CN106781503B - Method and apparatus for monitoring driving behavior - Google Patents
Method and apparatus for monitoring driving behavior Download PDFInfo
- Publication number
- CN106781503B CN106781503B CN201710046084.6A CN201710046084A CN106781503B CN 106781503 B CN106781503 B CN 106781503B CN 201710046084 A CN201710046084 A CN 201710046084A CN 106781503 B CN106781503 B CN 106781503B
- Authority
- CN
- China
- Prior art keywords
- driver
- value
- preset
- driving behavior
- label
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Classifications
-
- G—PHYSICS
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0129—Traffic data processing for creating historical data or processing based on historical data
Abstract
This application discloses the method and apparatus for monitoring driving behavior.One specific embodiment of the method includes: to be analyzed using driver behavior modeling model historical trajectory data, the driving behavior of monitoring objective driver;Driver behavior modeling model is obtained through the following steps: obtaining the historical trajectory data set and running cost data acquisition system of multiple drivers;Historical trajectory data set and running cost data acquisition system are handled respectively, determine the driving behavior parameter value and running cost parameter value of each driver;Respectively to each the first label of driver's label and the second label;Learn the driving behavior parameter value of the driving behavior parameter value of multiple drivers of the first label and multiple drivers of the second label using machine learning algorithm, extracts characteristic parameter, and above-mentioned driver behavior modeling model is determined based on characteristic parameter.The embodiment realizes the abundant excavation to the historical trajectory data of driver, carries out quantization monitoring to the driving behavior of driver.
Description
Technical field
This application involves field of computer technology, and in particular to driver behavior modeling field, more particularly to it is a kind of for supervising
The method and apparatus for surveying driving behavior.
Background technique
With economic society sustained and rapid development, masses' purchase vehicle rigid demand is vigorous, and car ownership continues in quickly increasing
Long trend.It is adapted with vehicle guaranteeding organic quantity rapid growth, the trend that increases substantially also is presented in vehicle driver quantity.Motor vehicle
And driver's quantity increases rapidly, while offering convenience to the production and living of people, also brings that cannot be neglected safety hidden
Suffer from.
Low driving age driver due to it is unskilled it is caused it is bad drive to be accustomed to, it is such as anxious to accelerate or anxious slow down;And the high driving age
The high speed joyride of driver etc. is bad to drive to be accustomed to, and can all endanger the life and health of pedestrian or driver.Therefore, how to quantify
The driving behavior of driver is monitored and is a problem to be solved.
Summary of the invention
The purpose of the application is to propose a kind of method and apparatus for monitoring driving behavior, to solve background above skill
The technical issues of art part is mentioned.
In a first aspect, the above method includes: to obtain to drive this application provides a kind of method for monitoring driving behavior
The historical trajectory data of behavior monitoring model and target driver;Using driver behavior modeling model to historical trajectory data into
Row analysis, the driving behavior of monitoring objective driver;Wherein, above-mentioned driver behavior modeling model is obtained through the following steps
: obtain the historical trajectory data set and running cost data acquisition system of multiple drivers;Historical trajectory data collection is handled respectively
Conjunction and running cost data acquisition system, determine the driving behavior parameter value and running cost parameter value of each driver;Will traveling at
This parameter value is labeled as the first label less than the driver of the first preset value, and running cost parameter value is greater than the second preset value
Driver is labeled as the second label;Learn the driving behavior of multiple drivers indicated by the first label using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by parameter value and the second label extracts the driving row for meeting preset condition
It is parameter as characteristic parameter, and above-mentioned driver behavior modeling model is determined based on characteristic parameter.
In some embodiments, historical trajectory data includes the location and time of multiple tracing points;And processing is gone through respectively
History track data set and running cost data acquisition system determine the driving behavior parameter value and running cost parameter of each driver
Value, comprising: determine the displacement and duration between any two are adjacent in historical trajectory data tracing point;It is timely according to displacement
It is long, each driver is determined in the driving behavior parameter value of each tracing point, and above-mentioned driving behavior parameter includes: speed and traveling angle
Degree.
In some embodiments, according to the displacement and duration between the adjacent tracing point of any two, each driving is determined
Driving behavior parameter value of the member in each tracing point, comprising: according to each driver in the velocity amplitude of each tracing point, determine and each drive
The velocity distribution curve for the person of sailing;Determine that velocity amplitude in velocity distribution curve is greater than the number of preset friction speed threshold value and every
The secondary duration greater than preset friction speed threshold value;According to velocity distribution curve, the cumulative distribution function of speed is determined;Really
Preset different probability is worth corresponding velocity amplitude in the cumulative distribution function of constant speed degree.
In some embodiments, above-mentioned driving behavior parameter further include: acceleration;And the rail adjacent according to any two
Displacement and duration between mark point determine each driver in the driving behavior parameter value of each tracing point, comprising: to be driven according to each
The velocity distribution curve for the person of sailing determines the acceleration of each tracing point;Determine the acceleration profile curve of each driver;It determines
Acceleration value is greater than the number of preset different acceleration rate thresholds and is greater than preset different acceleration in acceleration profile curve
Spend the first total duration of threshold value;Determine that acceleration value is less than zero in acceleration profile curve and absolute value is greater than preset difference and adds
The number and absolute value of threshold speed are greater than the second total duration of different acceleration rate thresholds;According to the acceleration of each driver
Distribution curve determines the cumulative distribution function of acceleration;Determine preset different probability value in the cumulative distribution function of acceleration
Corresponding acceleration value.
In some embodiments, above-mentioned driving behavior parameter includes continuous driving duration;And it is adjacent according to any two
Tracing point between displacement and duration, determine each driver in the driving behavior parameter value of each tracing point, comprising: for every
A driver determines that displacement is less than preset displacement and duration according to the displacement and duration between the adjacent tracing point of any two
Less than preset duration two adjacent track points at the time of;According at the time of identified tracing point and in historical trajectory data
At the time of first tracing point, at the time of the last one tracing point, at least one continuous driving duration of each driver is determined.
In some embodiments, according to the displacement and duration between the adjacent tracing point of any two, each driving is determined
Driving behavior parameter value of the member in each tracing point, comprising: for each driver, according to the traveling angle of each tracing point, really
Surely traveling angle is located at the velocity distribution curve within the scope of preset different angle;Determine that each velocity distribution curve is corresponding tired
Product distribution function;Determine the corresponding velocity amplitude of different probability in different cumulative distribution function.
In some embodiments, above-mentioned running cost parameter includes: the maintenance of the fuel consumption values and unit distance of unit distance
Cost value;And the driver by running cost parameter less than the first preset value is labeled as the first label, by running cost parameter
Driver greater than the second preset value is labeled as the second label, comprising: by the maintenance of the fuel consumption values of unit distance and unit distance
Cost value is respectively less than the driver of the first preset value labeled as the first label;By the fuel consumption values of unit distance and the dimension of unit distance
It repairs cost value and is all larger than the driver of the second preset value labeled as the second label.
In some embodiments, above-mentioned first label is third preset value, and above-mentioned second label is the 4th preset value;And
Learn driving behavior parameter value and the second label institute of multiple drivers indicated by the first label using machine learning algorithm
The driving behavior parameter value of the multiple drivers indicated extracts and meets the driving behavior parameter of preset condition as characteristic parameter,
It include: to be learnt using machine learning algorithm by the driving behavior parameter value and second of multiple drivers indicated by the first label
The driving behavior parameter value of multiple drivers indicated by label determines the P value of each driving behavior parameter;Calculate each driving row
For the Pearson correlation coefficients between parameter value and third preset value or the 4th preset value;Extract P value be greater than the 5th preset value and
Pearson correlation coefficients are greater than the driving behavior parameter of the 6th preset value as characteristic parameter.
In some embodiments, driver behavior modeling model is determined based on characteristic parameter, comprising: based on each characteristic parameter
P value and Pearson correlation coefficients determine the weight of each characteristic parameter;Based on each characteristic parameter and corresponding weight, driven
Behavior monitoring model.
