application may be implemented by providing a computer-readable storage medium. The computer readable storage medium has stored thereon computer program instructions; the computer program instructions, when executed by a processor, implement the steps of any of the driving behavior abnormality analysis methods of the above embodiments.

In combination with the driving behavior abnormality analysis method in the above embodiment, an embodiment of the present application may be implemented by providing a computer program product. The instructions in the computer program product, when executed by a processor of an electronic device, cause the electronic device to perform the steps of the driving behaviour abnormal analysis method as shown in any one of the embodiments of the first aspect.

It should be understood that the application is not limited to the particular arrangements and instrumentality described above and shown in the drawings. For the sake of brevity, a detailed description of known methods is omitted here. In the above embodiments, several specific steps are described and shown as examples. However, the method processes of the present application are not limited to the specific steps described and shown, and those skilled in the art can make various changes, modifications and additions, or change the order between steps, after appreciating the spirit of the present application.

The functional blocks shown in the above-described structural block diagrams may be implemented in hardware, software, firmware, or a combination thereof. When implemented in hardware, it may be, for example, an electronic circuit, an application specific integrated circuit (Application Specific Integrated Circuit, ASIC), suitable firmware, a plug-in, a function card, or the like. When implemented in software, the elements of the application are the programs or code segments used to perform the required tasks. The program or code segments may be stored in a machine readable medium or transmitted over transmission media or communication links by a data signal carried in a carrier wave. A "machine-readable medium" may include any medium that can store or transfer information. Examples of machine-readable media include electronic circuitry, semiconductor memory devices, ROM, flash memory, erasable ROM (EROM), floppy disks, CD-ROMs, optical disks, hard disks, fiber optic media, radio Frequency (RF) links, and the like. The code segments may be downloaded via computer networks such as the internet, intranets, etc.

It should also be noted that the exemplary embodiments mentioned in this disclosure describe some methods or systems based on a series of steps or devices. However, the present application is not limited to the order of the above-described steps, that is, the steps may be performed in the order mentioned in the embodiments, or may be performed in a different order from the order in the embodiments, or several steps may be performed simultaneously.

Aspects of the present disclosure are described above with reference to flowchart illustrations and/or block diagrams of methods, apparatus (systems) and computer program products according to embodiments of the disclosure. It will be understood that each block of the flowchart illustrations and/or block diagrams, and combinations of blocks in the flowchart illustrations and/or block diagrams, can be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, special purpose computer, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, enable the implementation of the functions/acts specified in the flowchart and/or block diagram block or blocks. Such a processor may be, but is not limited to being, a general purpose processor, a special purpose processor, an application specific processor, or a field programmable logic circuit. It will also be understood that each block of the block diagrams and/or flowchart illustration, and combinations of blocks in the block diagrams and/or flowchart illustration, can be implemented by special purpose hardware which performs the specified functions or acts, or combinations of special purpose hardware and computer instructions.

In the foregoing, only the specific embodiments of the present application are described, and it will be clearly understood by those skilled in the art that, for convenience and brevity of description, the specific working processes of the systems, modules and units described above may refer to the corresponding processes in the foregoing method embodiments, which are not repeated herein. It should be understood that the scope of the present application is not limited thereto, and any equivalent modifications or substitutions can be easily made by those skilled in the art within the technical scope of the present application, and they should be included in the scope of the present application.

Claims

1. A driving behavior abnormality analysis method, characterized by comprising:

acquiring a vehicle running state data set;

identifying abnormal driving behavior data in the vehicle running state data set, and matching corresponding abnormal labels for the abnormal driving behavior data, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors;

taking the abnormal driving behavior data as input data, taking an abnormal label corresponding to the abnormal driving behavior data as output data, and training a target network model to obtain a trained target network model;

The driving behavior of the vehicle is analyzed by inputting real-time running state data of the vehicle into the trained target network model, and an analysis result of the vehicle is obtained;

and screening the affected vehicles based on the current position of the vehicle and the abnormal label and pushing early warning information to the affected vehicles under the condition that the analysis result is the abnormal label.

2. The method of claim 1, wherein the target network model is a residual network model based on a self-attention mechanism, the target network model comprising 16 residual modules and 12 attention modules.

3. The method of claim 1, wherein prior to said inputting real-time operating state data of a vehicle to said trained target network model, said method further comprises:

and under the condition that the real-time running state data of the vehicle is obtained, carrying out normalization processing and denoising processing on the real-time running state data.