Second aspect, this application provides a kind of for monitoring the device of driving behavior, and above-mentioned apparatus includes: to obtain list
Member, for obtaining the historical trajectory data of driver behavior modeling model and target driver;Monitoring unit, for utilizing driving
Behavior monitoring model analyzes historical trajectory data, the driving behavior of monitoring objective driver;Wherein, driver behavior modeling
Model is obtained by model construction unit, and model construction unit includes: acquisition module, for obtaining going through for multiple drivers
History track data set and running cost data acquisition system;Processing module, for handling historical trajectory data set and traveling respectively
Cost data set determines the driving behavior parameter value and running cost parameter value of each driver;Mark module, for that will go
Driver of the cost parameter value less than the first preset value is sailed labeled as the first label, running cost parameter value is greater than second and is preset
The driver of value is labeled as the second label;Determining module is more indicated by the first label for being learnt using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by the driving behavior parameter value and the second label of a driver extracts symbol
The driving behavior parameter for closing preset condition determines above-mentioned driver behavior modeling model as characteristic parameter, and based on characteristic parameter.
In some embodiments, historical trajectory data includes the location and time of multiple tracing points;And processing module into
One step is used for: determining the displacement and duration between any two are adjacent in historical trajectory data tracing point;According to above-mentioned displacement
And duration, each driver is determined in the driving behavior parameter value of each tracing point, and above-mentioned driving behavior parameter includes: speed and row
Sail angle.
In some embodiments, processing module is further used for: according to each driver each tracing point velocity amplitude, really
The velocity distribution curve of fixed each driver;Determine that velocity amplitude is greater than the secondary of preset friction speed threshold value in velocity distribution curve
Number and the duration for being greater than preset friction speed threshold value every time;According to velocity distribution curve, the iterated integral of speed is determined
Cloth function;Determine the corresponding velocity amplitude of preset different probability value in the cumulative distribution function of speed.
In some embodiments, above-mentioned driving behavior parameter further include: acceleration;And processing module is further used for:
According to the velocity distribution curve of each driver, the acceleration of each tracing point is determined;Determine the acceleration point of each driver
Cloth curve;Determine that acceleration value is greater than the number of preset different acceleration rate thresholds and is greater than default in acceleration profile curve
Different acceleration rate thresholds the first total duration;Determine that acceleration value is less than zero in acceleration profile curve and absolute value is greater than in advance
If different acceleration rate thresholds number and absolute value be greater than different acceleration rate thresholds the second total duration;According to each driving
The acceleration profile curve of member, determines the cumulative distribution function of acceleration;It determines preset in the cumulative distribution function of acceleration
Different probability is worth corresponding acceleration value.
In some embodiments, above-mentioned driving behavior parameter includes: continuous driving duration;And processing module is further used
In: for each driver, according to the displacement and duration between the adjacent tracing point of any two, determine that displacement is less than default position
Move and duration be less than preset duration two adjacent track points at the time of;According at the time of identified tracing point and history rail
In mark data at the time of first tracing point, at the time of the last one tracing point, at least one for determining each driver is continuous
Drive duration.
In some embodiments, processing module is further used for: for each driver, according to the traveling of each tracing point
Angle determines that traveling angle is located at the velocity distribution curve within the scope of preset different angle;Determine each velocity distribution curve
Corresponding cumulative distribution function;Determine the corresponding velocity amplitude of different probability in different cumulative distribution function.
In some embodiments, above-mentioned running cost parameter includes: the maintenance of the fuel consumption values and unit distance of unit distance
Cost value;And mark module is further used for: the maintenance cost value of the fuel consumption values of unit distance and unit distance is respectively less than
The driver of first preset value is labeled as the first label;The maintenance cost value of the fuel consumption values of unit distance and unit distance is big
The second label is labeled as in the driver of the second preset value.
In some embodiments, above-mentioned first label is third preset value, and above-mentioned second label is the 4th preset value;And
Determining module is further used for: being learnt using machine learning algorithm by the driving behavior of multiple drivers indicated by the first label
The driving behavior parameter value of multiple drivers indicated by parameter value and the second label determines the P of each driving behavior parameter
Value;Calculate the Pearson correlation coefficients between each driving behavior parameter value and third preset value or the 4th preset value;It is big to extract P value
In the 5th preset value and Pearson correlation coefficients are greater than the driving behavior parameter of the 6th preset value as characteristic parameter.
In some embodiments, determining module is further used for: P value and pearson correlation system based on each characteristic parameter
Number, determines the weight of each characteristic parameter;Based on each characteristic parameter and corresponding weight, above-mentioned driver behavior modeling model is obtained.
Method and apparatus provided by the present application for monitoring driving behavior, pass through the historical track number to multiple drivers
It is respectively processed according to set and running cost parameter sets, obtains the driving behavior parameter value and running cost of each driver
Parameter value, be then according to the running cost parameter of each driver meet the driver of different condition to mark different labels, then
Learn the driving behavior parameter value of the driver of different labels using machine learning algorithm, extracts the driving row for meeting preset condition
It is parameter as characteristic parameter, driver behavior modeling model is determined based on features described above parameter, above-mentioned driving behavior is recycled to supervise
It surveys model to analyze the historical trajectory data of target driver, realizes the monitoring to target driver driving behavior.It is above-mentioned
Embodiment realizes the abundant excavation to the historical trajectory data of driver, carries out quantization monitoring to the driving behavior of driver.
Detailed description of the invention
By reading a detailed description of non-restrictive embodiments in the light of the attached drawings below, the application's is other
Feature, objects and advantages will become more apparent upon:
Fig. 1 is the flow chart according to one embodiment of the method for monitoring driving behavior of the application;
Fig. 2 is that this application can be applied to exemplary system architecture figures therein;
Fig. 3 is the process according to the determination driver behavior modeling model of the method for monitoring driving behavior of the application
Figure;
Fig. 4 is the driving behavior parameter according to each driver of determination of the method for monitoring driving behavior of the application
One embodiment flow chart;
Fig. 4 a is the schematic diagram according to the velocity distribution curve of the method for monitoring driving behavior of the application;
Fig. 5 is the structural schematic diagram according to one embodiment of the device for monitoring driving behavior of the application;
Fig. 6 is adapted for the structural representation of the computer system for the terminal device or server of realizing the embodiment of the present application
Figure.
Specific embodiment
The application is described in further detail with reference to the accompanying drawings and examples.It is understood that this place is retouched
The specific embodiment stated is used only for explaining related invention, rather than the restriction to the invention.It also should be noted that in order to
Convenient for description, part relevant to related invention is illustrated only in attached drawing.
It should be noted that in the absence of conflict, the features in the embodiments and the embodiments of the present application can phase
Mutually combination.The application is described in detail below with reference to the accompanying drawings and in conjunction with the embodiments.
Fig. 1 shows the flow chart 100 of one embodiment of the method for monitoring driving behavior according to the application.Such as
Shown in Fig. 1, the present embodiment for monitor driving behavior method the following steps are included:
Step 101, the historical trajectory data of driver behavior modeling model and target driver is obtained.
In the present embodiment, driver behavior modeling model can be the monitoring model constructed according to various machine learning algorithms,
Its corresponding relationship that may include historical trajectory data Yu driver behavior modeling result.That is, by a historical trajectory data
Input above-mentioned driver behavior modeling model, so that it may obtain a driver behavior modeling result.Above-mentioned monitoring result may include
At least one of below: the numerical value, the analysis and summary to driving behavior of various parameters (such as speed, acceleration) in driving procedure
(such as often hypervelocity, lane change are excessively frequent), to scoring of driving behavior etc..
Above-mentioned target driver can be driving behavior driver to be monitored.Such as when the driving behavior of the present embodiment is supervised
When surveying model applied to car insurance company, above-mentioned target driver be can be by the owner of insurance vehicle, and insurance company can
According to the monitoring result of the driving behavior of target driver, insurance premium is set;Or the driver behavior modeling mould when the present embodiment
When type is used for cartography company, above-mentioned target driver can be streetscape map acquisition person, and cartography company can be according to mesh
The monitoring result for marking the driving behavior of driver is managed each driver.
Above-mentioned historical trajectory data can be the data that target driver drives vehicle running path, and above-mentioned driving path can
It may include the position of each tracing point to be made of multiple tracing points, in above-mentioned historical trajectory data, travel angle and arrive
At the time of up to the tracing point etc..Above-mentioned position can be coordinate (such as GPS coordinate), can also be street information (such as XX
The city XX, the province area the XX street XX XX);Above-mentioned traveling angle can be tangent line and direct north of the tracing point on driving path
Angle, can also be the tracing point on driving path tangent line and a upper tracing point between the tangent line on driving path
Angle.It is understood that above-mentioned historical trajectory data is formed according to being arranged successively at the time of reaching each tracing point.
The method for monitoring driving behavior of the present embodiment is generally executed by terminal or server, above-mentioned terminal or clothes
Business device can be connect with vehicle communication.Terminal or server when obtaining above-mentioned historical trajectory data, can directly from vehicle
Each sensor connection storage device in obtain, the data being locally stored can also be obtained.
In the present embodiment, when server needs to obtain historical trajectory data from vehicle, corresponding system architecture diagram is such as
Shown in Fig. 2.In Fig. 2, system architecture 200 may include vehicle 201, network 202 and server 203.Network 202 is in vehicle
The medium of communication link is provided between 201 and server 203.Network 202 may include various connection types, such as wired, nothing
Line communication link or fiber optic cables etc..
It may include GPS chip on vehicle 201, can recorde the traveling-position of vehicle 201;It can also include that timing fills
It sets, at the time of can recorde vehicle 201 and reach each traveling-position.Certainly, on vehicle 201 can also include velocity sensor,
Acceleration transducer etc. records vehicle 201 in the travel speed and acceleration of each traveling-position respectively.
Server 203 can be to provide the server of various services, for example, to the historical trajectory data of vehicle 201 at
The background server of reason.The historical trajectory data of the available vehicle 201 of background server, and analyzed to obtain driver's
Driver behavior modeling result.
It should be noted that for monitoring the method for driving behavior generally by server provided by the embodiment of the present application
203 execute, correspondingly, device for monitoring driving behavior be generally arranged in server 203.
It should be understood that the number of vehicle, network and server in Fig. 2 is only schematical.It, can according to needs are realized
With any number of vehicle, network and server.
Fig. 1 is returned to analyze historical trajectory data using driver behavior modeling model, monitoring objective in step 102
The driving behavior of driver.
After getting above-mentioned driver behavior modeling model and historical trajectory data, driver behavior modeling model can use
Historical trajectory data is analyzed, the driving behavior of monitoring objective driver obtains monitoring result.Wherein, above-mentioned driving row
It for monitoring model is obtained by step 301 shown in Fig. 3~step 304, Fig. 3 is driving for monitoring according to the application
Sail the flow diagram 300 of the determination driver behavior modeling model of the method for behavior.
Step 301, the historical trajectory data set and running cost data acquisition system of multiple drivers are obtained.
When creating above-mentioned driver behavior modeling model, first have to the historical trajectory data set for obtaining multiple drivers and
Running cost parameter sets.Wherein, running cost data may include the expense spent in unit operating range, such as oil consumption expense
With, maintenance cost, the expense of traffic accident etc..
Step 302, historical trajectory data set and running cost data acquisition system are handled respectively, determine driving for each driver
Sail behavioral parameters value and running cost parameter value.
After obtaining above-mentioned historical data set and running cost data acquisition system, above two data set is handled respectively
It closes, with the driving behavior parameter and running cost parameter of each driver of determination.Above-mentioned processing may include calculating above-mentioned traveling
The mileage in path also may include the speed for calculating each tracing point, angle etc., can also include calculating above-mentioned driving path institute
The expense etc. of cost.Driving behavior parameter can be the various various parameters that can reflect driver's driving efficiency or habit, example
It such as can be speed, acceleration, traveling angle.Running cost parameter can be what reflection driver spent in unit distance
The various parameters of expense, such as can be oil consumption expense, maintenance cost.
Step 303, by running cost parameter value less than the first preset value driver be labeled as the first label, will traveling at
The driver that this parameter value is greater than the second preset value is labeled as the second label.
Running cost parameter value is being obtained, is being a part setting label of above-mentioned multiple drivers.Specifically, will traveling at
This parameter value is labeled as the first label less than the driver of the first preset value, and running cost parameter value is greater than the second preset value
Driver is labeled as the second label.Above-mentioned first label, the second label can indicate by different numerical value, can also be by different
Storage location indicates.It is understood that above-mentioned first preset value is less than or equal to the second preset value.
It, can also be first when setting label for above-mentioned driver in some optional implementations of the present embodiment
According to the running cost parameter value of each driver, ascending sequence is carried out.By running cost parameter value in above-mentioned sequence preceding
The driver of first preset ratio is labeled as the first label, by running cost parameter value in above-mentioned sequence in rear second preset ratio
Driver be labeled as the second label.
Step 304, learn the driving behavior parameter of multiple drivers indicated by the first label using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by value and the second label extracts the driving behavior ginseng for meeting preset condition
Number is used as characteristic parameter, and determines driver behavior modeling model based on characteristic parameter.
After setting label for the part driver in multiple drivers, first is labeled as using machine learning algorithm study
The driving behavior parameter value of the driving behavior parameter value of the driver of label and the driver labeled as the second label.Above-mentioned machine
Learning algorithm can be including logistic regression algorithm, support vector machines, random forest, decision tree etc..It is understood that this reality
It applies example not to be defined the type of machine learning algorithm, those skilled in the art can select machine according to practical application scene
The type of device learning algorithm.
Machine learning algorithm can go after learning the driving behavior parameter value of driver of different labels from multiple driving
To extract the driving behavior parameter for meeting preset condition in parameter as characteristic parameter, above-mentioned preset condition may include and first
The correlation of label or the second label is greater than the relevance between threshold value, each driving behavior parameter and is less than threshold value etc..It is then based on
It states characteristic parameter and determines that driver behavior modeling model, driver behavior modeling model can be to each characteristic ginseng value by various fortune
The model that a numerical value is exported after calculation is also possible to carry out each characteristic ginseng value on the model of analysis one report of output.
The method provided by the above embodiment for monitoring driving behavior of the application, passes through the history to multiple drivers
Track data set and running cost parameter sets are respectively processed, and obtain the driving behavior parameter value and row of each driver
Cost parameter value is sailed, is then to meet the driver of different condition to mark different marks according to the running cost parameter of each driver
Label, recycle machine learning algorithm to learn the driving behavior parameter value of the driver of different labels, and extraction meets preset condition
Driving behavior parameter determines driver behavior modeling model as characteristic parameter, based on features described above parameter, recycles above-mentioned driving
Behavior monitoring model analyzes the historical trajectory data of target driver, realizes the prison to target driver driving behavior
It surveys.Above example implements the abundant excavations of the historical trajectory data to driver, to the driving behavior amount of progress of driver
Change monitoring.
In some optional implementations of the present embodiment, above-mentioned running cost parameter may include the oil of unit distance
The maintenance cost value of consumption value and unit distance.Above-mentioned steps 303 specifically can be by following sub-step unshowned in Fig. 3 come real
It is existing:
The maintenance cost value of the fuel consumption values of unit distance and unit distance is respectively less than to driver's label of the first preset value
For the first label;The maintenance cost value of the fuel consumption values of unit distance and unit distance is all larger than to driver's mark of the second preset value
It is denoted as the second label.
The fuel consumption values of unit distance and the maintenance cost value of unit distance are smaller, illustrate the driving habit of the driver compared with
Good, the probability that traffic accident occurs is small, and oil consumption is less, sets the first label for such driver.The fuel consumption values of unit distance
And the maintenance cost value of unit distance is larger, illustrates that the driver's is poor, the probability that traffic accident occurs is big, and oil consumption compared with
It is more, the second label is set for such driver.In this implementation, above-mentioned first label and the second label can be respectively " 0 "
" 1 ".
Fig. 4 shows the driving behavior of each driver of determination of the method for monitoring driving behavior according to the application
The flow chart 400 of one embodiment of parameter, Fig. 4 a show the schematic diagram of velocity distribution curve.It is above-mentioned to go through in the present embodiment
History track data includes the location and time of multiple tracing points.As shown in Figure 4, the above method the following steps are included:
Step 401, the displacement and duration between any two are adjacent in historical trajectory data tracing point are determined.
Due to including multiple tracing points in historical trajectory data, and the location and time including each tracing point therefore can
With the displacement and duration between the every two adjacent tracing point of determination.