4. The method of claim 1, wherein the vehicle operating state data set includes vehicle operating state data reported at a plurality of reporting times when N vehicles traverse the same road segment, the vehicle operating state data including a speed value and an acceleration, the identifying abnormal driving behavior data in the vehicle operating state data set comprising:

Dividing the reporting time scores of each vehicle into different speed intervals based on a plurality of speed values reported by each vehicle in the N vehicles at a plurality of reporting times;

based on the acceleration corresponding to all reporting moments in each speed interval, determining an acceleration mean value and an acceleration standard deviation corresponding to each speed interval;

determining abnormal driving judgment conditions matched with each speed interval based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval;

screening target reporting moments meeting abnormal driving judgment conditions matched with the speed interval from all reporting moments falling into the speed interval to obtain abnormal driving time points, wherein the abnormal driving time points are reporting moments corresponding to target vehicles in the N vehicles;

selecting a first reporting time when acceleration starts to change and a second reporting time when acceleration ends to change from all reporting times in preset time before and after an abnormal driving time point of the target vehicle;

and intercepting the vehicle running state data of the target vehicle from the first reporting time to the second reporting time to obtain the abnormal driving behavior data.

5. The method according to claim 4, wherein the acceleration includes a longitudinal acceleration and/or a lateral acceleration, and the abnormal driving determination condition includes any one of:

a second sub-condition is that the longitudinal acceleration is less than a second longitudinal acceleration threshold;

a third sub-condition is that the lateral acceleration is greater than a first lateral acceleration threshold;

a fourth sub-condition is that the lateral acceleration is less than a second lateral acceleration threshold;

the first longitudinal acceleration threshold and the second longitudinal acceleration threshold are determined based on a longitudinal acceleration mean value and a longitudinal acceleration standard deviation corresponding to each speed interval, and the first transverse acceleration threshold and the second transverse acceleration threshold are determined based on a transverse acceleration mean value and a transverse acceleration standard deviation corresponding to each speed interval.

6. The method of claim 5, wherein the acceleration comprises a longitudinal acceleration, and the determining the abnormal driving judgment condition matched with each speed interval based on the acceleration average value and the acceleration standard deviation corresponding to each speed interval comprises:

Adding the acceleration mean value and the acceleration standard deviation of a preset multiple to obtain a first longitudinal acceleration threshold;

the acceleration mean value and the acceleration standard deviation of the preset multiple are subjected to difference to obtain a second longitudinal acceleration threshold value;

and determining that the abnormal driving judgment condition is that the longitudinal acceleration is larger than the first longitudinal acceleration threshold value or the longitudinal acceleration is smaller than the second longitudinal acceleration threshold value.

7. The method of claim 5, wherein the acceleration comprises a lateral acceleration, and the determining the abnormal driving determination condition matched to each speed interval based on the acceleration mean value and the acceleration standard deviation corresponding to each speed interval comprises:

performing difference between the acceleration mean value and an acceleration standard deviation of a preset multiple to obtain a second transverse acceleration threshold value;

and determining that the abnormal driving judgment condition is that the lateral acceleration is larger than the first lateral acceleration threshold value or the lateral acceleration is smaller than the second lateral acceleration threshold value.

8. The method of claim 5, wherein matching the abnormal driving behavior data with a corresponding abnormal tag comprises at least one of:

in order to meet the abnormal driving behavior data corresponding to the abnormal driving time point of the first sub-condition, an abnormal label of the abnormal acceleration behavior is matched;

when the target vehicle is located at the intersection of the road section at the abnormal driving time point, matching an abnormal tag of an abnormal turning behavior for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition;

and when the target vehicle is positioned in the road section at the abnormal driving time point, matching an abnormal label of the abnormal lane change behavior for abnormal driving behavior data corresponding to the abnormal driving time point meeting the third sub-condition.

9. A driving behavior abnormality analysis device, characterized by comprising:

the acquisition module is used for acquiring a vehicle running state data set;

the recognition matching module is used for recognizing abnormal driving behavior data in the vehicle running state data set and matching corresponding abnormal labels for the abnormal driving behavior data, wherein the abnormal labels comprise abnormal acceleration behaviors, abnormal deceleration behaviors, abnormal lane changing behaviors and abnormal turning behaviors;

The training module is used for training the target network model by taking the abnormal driving behavior data as input data and the abnormal labels corresponding to the abnormal driving behavior data as output data to obtain a trained target network model;

the analysis module is used for analyzing the driving behavior of the vehicle by inputting the real-time running state data of the vehicle into the trained target network model to obtain an analysis result of the vehicle;

and the early warning module is used for screening the affected vehicle based on the current position of the vehicle and the abnormal label and pushing early warning information to the affected vehicle under the condition that the analysis result is the abnormal label.

10. An electronic device, the electronic device comprising: a processor and a memory storing computer program instructions;

the processor, when executing the computer program instructions, implements the driving behavior abnormality analysis method according to any one of claims 1 to 8.

11. A computer readable storage medium, characterized in that the computer readable storage medium has stored thereon computer program instructions which, when executed by a processor, implement the steps of the driving behavior abnormality analysis method according to any one of claims 1-8.