In some optional implementations of the present embodiment, a when a length of fixed value between above-mentioned multiple tracing points,
For example, above-mentioned duration can be 2 seconds.
Step 402, according to above-mentioned displacement and duration, determine each driver in the driving behavior parameter value of each tracing point.
In the present embodiment, above-mentioned driving behavior parameter includes speed and traveling angle.Speed can be by the distance in displacement
It is determined with the ratio of duration, traveling angle can be determined by the angle of friendship degree and direct north in displacement.
Step 403, each driving behavior parameter value relevant to speed is determined.
Above-mentioned steps 403 can specifically be realized by following sub-step unshowned in Fig. 4:
According to each driver in the velocity amplitude of each tracing point, the velocity distribution curve of each driver is determined;Determine speed
Velocity amplitude in distribution curve is spent to be greater than the number of preset friction speed threshold value and be greater than preset friction speed threshold value every time
Duration;According to velocity distribution curve, the cumulative distribution function of speed is determined;It determines pre- in the cumulative distribution function of speed
If different probability be worth corresponding velocity amplitude.
After the velocity amplitude that each tracing point has been determined, it can determine that the speed of each driver is dependent variable, the time is certainly
The velocity distribution curve of variable.It can determine that velocity amplitude exceeds the number of friction speed threshold value in above-mentioned velocity distribution curve,
It can be determined simultaneously in the duration for exceeding above-mentioned threshold speed every time.Above-mentioned friction speed threshold value may include 50km/h,
80km/h, 120km/h etc..As shown in fig. 4 a, in velocity distribution curve, velocity amplitude is more than 50km/h 1 time, more than 80km/h 1
It is secondary, more than 120km/h 1 time, duration more than 50km/h is t6-t1, the duration more than 80km/h is t5-t2, surpass
The duration for crossing 120km/h is t4-t3.Then it according to above-mentioned velocity distribution curve, determines the cumulative distribution function of speed, tires out
Product distribution function is used to describe the probability distribution of variable.Preset different probability value can be determined in above-mentioned cumulative distribution function
Corresponding velocity amplitude, above-mentioned preset different probability value can be for from 0.2, until 0.95 terminates, step-length 0.05 be calculated separately
Probability corresponding velocity amplitude when being 0.2,0.25,0.3,0.35 ... 0.95.
Step 404, each driving behavior parameter value relevant to acceleration is determined.
In the present embodiment, driving behavior parameter further includes acceleration.It, can be according to each after velocity distribution curve has been determined
Slope at tracing point determines the acceleration of each tracing point, and above-mentioned steps 404 can specifically pass through following son unshowned in Fig. 4
Step is realized:
According to the velocity distribution curve of each driver, the acceleration of each tracing point is determined;Determine each driver's
Acceleration profile curve;Determine acceleration value in acceleration profile curve be greater than preset different acceleration rate thresholds number and
Greater than the first total duration of preset different acceleration rate thresholds;Determine that acceleration value is less than zero and absolute in acceleration profile curve
Value is greater than the second total duration of different acceleration rate thresholds greater than the number and absolute value of preset different acceleration rate thresholds;According to
The acceleration profile curve of each driver, determines the cumulative distribution function of acceleration;Determine the cumulative distribution function of acceleration
In preset different probability be worth corresponding acceleration value.
After velocity distribution curve has been determined, the acceleration of each tracing point can be determined according to the slope at each tracing point,
It may thereby determine that the acceleration profile curve of each driver.It is positive value due to accelerating brief acceleration, deceleration brief acceleration is
Negative value, in order to determine each driver with the presence or absence of anxious the case where accelerating or suddenly slowing down, it is first determined acceleration value is beyond preset
The number of different acceleration rate thresholds and the first total duration beyond above-mentioned preset different acceleration rate thresholds.Above-mentioned different adds
Threshold speed can be 10m/s2、15m/s2, the first total duration is to exceed 10m/s every time2Duration summation plus exceeding every time
15m/s2Duration summation.Then it determines and is less than the numbers of preset different acceleration rate thresholds by acceleration value (i.e. acceleration value is small
In zero, but absolute value is greater than the number of different acceleration rate thresholds) and it is less than the second total of above-mentioned different acceleration rate thresholds every time
Duration.It is understood that accelerating to slow down with anxious for anxious, acceleration rate threshold be may be the same or different.For example, right
18m/s can be respectively set in the minus threshold value of acceleration2、21m/s2, the second total duration is to be less than 18m/s every time2Duration
Summation is plus every time less than 21m/s2Duration summation.Further according to acceleration profile curve, the cumulative distribution letter of acceleration is determined
Number determines that preset different probability is worth corresponding acceleration value in the cumulative distribution function of acceleration.It is understood that this
Locating preset different probability value can be identical as the probability value in step 403, can also be different.For example, preset difference herein
Probability value can be from 0.8, until 1 terminates, step-length 0.05 calculates separately corresponding when probability is 0.85,0.9,0.95,1
Acceleration value.
It is understood that although the present embodiment schematically illustrates the value of above-mentioned threshold speed, acceleration rate threshold
Value and probability value value, but this is only schematical, and the present embodiment does not limit this, and those skilled in the art can
The value of above-mentioned threshold value is set according to practical application scene.
Process 400 in the present embodiment highlights the step of handling historical trajectory data.The present embodiment institute as a result,
The scheme of description can carry out a variety of careful processing to historical trajectory data, so as to the driving of the determination driver of quantization
Behavioral parameters.
In some optional implementations of the present embodiment, above-mentioned driving behavior parameter further includes continuous driving duration.
Further include unshowned following steps in Fig. 4 in the driving behavior parameter for determining each driver:
For each driver, according to the displacement and duration between the adjacent tracing point of any two, determine that displacement is less than
At the time of preset displacement and duration are less than two adjacent track points of preset duration;According at the time of identified tracing point and
In historical trajectory data at the time of first tracing point, at the time of the last one tracing point, at least the one of each driver is determined
A continuous driving duration.
In this implementation, when the displacement between two adjacent tracing points is less than preset displacement, and two tracing points
Between duration be greater than preset duration when, it can be assumed that driver parking rest.Above-mentioned preset displacement can be 5 meters, above-mentioned
Preset duration can be 20 minutes, this certain implementation does not limit the value of preset displacement and preset duration.Then basis
At the time of the tracing point of above-mentioned determination and in historical trajectory data at the time of first tracing point and the last one tracing point
Moment determines at least one continuous driving duration of each driver.
For example, 7200 tracing points are shared in historical trajectory data, the 1st tracing point at the time of is 9:30, the
It is 11:30 at the time of 3600 tracing points, is 11:50 at the time of the 3601st tracing point, the 7200th tracing point at the time of is
13:50.Then the driver includes two continuously driving durations, and first 2 hours a length of when continuously driving, second continuous driving
Duration is also 2 hours.
In some optional implementations of the present embodiment, determine that the driving behavior parameter of each driver further includes figure
Unshowned following steps in 4:
For each driver, according to the traveling angle of each tracing point, determine that traveling angle is located at preset different angles
Spend the velocity distribution curve in range;Determine the corresponding cumulative distribution function of each velocity distribution curve;Determine different iterated integrals
The corresponding velocity amplitude of different probability in cloth function.
In this implementation, driving habit of the driver in turning can be monitored according to historical trajectory data.
Angle of turn can be divided first, assert that angle of turn is small radian between 15 °~45 °, between 45 °~90 °
For medium radian, at 90 °, the above are big radians.Since each tracing point wraps traveling angle, angle position can will be travelled
Each tracing point between 15 °~45 ° is arranged sequentially in time, will travel each rail of the angle between 45 °~90 °
Mark point is arranged sequentially in time, and traveling angle is arranged sequentially in time in 90 ° or more of each tracing point, then
In conjunction with the speed of each tracing point, three velocity distribution curves are obtained.The corresponding iterated integral of each velocity distribution curve is determined respectively
Then cloth function determines the corresponding speed of different probability.Above-mentioned different probability can be with 0.8,0.85,0.9,0.95.It can be with
Understand, the corresponding probability value of each velocity distribution curve may be the same or different.
In some optional implementations of the present embodiment, above-mentioned driving behavior parameter can also include traveling angle with
The product of speed, referred to as turning intensity.Since the traveling angle and speed of each tracing point are it has been determined that the then turning of each tracing point
Intensity value also determines.According to the turning intensity value of each tracing point, it can determine that turning intensity value is greater than the ratio of preset threshold, together
Sample can also determine the cumulative distribution function of turning intensity, determine turn when probability value is 0.8,0.85,0.9,0.95 respectively
Curved intensity value.
In some optional implementations of the present embodiment, above-mentioned driving behavior parameter can also include speed and accelerate
The product of the absolute value of degree, referred to as acceleration power.Similarly with turning intensity, it is determined when the acceleration power value of each tracing point
's.It may thereby determine that the cumulative distribution function for accelerating power, determined respectively when probability value is 0.2,0.4,0.6,0.8,1
Accelerate power value.
After the above-mentioned processing to historical trajectory data, the driving behavior parameter value of each driver may include it is above-mentioned not
It is (including big with the corresponding velocity amplitude of probability, the number beyond friction speed threshold value and the corresponding acceleration value of duration, different probability
In zero and less than zero), beyond different acceleration rate thresholds number and total duration, less than the number of different acceleration rate thresholds and it is total when
Long, each continuous ratio for driving duration, the velocity amplitude of the corresponding each angular range of different probability, turning intensity value and being greater than preset threshold
Example, the corresponding each turning intensity value of different probability, the corresponding each acceleration power value of different probability.For being marked with driving for label
The person of sailing, using above-mentioned each driving behavior parameter value as independent variable, determines the coefficient value of each independent variable using label value as dependent variable.
The independent variable and dependent variable of each driver constitutes an equation, in order to determine the coefficient value of each dependent variable, it is desirable that equation
The number of formula is greater than the number of coefficient value to be determined, it is therefore desirable to which the number for being marked with the driver of label is greater than
The number of above-mentioned each driving behavior parameter value.In order to more fully consider each driving behavior parameter value to the shadow of driver's label
It rings, the number for the driver that can will be labeled as the first label is equal with labeled as the number of driver of the second label.
In step 304, learn the driving behavior parameter value and label value of each driver using machine learning algorithm, determine
The coefficient value of each driving behavior parameter.Later, it tests to each driving behavior parameter value, calculates each driving behavior parameter
P value.P value is smaller, shows that result is more significant.The pearson correlation system between each driving behavior parameter value and label value is calculated simultaneously
Number, Pearson correlation coefficients are the statistics for reflecting two linear variable displacement degrees of correlation.By P value be greater than the 5th preset value and
Pearson correlation coefficients are greater than the driving behavior parameter of the 6th preset value as characteristic parameter.It, can will be above-mentioned in the present embodiment
5th preset value is taken as 0.05, the 6th preset value is taken as 0.2, so that it is determined that being able to reflect multiple features ginseng of driving behavior
Number.
When determining driver behavior modeling model, the history of target driver can be determined based on above-mentioned multiple characteristic parameters
The corresponding driver behavior modeling of track data as a result, in order to protrude some parameter to the influence degree of driver behavior modeling result,
Weight can be set for each characteristic parameter.When determining the weight of each characteristic parameter, can be with P value and Pearson correlation coefficients
Independent variable substitutes into cost function, the value between [0,1] is obtained, using this value as the weight of each characteristic parameter.According to each feature
Parameter and its corresponding weight, obtain driver behavior modeling model.
It, can be by target driver when using the driving behavior of above-mentioned driver behavior modeling model monitoring target driver
Historical trajectory data import driver behavior modeling model, after above-mentioned each characteristic ginseng value is calculated, available target
The driver behavior modeling value of driver.Those skilled in the art can determine target driver according to this driver behavior modeling value
Driving behavior.
With further reference to Fig. 5, as the realization to method shown in above-mentioned each figure, this application provides one kind to drive for monitoring
One embodiment of the device of behavior is sailed, the Installation practice is corresponding with embodiment of the method shown in FIG. 1, which specifically may be used
To be applied in various electronic equipments.
As shown in figure 5, the device 500 for monitoring driving behavior of the present embodiment includes: acquiring unit 501, monitoring list
Member 502 and model construction unit 503.
Acquiring unit 501, for obtaining the historical trajectory data of driver behavior modeling model and target driver.
Monitoring unit 502, for being analyzed using driver behavior modeling model historical trajectory data, monitoring objective is driven
The driving behavior for the person of sailing.
Wherein, above-mentioned driver behavior modeling model is obtained by passing through model construction unit 503, above-mentioned model construction unit
503 include: to obtain module 5031, processing module 5032, mark module 5033 and determining module 5034.
Wherein, module 5031 is obtained, for obtaining the historical trajectory data set and running cost data of multiple drivers
Set.
Processing module 5032 determines each for handling historical trajectory data set and running cost data acquisition system respectively
The driving behavior parameter value and running cost parameter value of driver.
Mark module 5033, for the driver by running cost parameter value less than the first preset value labeled as the first mark
The driver that running cost parameter value is greater than the second preset value is labeled as the second label by label.
Determining module 5034, for learning the driving of multiple drivers indicated by the first label using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by behavioral parameters value and the second label, extraction meet driving for preset condition
Behavioral parameters are sailed as characteristic parameter, and driver behavior modeling model is determined based on characteristic parameter.
In some optional implementations of the present embodiment, above-mentioned historical trajectory data includes the position of multiple tracing points
And the moment.Above-mentioned processing module 5032 can be further used for: determine the tracing point that any two are adjacent in historical trajectory data
Between displacement and duration;According to displacement and duration, determine each driver in the driving behavior parameter value of each tracing point, it is above-mentioned
Driving behavior parameter includes: speed and traveling angle.
In some optional implementations of the present embodiment, above-mentioned processing module 5032 can also be further used for: root
According to each driver in the velocity amplitude of each tracing point, the velocity distribution curve of each driver is determined;Determine above-mentioned VELOCITY DISTRIBUTION
Velocity amplitude is greater than the number of preset friction speed threshold value and continues every time greater than preset friction speed threshold value in curve
Duration;According to velocity distribution curve, the cumulative distribution function of speed is determined;It determines in the cumulative distribution function of above-mentioned speed and presets
Different probability be worth corresponding velocity amplitude.
In some optional implementations of the present embodiment, above-mentioned driving behavior parameter further include: acceleration.Above-mentioned place
Reason module 5032 can also be further used for: according to the velocity distribution curve of each driver, determine the acceleration of each tracing point
Degree;Determine the acceleration profile curve of each driver;Determine that acceleration value is greater than preset difference in acceleration profile curve
The number of acceleration rate threshold and the first total duration greater than preset different acceleration rate thresholds;It determines in acceleration profile curve
Acceleration value is less than zero and absolute value is greater than the number of preset different acceleration rate thresholds and absolute value is greater than different acceleration
Second total duration of threshold value;According to the acceleration profile curve of each driver, the cumulative distribution function of acceleration is determined;It determines
Preset different probability is worth corresponding acceleration value in the cumulative distribution function of acceleration.
In some optional implementations of the present embodiment, above-mentioned driving behavior parameter further include: continuously drive duration.
Above-mentioned processing module 5032 can also be further used for: for each driver, according between the adjacent tracing point of any two
Displacement and duration, determine that above-mentioned displacement is less than two adjacent track points that preset displacement and above-mentioned duration are less than preset duration
Moment;According at the time of identified tracing point and in historical trajectory data at the time of first tracing point, the last one rail
At the time of mark point, at least one continuous driving duration of each driver is determined.
In some optional implementations of the present embodiment, above-mentioned processing module 5032 can also be further used for: right
In each driver, according to the traveling angle of each tracing point, determine that traveling angle is located within the scope of preset different angle
Velocity distribution curve;Determine the corresponding cumulative distribution function of each velocity distribution curve;It determines in different cumulative distribution function not
The corresponding velocity amplitude with probability.
In some optional implementations of the present embodiment, above-mentioned running cost parameter includes: the oil consumption of unit distance
The maintenance cost value of value and unit distance.Above-mentioned mark module 5033 can also be further used for: by the fuel consumption values of unit distance
And the maintenance cost value of unit distance is respectively less than the driver of the first preset value labeled as the first label;By the oil consumption of unit distance
The maintenance cost value of value and unit distance is all larger than the driver of the second preset value labeled as the second label.
In some optional implementations of the present embodiment, above-mentioned first label is third preset value, above-mentioned second mark
Label are the 4th preset value.Above-mentioned determining module 5034 can also be further used for: be marked using machine learning algorithm study by first
The driving behavior of multiple drivers indicated by the driving behavior parameter value and the second label of the indicated multiple drivers of label
Parameter value determines the P value of each driving behavior parameter;Calculate each driving behavior parameter value and third preset value or the 4th preset value
Between Pearson correlation coefficients;It extracts P value and is greater than the 5th preset value and Pearson correlation coefficients driving greater than the 6th preset value
Behavioral parameters are sailed as characteristic parameter.
In some optional implementations of the present embodiment, above-mentioned determining module 5034 can also be further used for: base
In the P value and Pearson correlation coefficients of each characteristic parameter, the weight of each characteristic parameter is determined;Based on each characteristic parameter and corresponding
Weight obtains above-mentioned driver behavior modeling model.
The application's is provided by the above embodiment for monitoring the device of driving behavior, by processing module to acquisition module
The historical trajectory data set and running cost parameter sets of the multiple drivers obtained is respectively processed, and obtains each driving
The driving behavior parameter value and running cost parameter value of member, then mark module is symbol according to the running cost parameter of each driver
The driver for closing different condition marks different labels, and determining module recycling machine learning algorithm learns the driving of different labels
The driving behavior parameter value of member is extracted the driving behavior parameter for meeting preset condition as characteristic parameter, is joined based on features described above
Number determines driver behavior modeling model, the target that monitoring unit recycles above-mentioned driver behavior modeling model to obtain acquiring unit
The historical trajectory data of driver is analyzed, and realizes the monitoring to target driver driving behavior.Above example implements
Abundant excavation to the historical trajectory data of driver carries out quantization monitoring to the driving behavior of driver.
It should be appreciated that for monitor the unit 501 recorded in the device 500 of driving behavior to unit 502 respectively with reference
Each step in method described in Fig. 1 is corresponding, and module 5031 to module 5034 is respectively and in method described in Fig. 1
Each step is corresponding.It is equally applicable to as a result, above with respect to the operation and feature of the method description for monitoring driving behavior
Device 500 and unit wherein included, details are not described herein.The corresponding units of device 500 can in server or terminal
Unit cooperates to realize the scheme of the embodiment of the present application.
Below with reference to Fig. 6, it illustrates the calculating of the terminal device or server that are suitable for being used to realize the embodiment of the present application
The structural schematic diagram of machine system 600.
As shown in fig. 6, computer system 600 includes central processing unit (CPU) 601, it can be read-only according to being stored in
Program in memory (ROM) 602 or be loaded into the program in random access storage device (RAM) 603 from storage section 608 and
Execute various movements appropriate and processing.In RAM 603, also it is stored with system 600 and operates required various programs and data.
CPU 601, ROM 602 and RAM 603 are connected with each other by bus 604.Input/output (I/O) interface 605 is also connected to always
Line 604.
I/O interface 605 is connected to lower component: the importation 606 including keyboard, mouse etc.;It is penetrated including such as cathode
The output par, c 607 of spool (CRT), liquid crystal display (LCD) etc. and loudspeaker etc.;Storage section 608 including hard disk etc.;
And the communications portion 609 of the network interface card including LAN card, modem etc..Communications portion 609 via such as because
The network of spy's net executes communication process.Driver 610 is also connected to I/O interface 605 as needed.Detachable media 611, such as
Disk, CD, magneto-optic disk, semiconductor memory etc. are mounted on as needed on driver 610, in order to read from thereon
Computer program be mounted into storage section 608 as needed.
Particularly, in accordance with an embodiment of the present disclosure, it may be implemented as computer above with reference to the process of flow chart description
Software program.For example, embodiment of the disclosure includes a kind of computer program product comprising be tangibly embodied in machine readable
Computer program on medium, above-mentioned computer program include the program code for method shown in execution flow chart.At this
In the embodiment of sample, which can be downloaded and installed from network by communications portion 609, and/or from removable
Medium 611 is unloaded to be mounted.When the computer program is executed by central processing unit (CPU) 601, execute in the present processes
The above-mentioned function of limiting.
Flow chart and block diagram in attached drawing are illustrated according to the system of the various embodiments of the application, method and computer journey
The architecture, function and operation in the cards of sequence product.In this regard, each box in flowchart or block diagram can generation
A part of one module, program segment or code of table, a part of above-mentioned module, program segment or code include one or more
Executable instruction for implementing the specified logical function.It should also be noted that in some implementations as replacements, institute in box
The function of mark can also occur in a different order than that indicated in the drawings.For example, two boxes succeedingly indicated are practical
On can be basically executed in parallel, they can also be executed in the opposite order sometimes, and this depends on the function involved.Also it wants
It is noted that the combination of each box in block diagram and or flow chart and the box in block diagram and or flow chart, Ke Yiyong
The dedicated hardware based system of defined functions or operations is executed to realize, or can be referred to specialized hardware and computer
The combination of order is realized.
Being described in unit involved in the embodiment of the present application can be realized by way of software, can also be by hard
The mode of part is realized.Described unit also can be set in the processor, for example, can be described as: a kind of processor packet
Include acquiring unit, monitoring unit and model construction unit.Wherein, the title of these units is not constituted to this under certain conditions
The restriction of unit itself, for example, acquiring unit is also described as " obtaining driver behavior modeling model and target driver
Historical trajectory data unit ".
As on the other hand, present invention also provides a kind of nonvolatile computer storage media, the non-volatile calculating
Machine storage medium can be nonvolatile computer storage media included in device described in above-described embodiment;It can also be with
It is individualism, without the nonvolatile computer storage media in supplying terminal.Above-mentioned nonvolatile computer storage media
It is stored with one or more program, when said one or multiple programs are executed by an equipment, so that the equipment: obtaining
Take the historical trajectory data of driver behavior modeling model and target driver;Using driver behavior modeling model to historical track
Data are analyzed, the driving behavior of monitoring objective driver;Wherein, above-mentioned driver behavior modeling model through the following steps that
It obtains: obtaining the historical trajectory data set and running cost data acquisition system of multiple drivers;Historical track number is handled respectively
According to set and running cost data acquisition system, the driving behavior parameter value and running cost parameter value of each driver are determined;It will go
Driver of the cost parameter value less than the first preset value is sailed labeled as the first label, running cost parameter value is greater than second and is preset
The driver of value is labeled as the second label;Learn the driving of multiple drivers indicated by the first label using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by behavioral parameters value and the second label, extraction meet driving for preset condition
Behavioral parameters are sailed as characteristic parameter, and driver behavior modeling model is determined based on features described above parameter.
Above description is only the preferred embodiment of the application and the explanation to institute's application technology principle.Those skilled in the art
Member is it should be appreciated that invention scope involved in the application, however it is not limited to technology made of the specific combination of above-mentioned technical characteristic
Scheme, while should also cover in the case where not departing from foregoing invention design, it is carried out by above-mentioned technical characteristic or its equivalent feature
Any combination and the other technical solutions formed.Such as features described above has similar function with (but being not limited to) disclosed herein
Can technical characteristic replaced mutually and the technical solution that is formed.
Claims (18)
1. a kind of method for monitoring driving behavior, which is characterized in that the described method includes:
Obtain the historical trajectory data of driver behavior modeling model and target driver;
The historical trajectory data is analyzed using the driver behavior modeling model, monitors driving for the target driver
Sail behavior;
Wherein, the driver behavior modeling model is obtained through the following steps:
Obtain the historical trajectory data set and running cost data acquisition system of multiple drivers;
The historical trajectory data set and the running cost data acquisition system are handled respectively, determine the driving row of each driver
For parameter value and running cost parameter value;
Driver by the running cost parameter value less than the first preset value is labeled as the first label, and the running cost is joined
The driver that numerical value is greater than the second preset value is labeled as the second label;
Learn driving behavior parameter value and the institute of multiple drivers indicated by first label using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by the second label is stated, the driving behavior parameter work for meeting preset condition is extracted
It is characterized parameter, and the driver behavior modeling model is determined based on the characteristic parameter.
2. the method according to claim 1, wherein the historical trajectory data includes the position of multiple tracing points
And the moment;And
It is described to handle the historical trajectory data set and the running cost data acquisition system respectively, determine driving for each driver
Sail behavioral parameters value and running cost parameter value, comprising:
Determine the displacement and duration between any two are adjacent in the historical trajectory data tracing point;
According to the displacement and the duration, determine each driver in the driving behavior parameter value of each tracing point, the driving
Behavioral parameters include: speed and traveling angle.
3. according to the method described in claim 2, it is characterized in that, between the tracing point adjacent according to any two
Displacement and duration, determine each driver in the driving behavior parameter value of each tracing point, comprising:
According to each driver in the velocity amplitude of each tracing point, the velocity distribution curve of each driver is determined;
Determine that velocity amplitude is greater than the number of preset friction speed threshold value and is greater than every time described in the velocity distribution curve
The duration of preset friction speed threshold value;
According to the velocity distribution curve, the cumulative distribution function of speed is determined;
Determine the corresponding velocity amplitude of preset different probability value in the cumulative distribution function of the speed.
4. according to the method described in claim 3, it is characterized in that, the driving behavior parameter further include: acceleration;And
Displacement and duration between the tracing point adjacent according to any two determine each driver in each tracing point
Driving behavior parameter value, comprising:
According to the velocity distribution curve of each driver, the acceleration of each tracing point is determined;
Determine the acceleration profile curve of each driver;
Determine that acceleration value is greater than the number of preset different acceleration rate thresholds and is greater than institute in the acceleration profile curve
State the first total duration of preset different acceleration rate thresholds;
Determine that acceleration value is less than zero in the acceleration profile curve and absolute value is greater than preset different acceleration rate thresholds
Number and the absolute value are greater than the second total duration of the different acceleration rate thresholds;
According to the acceleration profile curve of each driver, the cumulative distribution function of acceleration is determined;
Determine the corresponding acceleration value of preset different probability value in the cumulative distribution function of the acceleration.
5. according to the method described in claim 2, it is characterized in that, the driving behavior parameter includes continuous driving duration;With
And
Displacement and duration between the tracing point adjacent according to any two determine each driver in each tracing point
Driving behavior parameter value, comprising:
The displacement is determined according to the displacement and duration between the adjacent tracing point of any two for each driver
Less than preset displacement and at the time of the duration is less than two adjacent track points of preset duration;
According at the time of identified tracing point and in the historical trajectory data at the time of first tracing point, the last one
At the time of tracing point, at least one continuous driving duration of each driver is determined.
6. according to the method described in claim 2, it is characterized in that, between the tracing point adjacent according to any two
Displacement and duration, determine each driver in the driving behavior parameter value of each tracing point, comprising:
For each driver, according to the traveling angle of each tracing point, determine that the traveling angle is located at preset different angles
Spend the velocity distribution curve in range;
Determine the corresponding cumulative distribution function of each velocity distribution curve;
Determine the corresponding velocity amplitude of different probability in the different cumulative distribution function.
7. method according to claim 1-6, which is characterized in that the running cost parameter include: unit away from
From fuel consumption values and unit distance maintenance cost value;And
The driver by the running cost parameter less than the first preset value is labeled as the first label, by the running cost
The driver that parameter is greater than the second preset value is labeled as the second label, comprising:
The maintenance cost value of the fuel consumption values of the unit distance and the unit distance is respectively less than to the driver of the first preset value
Labeled as the first label;
The maintenance cost value of the fuel consumption values of the unit distance and the unit distance is all larger than to the driver of the second preset value
Labeled as the second label.
8. method according to claim 1-6, which is characterized in that first label is third preset value, institute
Stating the second label is the 4th preset value;And
It is described using machine learning algorithm learn the driving behavior parameter value of multiple drivers indicated by first label with
And the driving behavior parameter value of multiple drivers indicated by second label, extract the driving behavior ginseng for meeting preset condition
Number is used as characteristic parameter, comprising:
Using machine learning algorithm study by the driving behavior parameter value of multiple drivers indicated by first label and
The driving behavior parameter value of multiple drivers indicated by second label determines the P value of each driving behavior parameter;
Calculate the Pearson correlation coefficients between each driving behavior parameter value and the third preset value or the 4th preset value;
Extract the driving behavior parameter that the P value is greater than the 6th preset value greater than the 5th preset value and the Pearson correlation coefficients
As characteristic parameter.
9. according to the method described in claim 8, it is characterized in that, described determine the driving behavior based on the characteristic parameter
Monitoring model, comprising:
P value and Pearson correlation coefficients based on each characteristic parameter, determine the weight of each characteristic parameter;
Based on each characteristic parameter and corresponding weight, the driver behavior modeling model is obtained.
10. a kind of for monitoring the device of driving behavior, which is characterized in that described device includes:
Acquiring unit, for obtaining the historical trajectory data of driver behavior modeling model and target driver;
Monitoring unit, for being analyzed using the driver behavior modeling model the historical trajectory data, described in monitoring
The driving behavior of target driver;
Wherein, the driver behavior modeling model is obtained by model construction unit, and the model construction unit includes:
Module is obtained, for obtaining the historical trajectory data set and running cost data acquisition system of multiple drivers;
Processing module determines each for handling the historical trajectory data set and the running cost data acquisition system respectively
The driving behavior parameter value and running cost parameter value of driver;
Mark module is labeled as the first label for the driver by the running cost parameter value less than the first preset value, will
The driver that the running cost parameter value is greater than the second preset value is labeled as the second label;
Determining module, for learning the driving behavior of multiple drivers indicated by first label using machine learning algorithm
The driving behavior parameter value of multiple drivers indicated by parameter value and second label, extraction meet driving for preset condition
Behavioral parameters are sailed as characteristic parameter, and the driver behavior modeling model is determined based on the characteristic parameter.
11. device according to claim 10, which is characterized in that the historical trajectory data includes the position of multiple tracing points
It sets and the moment;And
The processing module is further used for:
Determine the displacement and duration between any two are adjacent in the historical trajectory data tracing point;
According to the displacement and the duration, determine each driver in the driving behavior parameter value of each tracing point, the driving
Behavioral parameters include: speed and traveling angle.
12. device according to claim 11, which is characterized in that the processing module is further used for:
According to each driver in the velocity amplitude of each tracing point, the velocity distribution curve of each driver is determined;
Determine that velocity amplitude is greater than the number of preset friction speed threshold value and is greater than every time described in the velocity distribution curve
The duration of preset friction speed threshold value;
According to the velocity distribution curve, the cumulative distribution function of speed is determined;
Determine the corresponding velocity amplitude of preset different probability value in the cumulative distribution function of the speed.
13. device according to claim 12, which is characterized in that the driving behavior parameter further include: acceleration;And
The processing module is further used for:
According to the velocity distribution curve of each driver, the acceleration of each tracing point is determined;
Determine the acceleration profile curve of each driver;
Determine that acceleration value is greater than the number of preset different acceleration rate thresholds and is greater than institute in the acceleration profile curve
State the first total duration of preset different acceleration rate thresholds;
Determine that acceleration value is less than zero in the acceleration profile curve and absolute value is greater than preset different acceleration rate thresholds
Number and the absolute value are greater than the second total duration of the different acceleration rate thresholds;
According to the acceleration profile curve of each driver, the cumulative distribution function of acceleration is determined;
Determine the corresponding acceleration value of preset different probability value in the cumulative distribution function of the acceleration.
14. device according to claim 11, which is characterized in that the driving behavior parameter includes: continuous driving duration;
And
The processing module is further used for:
The displacement is determined according to the displacement and duration between the adjacent tracing point of any two for each driver
Less than preset displacement and at the time of the duration is less than two adjacent track points of preset duration;
According at the time of identified tracing point and in the historical trajectory data at the time of first tracing point, the last one
At the time of tracing point, at least one continuous driving duration of each driver is determined.
15. device according to claim 11, which is characterized in that the processing module is further used for:
For each driver, according to the traveling angle of each tracing point, determine that the traveling angle is located at preset different angles
Spend the velocity distribution curve in range;
Determine the corresponding cumulative distribution function of each velocity distribution curve;
Determine the corresponding velocity amplitude of different probability in the different cumulative distribution function.
16. the described in any item devices of 0-15 according to claim 1, which is characterized in that the running cost parameter includes: unit
The fuel consumption values of distance and the maintenance cost value of unit distance;And
The mark module is further used for:
The maintenance cost value of the fuel consumption values of the unit distance and the unit distance is respectively less than to the driver of the first preset value
Labeled as the first label;
The maintenance cost value of the fuel consumption values of the unit distance and the unit distance is all larger than to the driver of the second preset value
Labeled as the second label.
17. the described in any item devices of 0-15 according to claim 1, which is characterized in that first label is third preset value,
Second label is the 4th preset value;And
The determining module is further used for:
Using machine learning algorithm study by the driving behavior parameter value of multiple drivers indicated by first label and
The driving behavior parameter value of multiple drivers indicated by second label determines the P value of each driving behavior parameter;
Calculate the Pearson correlation coefficients between each driving behavior parameter value and the third preset value or the 4th preset value;
Extract the driving behavior parameter that the P value is greater than the 6th preset value greater than the 5th preset value and the Pearson correlation coefficients
As characteristic parameter.
18. device according to claim 17, which is characterized in that the determining module is further used for:
P value and Pearson correlation coefficients based on each characteristic parameter, determine the weight of each characteristic parameter;
Based on each characteristic parameter and corresponding weight, the driver behavior modeling model is obtained.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710046084.6A CN106781503B (en) | 2017-01-22 | 2017-01-22 | Method and apparatus for monitoring driving behavior |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201710046084.6A CN106781503B (en) | 2017-01-22 | 2017-01-22 | Method and apparatus for monitoring driving behavior |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106781503A CN106781503A (en) | 2017-05-31 |
CN106781503B true CN106781503B (en) | 2019-04-05 |
Family
ID=58945223
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201710046084.6A Active CN106781503B (en) | 2017-01-22 | 2017-01-22 | Method and apparatus for monitoring driving behavior |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN106781503B (en) |
Families Citing this family (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107657812B (en) * | 2017-08-29 | 2021-10-29 | 耀灵人工智能(浙江)有限公司 | Method for dynamically planning parking space according to traffic flow prediction |
CN109543497A (en) * | 2017-09-20 | 2019-03-29 | 顾泽苍 | A kind of construction method of more purposes control machine learning model suitable for automatic Pilot |
CN108806021B (en) * | 2018-06-12 | 2020-11-06 | 重庆大学 | Electric vehicle target road section energy consumption prediction method based on physical model and road characteristic parameters |
CN109151773B (en) * | 2018-07-17 | 2021-10-19 | Oppo广东移动通信有限公司 | Information prompting method and related product |
CN109102194B (en) * | 2018-08-20 | 2021-06-18 | 安徽佳略信息科技有限公司 | Driving behavior scoring method based on global positioning system and inertial sensor |
CN109493448A (en) * | 2018-09-26 | 2019-03-19 | 平安科技(深圳)有限公司 | Driving recording data processing method, device, computer equipment and storage medium |
JP7169831B2 (en) * | 2018-09-27 | 2022-11-11 | 株式会社Subaru | MOBILE OBJECT MONITORING DEVICE, VEHICLE CONTROL SYSTEM AND TRAFFIC SYSTEM USING THE SAME |
CN111489460B (en) * | 2019-01-28 | 2022-09-23 | 北京嘀嘀无限科技发展有限公司 | Travel data processing method, travel data processing device, navigation equipment and computer storage medium |
CN109919293B (en) * | 2019-02-22 | 2021-02-26 | 深兰科技(上海)有限公司 | Dangerous driving judgment method and device |
CN110288826B (en) * | 2019-05-24 | 2020-11-13 | 浙江工业大学 | Traffic control subregion clustering division method based on multi-source data fusion and MILP |
CN112319489B (en) * | 2020-11-18 | 2022-03-04 | 三一重型装备有限公司 | Driving behavior monitoring method, driving behavior monitoring system, server and storage medium |
CN113762755A (en) * | 2021-08-30 | 2021-12-07 | 一汽解放汽车有限公司 | Method and device for pushing driver analysis report, computer equipment and storage medium |
CN115662184B (en) * | 2022-09-09 | 2023-11-24 | 湖南大学 | Vehicle driving risk assessment method |
CN116749989B (en) * | 2023-07-24 | 2023-11-07 | 吉林大学 | Method, device, equipment and storage medium for identifying turning driving habit of driver |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103456167A (en) * | 2013-09-17 | 2013-12-18 | 重庆大学 | Good driving skill parameter obtaining method based on critical areas |
EP3073450A1 (en) * | 2015-03-27 | 2016-09-28 | Tata Consultancy Services Limited | System and method for monitoring driving behavior of a driver |
CN106056162A (en) * | 2016-06-07 | 2016-10-26 | 浙江大学 | A traffic safety credit scoring method based on GPS track and traffic law-violation records |
CN106203735A (en) * | 2016-07-27 | 2016-12-07 | 北京工业大学 | A kind of automobile driver driving behavior energy consumption characters measuring method |
-
2017
- 2017-01-22 CN CN201710046084.6A patent/CN106781503B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN103456167A (en) * | 2013-09-17 | 2013-12-18 | 重庆大学 | Good driving skill parameter obtaining method based on critical areas |
EP3073450A1 (en) * | 2015-03-27 | 2016-09-28 | Tata Consultancy Services Limited | System and method for monitoring driving behavior of a driver |
CN106056162A (en) * | 2016-06-07 | 2016-10-26 | 浙江大学 | A traffic safety credit scoring method based on GPS track and traffic law-violation records |
CN106203735A (en) * | 2016-07-27 | 2016-12-07 | 北京工业大学 | A kind of automobile driver driving behavior energy consumption characters measuring method |
Also Published As
Publication number | Publication date |
---|---|
CN106781503A (en) | 2017-05-31 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106781503B (en) | Method and apparatus for monitoring driving behavior | |
CN103364006B (en) | For determining the system and method for vehicle route | |
CN112465395A (en) | Multi-dimensional comprehensive evaluation method and device for automatically-driven automobile | |
CN108074396A (en) | The evaluation method that drives safely and system | |
US20190310649A1 (en) | System and method for a framework of robust and safe reinforcement learning application in real world autonomous vehicle application | |
CN106347359A (en) | Method and device for operating autonomous vehicle | |
CN106314438A (en) | Method and system for detecting abnormal track in driver driving track | |
CN107665520A (en) | A kind of driving monitoring method and device based on vehicle operation data | |
CN105527110B (en) | The appraisal procedure and device of automobile fuel ecomomy | |
CN109993966A (en) | A kind of method and device of building user portrait | |
Singh et al. | Deep learning in the automotive industry: Recent advances and application examples | |
US20230251665A1 (en) | Systems and methods for evaluating autonomous vehicle software interactions for proposed trips | |
CN112508164A (en) | End-to-end automatic driving model pre-training method based on asynchronous supervised learning | |
Wu et al. | Clustering of several typical behavioral characteristics of commercial vehicle drivers based on GPS data mining: Case study of highways in China | |
Gatteschi et al. | Comparing algorithms for aggressive driving event detection based on vehicle motion data | |
LU100760B1 (en) | Vehicular motion assessment method | |
CN105719359B (en) | The acquisition methods and device of information of vehicles | |
CN114492544B (en) | Model training method and device and traffic incident occurrence probability evaluation method and device | |
CN114495421B (en) | Intelligent open type road construction operation monitoring and early warning method and system | |
CN114920122A (en) | Escalator fault early warning method and system based on transfer learning | |
Wijayasekara et al. | Data driven fuel efficient driving behavior feedback for fleet vehicles | |
SB et al. | Real-time pothole detection using YOLOv5 algorithm: a feasible approach for intelligent transportation systems | |
WO2019034233A1 (en) | System and method for vehicle modelling and simulation | |
Ma et al. | A Scoping Review of Energy-Efficient Driving Behaviors and Applied State-of-the-Art AI Methods | |
Sun | A method for judging abnormal driving behaviors in diversion areas based on group intelligence perception of internet of vehicles. |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |