CN107909678A - One kind driving risk evaluating method and system - Google Patents
One kind driving risk evaluating method and system Download PDFInfo
- Publication number
- CN107909678A CN107909678A CN201711222806.5A CN201711222806A CN107909678A CN 107909678 A CN107909678 A CN 107909678A CN 201711222806 A CN201711222806 A CN 201711222806A CN 107909678 A CN107909678 A CN 107909678A
- Authority
- CN
- China
- Prior art keywords
- information
- driving
- vehicle
- acceleration
- axis
- 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.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0841—Registering performance data
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N20/00—Machine learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0635—Risk analysis of enterprise or organisation activities
-
- G—PHYSICS
- G07—CHECKING-DEVICES
- G07C—TIME OR ATTENDANCE REGISTERS; REGISTERING OR INDICATING THE WORKING OF MACHINES; GENERATING RANDOM NUMBERS; VOTING OR LOTTERY APPARATUS; ARRANGEMENTS, SYSTEMS OR APPARATUS FOR CHECKING NOT PROVIDED FOR ELSEWHERE
- G07C5/00—Registering or indicating the working of vehicles
- G07C5/08—Registering or indicating performance data other than driving, working, idle, or waiting time, with or without registering driving, working, idle or waiting time
- G07C5/0808—Diagnosing performance data
Abstract
The invention discloses one kind driving risk evaluating method and system, it is related to traffic risk evaluation areas.This method includes:The driving information of collection vehicle;Driving information is handled, obtains the abnormal driving information of vehicle;The accident information occurred is obtained, the machine learning model for being used for evaluating driving risk is established according to abnormal driving information and accident information;Evaluated according to driving risk of the machine learning model to vehicle.A kind of driving risk evaluating method provided by the invention and system, it can obtain the driving risk evaluation result with very pinpoint accuracy, risk assessment objectively accurately can be carried out to evaluated vehicle, meet the demand that becomes more meticulous to risk assessment, timely driver instructor correction bad steering behavior, reduces accident risk.
Description
Technical field
The present invention relates to traffic risk evaluation areas, more particularly to a kind of driving risk evaluating method and system.
Background technology
At present, all it is using static data as feature evaluation risk, these features are past for the appraisal procedure for risk of driving a vehicle
It is past simply relatively low with accident indirect correlation, correlation.For example, currently with driving behavior assessment risk method in it is most important
Reference index is distance travelled, its core concept is that the distance travelled of user is higher, and risk of being in danger is higher.
But rely on distance travelled and determine driving risk, have the shortcomings that assessment is inaccurate.For example, when the row of two drivers
Sail mileage it is identical when, according to existing methods of risk assessment, the driving risk of two people should be identical, but actually proves and differ
Fixed correct, existing appraisal procedure accurately can not objectively react driving risk, and only relying on distance travelled cannot meet to work as
The preceding demand that becomes more meticulous to risk assessment.
The content of the invention
The technical problems to be solved by the invention are in view of the deficiencies of the prior art, there is provided one kind driving risk evaluating method
And system.
The technical solution that the present invention solves above-mentioned technical problem is as follows:
One kind driving risk evaluating method, including:
The driving information of collection vehicle;
The driving information is handled, obtains the abnormal driving information of the vehicle;
The accident information occurred is obtained, is established according to the abnormal driving information and the accident information and is used to evaluate row
The machine learning model of car risk;
Evaluated according to driving risk of the machine learning model to the vehicle.
The beneficial effects of the invention are as follows:A kind of driving risk evaluating method provided by the invention, is expert at by collection vehicle
Various abnormal driving information during sailing, and combine the accident information occurred and be modeled, believed by the accident occurred
Cease and the abnormal driving information of vehicle is matched and analyzed, the driving risk assessment knot with very pinpoint accuracy can be obtained
Fruit, accurately objectively can carry out risk assessment to evaluated vehicle, meet the demand that becomes more meticulous to risk assessment, instruct in time
Driver corrects bad steering behavior, reduces accident risk.
Based on the above technical solutions, the present invention can also be improved as follows.
Further, the driving information includes:The velocity information and acceleration information of vehicle, the row of the collection vehicle
Information is sailed to specifically include:
Obtain the velocity information in vehicle travel process in real time by OBD equipment;
The acceleration information of all directions in the vehicle travel process is obtained by three axis accelerometer, wherein, with described
The center of vehicle is origin, and headstock direction is forward direction, and car door direction is lateral.
It is using the above-mentioned further beneficial effect of scheme:During OBD equipment in real time collection vehicle traveling
Velocity information, since its source is the ECU of vehicle, has the advantages that data accurately and reliably and frequency acquisition is easy to control.And
The fatigue driving event continuously driven for a long time can be identified by picking rate data time series.
Further, the acceleration information that all directions in the vehicle travel process are obtained by three axis accelerometer
Before, further include:
The spin matrix for being used for correcting the three axis accelerometer is calculated, three axis is accelerated according to the spin matrix
The x, y, z axis of degree meter carries out rotation processing, obtains x ', y ', z ' axis;
The x ', y ', z ' axis are calibrated respectively, and the forward direction using the x ' axis after calibration and y ' axis as the vehicle
With it is lateral.
It is using the above-mentioned further beneficial effect of scheme:By during the form of vehicle, to three axis accelerometer
Each axis be modified, can be when subsequent acquisition acceleration information, according to the information of orientation by its projective transformation to car
On three axis of itself, ensure vehicle traveling when on acclive inclined-plane, actual acquisition acceleration direction, numerical value are still accurate
Really.
Further, the abnormal driving information of the vehicle includes:Driving range information, fatigue driving information, anxious acceleration
Information, anxious deceleration, zig zag information and sudden turn of events road information, it is described that the driving information is handled, obtain the car
Abnormal driving information specifically include:
Integral Processing is carried out to the velocity information, obtains the driving range information;
According to the velocity information and the Velocity Time sequence of the velocity information, the fatigue driving information is obtained;
According to the acceleration information of vehicle forward direction, the peak of the travel direction acceleration time series of the vehicle is identified
Value and valley, the anxious acceleration information and the anxious deceleration are obtained according to the peak value and the valley;
According to the lateral acceleration information of the vehicle, peak value and the paddy of the vehicle side acceleration time series are identified
Value, obtains the zig zag information, and obtain sudden turn of events road information with reference to the time series according to the peak value and valley.
It is using the above-mentioned further beneficial effect of scheme:Driving range is obtained by carrying out Integral Processing to velocity information
Information, it is possible to increase the precision of driving range information.All it is with GPS positioning technology due to traditional driving range computational methods
To estimate driving range, driving locus is formed using the discrete location longitude and latitude degrees of data sampled in vehicle travel process, most
Path length is calculated afterwards obtains distance travelled.This method is restricted be subject to several respects:If the frequency of sampling is too low, track estimation
With regard to inaccuracy;Individual lot satellite-signal is bad so that position loses or drifts about, and influences track estimation.And set by OBD
The standby velocity information collected calculates driving range, have the advantages that data accurately, strong interference immunity, in discrete sampling speed
The complicated calculations such as conversion, track identification relatively simple, without being related to GPS are realized in equipment, systematic error is smaller.
And each axle acceleration value collected by three axis accelerometer, with reference to the change of instantaneous G values, can accurately identify
Every abnormal driving event, the method such as anxious acceleration, anxious deceleration, zig zag, sudden turn of events road are simply easily achieved, round-the-clock can supervise
Control.
Further, described established according to the abnormal driving information and the accident information is used to evaluate driving risk
After machine learning model, further include:
By cross validation method, training is updated to the machine learning model.
The another technical solution that the present invention solves above-mentioned technical problem is as follows:
One kind driving Risk Evaluating System, including:
Information collecting device, the driving information for collection vehicle;
Processor, for handling the driving information, obtains the abnormal driving information of the vehicle;
Server, for obtaining the accident information occurred, builds according to the abnormal driving information and the accident information
Stand for evaluate driving risk machine learning model, and according to driving risk of the machine learning model to the vehicle into
Row evaluation.
Further, the driving information includes:The velocity information and acceleration information of vehicle, described information harvester
Including:
OBD equipment, for obtaining the velocity information in vehicle travel process in real time;
Three axis accelerometer, for obtaining the acceleration information of all directions in the vehicle travel process, wherein, with described
The center of vehicle is origin, and headstock direction is forward direction, and car door direction is lateral.
Further, the processor includes:
Spin matrix computing unit, for calculating the spin matrix for being used for correcting the three axis accelerometer, according to described
Spin matrix carries out rotation processing to the x, y, z axis of the three axis accelerometer, obtains x ', y ', z ' axis;
Reference axis corrects unit, for being calibrated respectively to the x ', y ', z ' axis, and by the x ' axis and y ' after calibration
Axis is as the positive and lateral of the vehicle.
Further, the abnormal driving information of the vehicle includes:Driving range information, fatigue driving information, anxious acceleration
Information, anxious deceleration, zig zag information and sudden turn of events road information, the processor further include:
Driving range computing unit, for carrying out Integral Processing to the velocity information, obtains the driving range information;
Fatigue driving judging unit, for the Velocity Time sequence according to the velocity information and the velocity information,
Obtain the fatigue driving information;
It is anxious to accelerate anxious deceleration judging unit, for the acceleration information according to vehicle forward direction, identify the vehicle
The peak value and valley of travel direction acceleration time series, according to the peak value and the valley obtain it is described it is anxious accelerate information and
The urgency deceleration;
Take a sudden turn sudden turn of events road judging unit, for according to the lateral acceleration information of the vehicle, identifying the vehicle side
To the peak value and valley of acceleration time series, the zig zag information is obtained according to the peak value and valley, and with reference to described
Time series obtains sudden turn of events road information.
Further, the server is additionally operable to by cross validation method, and the machine learning model is updated
Training.
The beneficial effects of the invention are as follows:A kind of driving Risk Evaluating System provided by the invention, passes through OBD equipment and three axis
The accelerometer various abnormal driving information of collection vehicle in the process of moving in real time, and combine the accident information that has occurred into
Row modeling, is matched and is analyzed to the abnormal driving information of vehicle by the accident information occurred, can obtain having very
The driving risk evaluation result of pinpoint accuracy, accurately objectively can carry out risk assessment to evaluated vehicle, meet to risk
The demand that becomes more meticulous of assessment, timely driver instructor correct bad steering behavior, reduce accident risk.
The advantages of aspect that the present invention adds, will be set forth in part in the description, and will partly become from the following description
Obtain substantially, or recognized by present invention practice.
Brief description of the drawings
Fig. 1 is a kind of flow diagram of risk evaluating method of driving a vehicle provided by one embodiment of the present invention;
Fig. 2 is a kind of flow diagram for driving risk evaluating method that another embodiment of the present invention provides;
Fig. 3 is a kind of structural framing figure for driving Risk Evaluating System that another embodiment of the present invention provides.
Embodiment
The principle and features of the present invention will be described below with reference to the accompanying drawings, and the given examples are served only to explain the present invention, and
It is non-to be used to limit the scope of the present invention.
As shown in Figure 1, for a kind of flow diagram for risk evaluating method of driving a vehicle provided by one embodiment of the present invention, should
Method includes:
S1, the driving information of collection vehicle.
It should be noted that the driving information of collection refers to running time of the vehicle during certain traveling, traveling
Distance, real time speed information, speed change information, acceleration changes information, travel direction changes information etc., by gathering these
Driving information, helps comprehensively to assess driving risk.
Collection for driving information, actually there is a variety of implementations, for example, can be obtained by vehicle-mounted OBD system
These driving informations of vehicle in the process of moving, using OBD system obtain these driving informations benefit be obtain data
Promptly and accurately, and extra transformation need not be carried out to vehicle, has the advantages that cost is low.
In another example the equipment such as velocity sensor, acceleration transducer, GPS positioning device can also be set on vehicle, come
Gather these running datas.
S2, handles driving information, obtains the abnormal driving information of vehicle.
It should be noted that abnormal driving information is referred to exceeding the fatigue driving of default driving time, anxious acceleration, suddenly subtracted
The driving information such as speed, zig zag, sudden turn of events road, climbing, descending.
For example, can be by judging the travel-time information collected, when more than preset time, just obtaining this time
Drive as fatigue driving, and record the time of fatigue driving, the abnormal driving information as fatigue driving.
In another example can be by judging the velocity information collected, when in preset time, the change of speed
Change exceedes preset value, illustrates that vehicle has the anxious behavior for accelerating or suddenly slowing down, then can record the letter such as change value of speed
Breath, as the anxious abnormal driving information for accelerating or suddenly slowing down.
Driving information is handled, is that default abnormal driving information, its processing form have a variety of in order to obtain.
For example, it is simple digital filtering, gliding smoothing filtering and resampling conversion etc. can be carried out to the driving information collected
Single processing, then pass through default detection algorithm, it is possible to obtain the abnormal driving information of vehicle.
S3, obtains the accident information occurred, is established according to abnormal driving information and accident information and is used to evaluate driving wind
The machine learning model of danger.
It should be noted that in the accident information occurred, include before accident generation, the driving information of vehicle.
By counting in the accident information occurred, the average frequency that behavior occurs respectively is travelled during traveling, to different
Normal driving information is trained and learns, and establishes model.
For example, grader can be realized by SVM (Support Vector Machine, support vector machines), by sample
It is divided into two class of low-risk and excessive risk.
Preferably, the machine such as Logistics Regression, Random Forest, Ada Boost, ANN can also be passed through
Device learning model is realized.
Preferably due to which the data characteristics dimension of driving information is not very high, RBF (Radial Basis can be used
Function, Radial basis kernel function) SVM.
In another example abnormal driving information can also be learnt and be classified by regression algorithm, each driving information is obtained
Classification results.
In another example abnormal driving information can also be trained and be learnt by neural learning network, make it to specific
Vector it is sensitive.
S4, is evaluated according to driving risk of the machine learning model to vehicle.
After machine learning model is obtained, the vehicle that collects in real time can be travelled by machine learning model behavior into
Row monitoring and analysis, the driving risk to the vehicle are assessed, and the output of machine learning model is the result is that the driving wind of vehicle
Dangerous assessment result.
Preferably, driver can be prompted according to the evaluation result that evaluation obtains, prompts driver to change in time bad
Driving behavior, reduces the risk that accident occurs.
A kind of driving risk evaluating method provided in this embodiment, passes through the various exceptions of collection vehicle in the process of moving
Driving information, and combine the accident information that has occurred and be modeled, the abnormal driving by the accident information that has occurred to vehicle
Information is matched and analyzed, and can obtain the driving risk evaluation result with very pinpoint accuracy, can be accurately objectively right
Evaluated vehicle carries out risk assessment, meets the demand that becomes more meticulous to risk assessment, and timely driver instructor corrects bad steering row
To reduce accident risk.
As shown in Fig. 2, a kind of flow diagram of the driving risk evaluating method provided for another embodiment of the present invention, should
Method includes:
S1, the driving information of collection vehicle.
Preferably, driving information includes:The velocity information and acceleration information of vehicle.
Preferably, step S1 can specifically include:
S11, the velocity information in vehicle travel process is obtained by OBD equipment in real time.
For example, can be by the way that car can be obtained from the OBD equipment on vehicle with the interface of OBD equipment progress data exchange
Traveling during velocity information.
S12, the acceleration information of all directions in vehicle travel process is obtained by three axis accelerometer, wherein, with vehicle
Center be origin, headstock direction is forward direction, and car door direction is lateral.
By OBD equipment in real time collection vehicle traveling during velocity information, since its source is the ECU of vehicle,
There are data accurately and reliably and frequency acquisition is easy to control.And it can be known by picking rate data time series
Do not go out the fatigue driving event continuously driven for a long time.
Preferably, when installing accelerometer, due to cannot be guaranteed that its x, y, z axis is all positive, lateral and vertical with vehicle
Directly to coincidence, therefore, before step S12, the process calibrated to three axis accelerometer can also be included, specifically included:
Calculate the spin matrix for being used for correcting three axis accelerometer, the x, y, z according to spin matrix to three axis accelerometer
Axis carries out rotation processing, obtains x ', y ', z ' axis.
It should be noted that the x ' after rotation processing, y ', z ' axis are carried out, should be positive, lateral and vertical with vehicle in theory
Directly to coincidence, and if there is the gradient on the road surface that vehicle is stopped when static, still it cannot be guaranteed that x ', y ', z ' axis and vehicle
Positive, lateral and vertically to coincidence, therefore, it is necessary to x ', y ', z ' axis of accelerometer are further calibrated.
The acceleration magnitude arrived according to three axis accelerometer multi collect, is calculated the acceleration magnitude collected every time respectively
Resultant acceleration value in x ' y ' planes.
Obtained resultant acceleration value can be incident upon in x ' y ' planes in dots, when obtained all the points are flat in x ' y '
When being evenly distributed in four quadrants in face, the calibration to z ' axis is completed.
Then the resultant acceleration direction that the accelerometer of real-time collection vehicle in the process of moving collects is in the horizontal plane
Projection, according to the project linear of generation fitting be in line, as x ' axis, complete the calibration to x ' axis.
After the direction of x ' axis and z ' axis determines, the direction of y ' axis also determines therewith, completes calibration.
When skewness of the obtained all the points in four quadrants of x ' y ' planes, spin matrix is recalculated,
Until projected position meets preset condition.
Preferably, the fixing situation of itself can also be judged by the statistics to acceleration information.For example, it can pass through
The acceleration energy and frequency domain characteristic under transport condition are analyzed, to judge fixing situation.
Accelerometer, which fixes shakiness, will cause extra system noise to introduce, and acceleration information collection be caused necessarily dry
Disturb.In general during normal vehicle operation, the energy comparison of acceleration is small, it is concentrated mainly on low frequency portion on frequency domain
Point, only a acceleration of not worrying, anxious deceleration, zig zag etc. correspond to high frequency section.If equipment generates rolling because of fixed shakiness
It is dynamic, the obvious of acceleration energy will be caused to rise, can be embodied a concentrated reflection of in higher frequency range on frequency domain.Therefore, analysis is passed through
The feature of energy and frequency domain, can accurately judge the fixed condition of equipment.User is reminded to carry out phase when fixation is unreliable
It should check, ensure the reliable, stable of equipment, improve the accuracy of gathered data.
By during the form of vehicle, being modified, can add when subsequent acquisition to each axis of three axis accelerometer
During speed data, according to the information of orientation vehicle traveling on its projective transformation to vehicle three axis in itself, will be ensured in band
When on acclive inclined-plane, actual acquisition acceleration direction, numerical value are still accurate.
S2, handles driving information, obtains the abnormal driving information of vehicle.
Preferably, the abnormal driving information of vehicle includes:Driving range information, fatigue driving information, anxious acceleration information, urgency
Deceleration, zig zag information and sudden turn of events road letter.
Preferably, in step S2, can specifically include:
S21, carries out Integral Processing to velocity information, obtains driving range information.
For example, driving range information can be calculated by the following formula:
Mile=∫ Vdt.
S22, according to velocity information and the Velocity Time sequence of velocity information, obtains fatigue driving information.
For example, after carrying out simple digital filtering and resampling conversion to the Velocity Time sequence collected, input is set
In the finite state machine model counted, it is possible to when detecting recognition speed more than setting speed and continuously driving the time more than setting
Between driving behavior.The parameters such as state transition function and speed, time threshold by rationally setting finite state machine, it is possible to
Accurately identify fatigue driving event, and can be with real-time report to server.
S23, according to the acceleration information of vehicle forward direction, identify the travel direction acceleration time series of vehicle peak value and
Valley, anxious acceleration information and anxious deceleration are obtained according to peak value and valley.
For example, the vehicle acceleration information that accelerometer is gathered in real time, filters by simple gliding smoothing and adopts again
After sample conversion, by the local extremum (peak value/valley) for identifying vehicle y direction acceleration time series, it is possible to obtain urgency
Accelerate information and anxious deceleration.If the direction of acceleration is forward, predetermined threshold value is exceeded in the absolute value short time, it is corresponding to worry
Accelerated events;If the direction of acceleration is backward, the absolute value short time exceedes predetermined threshold value, then corresponding worried deceleration event
S24, according to the lateral acceleration information of vehicle, identifies the peak value and valley of vehicle side acceleration time series,
Zig zag information is obtained according to peak value and valley, and binding time sequence obtains sudden turn of events road information.
It should be noted that according to the lateral acceleration information of vehicle, the lateral acceleration of vehicle can be obtained, it is assumed that with
The left side of vehicle is just, then when the side acceleration values of vehicle is just and when its absolute value is more than default side acceleration values,
So it is considered that vehicle is to left sharp turn;Lateral accelerate is preset when the side acceleration values of vehicle are more than for negative and its absolute value
During angle value, then it is considered that vehicle sharp right-hand bend.
An acceleration rate threshold a can be set, it is assumed that the lateral acceleration of vehicle is b, ︱ b ︱ > ︱ a ︱ ≠ 0, and default
In time, when the acceleration of vehicle changes to-b from+b, then vehicle can be obtained in sudden turn of events road to the left;When the acceleration of vehicle
When changing to+b from-b, then vehicle can be obtained in sudden turn of events road to the right.
Preferably, the positive acceleration direction of vehicle can also be obtained, after detecting that zig zag behavior occurs for vehicle, is sentenced
The positive acceleration direction of disconnected vehicle, if the positive acceleration direction of vehicle does not change, it is believed that there occurs sudden turn of events road.
Driving range information is obtained by carrying out Integral Processing to velocity information, it is possible to increase the essence of driving range information
Degree.Due to traditional driving range computational methods, all it is that driving range is estimated with GPS positioning technology, is run over using vehicle
The discrete location longitude and latitude degrees of data sampled in journey forms driving locus, finally calculates path length and obtains distance travelled.This
Kind method is restricted be subject to several respects:If the frequency of sampling is too low, track estimation is just inaccurate.Individual lot satellite-signal is not
It is good so that position loses or drifts about, and influences track estimation.And the velocity information collected by OBD equipment is calculated and driven
Mileage, has the advantages that data are accurate, strong interference immunity, realized in the equipment of discrete sampling speed it is relatively simple, without relating to
And complicated calculations, the systematic error such as the conversion of GPS, track identification are smaller.
And each axle acceleration value collected by three axis accelerometer, with reference to the change of instantaneous G values, can accurately identify
Every abnormal driving event, the method such as anxious acceleration, anxious deceleration, zig zag, sudden turn of events road are simply easily achieved, round-the-clock can supervise
Control.
S3, obtains the accident information occurred, is established according to abnormal driving information and accident information and is used to evaluate driving wind
The machine learning model of danger.
Preferably, before step S3, the abnormal driving information for exceeding predetermined threshold value with the accident degree of correlation can also be verified,
To improve the accuracy of machine learning model and learning efficiency.
For example, the method that can be analyzed by single-factor, analyzes the whole abnormal driving information collected.Example
Such as, a degree of discrete processes are carried out in the distribution of the studied factor first, for example mileage is pressed and is segmented per 1000km,
Then the accident index of the sample to being distributed in each section counts, for example, can count in each accident, the exception of appearance
The average accident frequency of driving information, the contact of ultimate analysis between the two, it is for instance possible to use relevant function method, regression analysis
Method etc..Finally, by statistics, the spies such as driving range, speed, anxious acceleration, anxious deceleration, zig zag, sudden turn of events road, fatigue driving are obtained
Sign is closer with contacting for accident.
Preferably, before step S3, the process pre-processed to the driving information collected can also be included.
For example, can abandon containing missing values, exceptional value, outlier sample, every driving information is standardized
Processing, removes dimension.In initial data, accident relative recording is relatively fewer, and performance may be influenced by being directly used in training pattern,
Thus it is possible to the priori features based on single-factor analysis verification before, using unsupervised clustering method, by similar in feature
Data sample collects together, then using situation statistical indicator of being in danger of all categories, for example, can be average accident frequency,
Again it is marked for the reference target of each sample.The method of evaluation Clustering Effect is to calculate of all categories situation statistics of being in danger to refer to
Mark whether significant difference, for example, can utilize hypothesis testing complete.
Preferably, in step S3, the machine learning model for being used for evaluating driving risk can be established by SVM.Below into
Row describes in detail.
For example, grader can be realized using SVM, sample is divided into (1) two class of low-risk (0) and excessive risk.SVM is former
It can guarantee that structure risk is minimum in reason, there is outstanding general property and robustness, it is not high to the scale requirements of data set sample size, it is special
Not Shi He current data Finite Samples scene.It is, for example, possible to use the SVM of RBF.
The model mainly has two parameters:Penalty coefficient C and kernel functional parameter gamma, it is first determined both scopes, will
Both are supplied to model at various possible coefficient combinations, carry out learning test to data respectively, find performance highest model correspondence
Parameter is as optimized parameter.
Whole data set is divided into training set and test set in proportion first.Wherein training set is used for model learning, training
Obtained model is predicted the data of test set, and prediction result is compared statistics with the real marking of test set, calculates
F1score is obtained, for scoring model performance.
Wherein, the computational methods of each index are:
Wherein, TruePositive (true positives) representative model is categorized as excessive risk, and real marking is also excessive risk
Sample size;FalsePositive (false positive) representative model is categorized as low-risk, and the sample number that real marking is excessive risk
Amount;FalseNegative (false negative) representative model is categorized as low-risk, and real marking is also the sample size of low-risk;
When comparing two model performances, the height of F1-score is subject to, the model performance higher of F1-score higher, selects performance
The model of higher is as final machine learning model.
Preferably, after machine learning model is obtained, can also by cross validation method, to machine learning model into
Row renewal training.
In study, test, in order to further improve its general property, reduce the influence of over-fitting, can be to machine learning mould
Type is updated training.
It is for instance possible to use the method for cross validation (cross validation) is determining optimal ginseng by parameter optimization
After array is closed, the model that is obtained using the parameter training tests test set data, records final performance indicator.If
Index not yet reaches requirement, then redefines feature, iterative learning testing procedure, until index is met the requirements.
S4, is evaluated according to driving risk of the machine learning model to vehicle.
It should be noted that when being evaluated using driving risk of the machine learning model to vehicle, will be stored with first
The equipment of the model is installed on vehicle, with the driving information of true collection vehicle, records the vehicle traveling number in a period of time
According to.Again from accumulation extracting data individual features, input in housebroken machine learning model and calculated and judged, export
As a result the degree of risk of prediction is reflected, for example, low-risk, excessive risk can be divided into.
Preferably, the degree of risk of many levels according to the actual requirements, can also be set, represent different risk etc. respectively
Level.For example, 10 risk class can be set, risk increase is represented successively from 1 to 10.
A kind of driving risk evaluating method provided in this embodiment, passes through the various exceptions of collection vehicle in the process of moving
Driving information, and combine the accident information that has occurred and be modeled, the abnormal driving by the accident information that has occurred to vehicle
Information is matched and analyzed, and can obtain the driving risk evaluation result with very pinpoint accuracy, can be accurately objectively right
Evaluated vehicle carries out risk assessment, meets the demand that becomes more meticulous to risk assessment, and timely driver instructor corrects bad steering row
To reduce accident risk.
By the velocity information of OBD equipment collection vehicles, and the distance travelled information of vehicle, obtained traveling are obtained accordingly
Mileage is more accurate, and three axis accelerometer is calibrated in real time, vehicle traveling also can when having on acclive road
Acceleration information is gathered exactly.
As shown in figure 3, a kind of structural framing figure of the driving Risk Evaluating System provided for another embodiment of the present invention, should
System includes:
Information collecting device 1, the driving information for collection vehicle.
Preferably, driving information includes:The velocity information and acceleration information of vehicle.
Processor 2, for handling driving information, obtains the abnormal driving information of vehicle.
Preferably, the abnormal driving information of vehicle includes:Driving range information, fatigue driving information, anxious acceleration information, urgency
Deceleration, zig zag information and sudden turn of events road information.
It should be noted that information collecting device 1 and processor 2 can be arranged on vehicle, pass through OBD interfaces etc. and car
OBD equipment connection.
Information collecting device 1 can include a variety of equipment for being capable of collection vehicle driving information, such as, OBD equipment, speed
Sensor, accelerometer etc..
Preferably, information collecting device 1 includes:
OBD equipment 11, for obtaining the velocity information in vehicle travel process in real time.
Three axis accelerometer 12, for obtaining the acceleration information of all directions in vehicle travel process, wherein, with vehicle
Center is origin, and headstock direction is forward direction, and car door direction is lateral.
Preferably, processor 2 includes:
Spin matrix computing unit 21, for calculating the spin matrix for being used for correcting three axis accelerometer, according to spin moment
Battle array carries out rotation processing to the x, y, z axis of three axis accelerometer, obtains x ', y ', z ' axis.
Reference axis corrects unit 22, for being calibrated respectively to the x ', y ', z ' axis, and by the x ' axis after calibration and
Y ' axis is as the positive and lateral of the vehicle.
Preferably, processor 2 further includes:
Driving range computing unit 23, for carrying out Integral Processing to velocity information, obtains driving range information.
Fatigue driving judging unit 24, for the Velocity Time sequence according to velocity information and velocity information, obtains tired
Labor driving information.
It is anxious to accelerate anxious deceleration judging unit 25, for the acceleration information according to vehicle forward direction, identify the traveling side of vehicle
To the peak value and valley of acceleration time series, anxious acceleration information and anxious deceleration are obtained according to peak value and valley.
Take a sudden turn sudden turn of events road judging unit 26, for laterally being accelerated according to the lateral acceleration information of vehicle, identification vehicle
The peak value and valley of time series are spent, zig zag information is obtained according to peak value and valley, and binding time sequence obtains sudden turn of events road
Information.
Server 3, for obtaining the accident information occurred, establishes according to abnormal driving information and accident information and is used to comment
The machine learning model of valency driving risk, and evaluated according to driving risk of the machine learning model to vehicle.
Preferably, server 3 is additionally operable to increase new abnormal driving information and accident information, and machine learning model is carried out
Renewal training.
A kind of driving Risk Evaluating System provided in this embodiment, is adopted in real time by OBD equipment and three axis accelerometer
Collect the various abnormal driving information of vehicle in the process of moving, and combine the accident information occurred and be modeled, by having sent out
Raw accident information is matched and analyzed to the abnormal driving information of vehicle, can obtain the driving wind with very pinpoint accuracy
Dangerous assessment result, accurately objectively can carry out risk assessment to evaluated vehicle, meet the demand that becomes more meticulous to risk assessment,
Timely driver instructor correction bad steering behavior, reduces accident risk.
Reader should be understood that in the description of this specification, reference term " one embodiment ", " some embodiments ", " show
The description of example ", " specific example " or " some examples " etc. mean to combine the specific features of the embodiment or example description, structure,
Material or feature are contained at least one embodiment of the present invention or example.In the present specification, above-mentioned term is shown
The statement of meaning property need not be directed to identical embodiment or example.Moreover, particular features, structures, materials, or characteristics described
It may be combined in any suitable manner in any one or more of the embodiments or examples.In addition, without conflicting with each other, this
The technical staff in field can be by the different embodiments or example described in this specification and different embodiments or exemplary spy
Sign is combined and combines.
It is apparent to those skilled in the art that for convenience of description and succinctly, the dress of foregoing description
The specific work process with unit is put, may be referred to the corresponding process in preceding method embodiment, details are not described herein.
In several embodiments provided herein, it should be understood that disclosed apparatus and method, can pass through it
Its mode is realized.For example, device embodiment described above is only schematical, for example, the division of unit, is only
A kind of division of logic function, can there is an other dividing mode when actually realizing, for example, multiple units or component can combine or
Person is desirably integrated into another system, or some features can be ignored, or does not perform.
The unit illustrated as separating component may or may not be physically separate, be shown as unit
Component may or may not be physical location, you can with positioned at a place, or can also be distributed to multiple networks
On unit.Some or all of unit therein can be selected to realize the mesh of the embodiment of the present invention according to the actual needs
's.
In addition, each functional unit in each embodiment of the present invention can be integrated in a processing unit, can also
It is that unit is individually physically present or two or more units integrate in a unit.It is above-mentioned integrated
Unit can both be realized in the form of hardware, can also be realized in the form of SFU software functional unit.
If integrated unit is realized in the form of SFU software functional unit and is used as independent production marketing or in use, can
To be stored in a computer read/write memory medium.Based on such understanding, technical scheme substantially or
Say that the part to contribute to the prior art, or all or part of the technical solution can be embodied in the form of software product
Out, which is stored in a storage medium, including some instructions are used so that a computer equipment
(can be personal computer, server, or network equipment etc.) performs all or part of each embodiment method of the present invention
Step.And foregoing storage medium includes:USB flash disk, mobile hard disk, read-only storage (ROM, Read-OnlyMemory), deposit at random
Access to memory (RAM, RandomAccessMemory), magnetic disc or CD etc. are various can be with the medium of store program codes.
More than, it is only embodiment of the invention, but protection scope of the present invention is not limited thereto, and it is any to be familiar with
Those skilled in the art the invention discloses technical scope in, various equivalent modifications or substitutions can be readily occurred in,
These modifications or substitutions should be covered by the protection scope of the present invention.Therefore, protection scope of the present invention should be wanted with right
Subject to the protection domain asked.
Claims (10)
1. one kind driving risk evaluating method, it is characterised in that including:
The driving information of collection vehicle;
The driving information is handled, obtains the abnormal driving information of the vehicle;
The accident information occurred is obtained, is established according to the abnormal driving information and the accident information and is used to evaluate driving wind
The machine learning model of danger;
Evaluated according to driving risk of the machine learning model to the vehicle.
2. driving risk evaluating method according to claim 1, it is characterised in that the driving information includes:Vehicle
Velocity information and acceleration information, the driving information of the collection vehicle specifically include:
Obtain the velocity information in vehicle travel process in real time by OBD equipment;
The acceleration information of all directions in the vehicle travel process is obtained by three axis accelerometer, wherein, with the vehicle
Center be origin, headstock direction is forward direction, and car door direction is lateral.
3. driving risk evaluating method according to claim 2, it is characterised in that described to be obtained by three axis accelerometer
In the vehicle travel process before the acceleration information of all directions, further include:
The spin matrix for being used for correcting the three axis accelerometer is calculated, according to the spin matrix to the three axis accelerometer
X, y, z axis carry out rotation processing, obtain x ', y ', z ' axis;
The x ', y ', z ' axis are calibrated respectively, and using forward direction and side of the x ' axis and y ' axis after calibration as the vehicle
To.
4. the driving risk evaluating method according to Claims 2 or 3, it is characterised in that the abnormal driving letter of the vehicle
Breath includes:Driving range information, fatigue driving information, anxious acceleration information, anxious deceleration, zig zag information and sudden turn of events road letter
Breath, described that the driving information is handled, the abnormal driving information for obtaining the vehicle specifically includes:
Integral Processing is carried out to the velocity information, obtains the driving range information;
According to the velocity information and the Velocity Time sequence of the velocity information, the fatigue driving information is obtained;
According to the acceleration information of vehicle forward direction, identify the travel direction acceleration time series of the vehicle peak value and
Valley, the anxious acceleration information and the anxious deceleration are obtained according to the peak value and the valley;
According to the lateral acceleration information of the vehicle, the peak value and valley of the vehicle side acceleration time series are identified,
The zig zag information is obtained according to the peak value and valley, and sudden turn of events road information is obtained with reference to the time series.
5. driving risk evaluating method according to any one of claim 1 to 3, it is characterised in that described in the basis
Abnormal driving information and the accident information are established after the machine learning model for being used for evaluating driving risk, are further included:
By cross validation method, training is updated to the machine learning model.
6. one kind driving Risk Evaluating System, it is characterised in that including:
Information collecting device, the driving information for collection vehicle;
Processor, for handling the driving information, obtains the abnormal driving information of the vehicle;
Server, for obtaining the accident information occurred, establishes according to the abnormal driving information and the accident information and uses
In the machine learning model of evaluation driving risk, and commented according to driving risk of the machine learning model to the vehicle
Valency.
7. driving Risk Evaluating System according to claim 6, it is characterised in that the driving information includes:Vehicle
Velocity information and acceleration information, described information harvester include:
OBD equipment, for obtaining the velocity information in vehicle travel process in real time;
Three axis accelerometer, for obtaining the acceleration information of all directions in the vehicle travel process, wherein, with the vehicle
Center be origin, headstock direction is forward direction, and car door direction is lateral.
8. driving Risk Evaluating System according to claim 7, it is characterised in that the processor includes:
Spin matrix computing unit, for calculating the spin matrix for being used for correcting the three axis accelerometer, according to the rotation
Matrix carries out rotation processing to the x, y, z axis of the three axis accelerometer, obtains x ', y ', z ' axis;
Reference axis corrects unit, makees for being calibrated respectively to the x ', y ', z ' axis, and by the x ' axis after calibration and y ' axis
For the positive and lateral of the vehicle.
9. the driving Risk Evaluating System according to claim 7 or 8, it is characterised in that the abnormal driving letter of the vehicle
Breath includes:Driving range information, fatigue driving information, anxious acceleration information, anxious deceleration, zig zag information and sudden turn of events road letter
Breath, the processor further include:
Driving range computing unit, for carrying out Integral Processing to the velocity information, obtains the driving range information;
Fatigue driving judging unit, for the Velocity Time sequence according to the velocity information and the velocity information, obtains
The fatigue driving information;
It is anxious to accelerate anxious deceleration judging unit, for the acceleration information according to vehicle forward direction, identify the traveling of the vehicle
The peak value and valley of directional acceleration time series, the anxious information and described of accelerating is obtained according to the peak value and the valley
Anxious deceleration;
Take a sudden turn sudden turn of events road judging unit, for according to the lateral acceleration information of the vehicle, identifying that the vehicle laterally adds
The peak value and valley of Velocity Time sequence, the zig zag information is obtained according to the peak value and valley, and with reference to the time
Sequence obtains sudden turn of events road information.
10. the driving Risk Evaluating System according to any one of claim 6 to 8, it is characterised in that the server is also
For by cross validation method, training to be updated to the machine learning model.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711222806.5A CN107909678A (en) | 2017-11-29 | 2017-11-29 | One kind driving risk evaluating method and system |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201711222806.5A CN107909678A (en) | 2017-11-29 | 2017-11-29 | One kind driving risk evaluating method and system |
Publications (1)
Publication Number | Publication Date |
---|---|
CN107909678A true CN107909678A (en) | 2018-04-13 |
Family
ID=61847892
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201711222806.5A Pending CN107909678A (en) | 2017-11-29 | 2017-11-29 | One kind driving risk evaluating method and system |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN107909678A (en) |
Cited By (25)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109094482A (en) * | 2018-07-25 | 2018-12-28 | 哈尔滨工业大学 | Vehicle operation risk assessment information acquisition system, risk evaluating system and method |
CN109118055A (en) * | 2018-07-19 | 2019-01-01 | 众安信息技术服务有限公司 | A kind of driving behavior methods of marking and device |
CN109583776A (en) * | 2018-12-05 | 2019-04-05 | 百度在线网络技术(北京)有限公司 | A kind of vehicle body-sensing evaluating method, device, electronic equipment, medium and vehicle |
CN109840612A (en) * | 2018-07-24 | 2019-06-04 | 上海赢科信息技术有限公司 | User's driving behavior analysis method and system |
CN109872415A (en) * | 2018-12-28 | 2019-06-11 | 北京理工大学 | A kind of vehicle speed estimation method neural network based and system |
CN110197588A (en) * | 2019-06-03 | 2019-09-03 | 长安大学 | A kind of truck driving behavior appraisal procedure and device based on GPS track data |
CN110210980A (en) * | 2018-06-15 | 2019-09-06 | 腾讯科技(深圳)有限公司 | A kind of driving behavior appraisal procedure, device and storage medium |
CN110415510A (en) * | 2018-04-28 | 2019-11-05 | 江苏骏环昇旺科技产业股份有限公司 | A kind of system and judgment method for being judged to drive risk according to traffic violation |
CN110634288A (en) * | 2019-08-30 | 2019-12-31 | 上海电科智能系统股份有限公司 | Multi-dimensional urban traffic abnormal event identification method based on ternary Gaussian mixture model |
CN110751586A (en) * | 2019-02-21 | 2020-02-04 | 北京嘀嘀无限科技发展有限公司 | Order travel abnormity identification method and system |
CN110758398A (en) * | 2018-07-10 | 2020-02-07 | 阿里巴巴集团控股有限公司 | Driving risk detection method and device |
CN110949398A (en) * | 2019-11-28 | 2020-04-03 | 同济大学 | Method for detecting abnormal driving behavior of first-vehicle drivers in vehicle formation driving |
CN111105110A (en) * | 2018-10-25 | 2020-05-05 | 北京嘀嘀无限科技发展有限公司 | Driving risk determination method, device, medium and computing equipment |
CN111131617A (en) * | 2019-12-28 | 2020-05-08 | 长安大学 | Driving behavior analysis and feedback method based on smart phone |
CN111178452A (en) * | 2020-01-02 | 2020-05-19 | 平安科技(深圳)有限公司 | Driving risk identification method, electronic device and readable storage medium |
CN111259935A (en) * | 2020-01-09 | 2020-06-09 | 斑马网络技术有限公司 | Vehicle accident recognition method, device, equipment and storage medium |
CN111382883A (en) * | 2018-12-29 | 2020-07-07 | 北京嘀嘀无限科技发展有限公司 | Network appointment vehicle order-receiving authority control method, device and equipment |
WO2020169052A1 (en) * | 2019-02-21 | 2020-08-27 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for driving condition identification |
CN112396096A (en) * | 2020-11-07 | 2021-02-23 | 山东派蒙机电技术有限公司 | SVM-based method, system and medium for establishing driving risk identification model |
CN112613786A (en) * | 2020-12-30 | 2021-04-06 | 武汉理工大学 | Individualized and differentiated driving risk evaluation method based on driving safety event |
CN114072863A (en) * | 2019-07-08 | 2022-02-18 | 株式会社日立高新技术 | Accident risk diagnosis method, accident risk diagnosis device, and accident risk diagnosis system |
US11312374B2 (en) | 2020-01-31 | 2022-04-26 | International Business Machines Corporation | Prediction of accident risk based on anomaly detection |
CN114670752A (en) * | 2020-12-25 | 2022-06-28 | 西安墨科通讯科技有限公司 | Vehicle behavior data analysis and alarm method and safety device |
CN114707915A (en) * | 2022-06-06 | 2022-07-05 | 苏交科集团股份有限公司 | Vehicle transportation risk assessment system and method |
CN114842571A (en) * | 2021-02-02 | 2022-08-02 | 深圳市易流科技股份有限公司 | Method and device for determining driving behavior data |
Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101937421A (en) * | 2009-07-03 | 2011-01-05 | 上海大潮电子技术有限公司 | Method for collecting real-time operation information of vehicle for operation security risk assessment |
CN103818327A (en) * | 2013-11-22 | 2014-05-28 | 深圳先进技术研究院 | Method and device for analyzing driving behaviors |
WO2014207558A2 (en) * | 2013-06-27 | 2014-12-31 | Scope Technologies Holdings Limited | Onboard vehicle accident detection and damage estimation system and method of use |
CN105510632A (en) * | 2015-11-24 | 2016-04-20 | 上海汽车集团股份有限公司 | Method and apparatus for obtaining automobile acceleration data |
CN105799709A (en) * | 2016-03-10 | 2016-07-27 | 深圳市元征科技股份有限公司 | Method and device for recognizing sudden turn of vehicle |
-
2017
- 2017-11-29 CN CN201711222806.5A patent/CN107909678A/en active Pending
Patent Citations (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN101937421A (en) * | 2009-07-03 | 2011-01-05 | 上海大潮电子技术有限公司 | Method for collecting real-time operation information of vehicle for operation security risk assessment |
WO2014207558A2 (en) * | 2013-06-27 | 2014-12-31 | Scope Technologies Holdings Limited | Onboard vehicle accident detection and damage estimation system and method of use |
CN103818327A (en) * | 2013-11-22 | 2014-05-28 | 深圳先进技术研究院 | Method and device for analyzing driving behaviors |
CN105510632A (en) * | 2015-11-24 | 2016-04-20 | 上海汽车集团股份有限公司 | Method and apparatus for obtaining automobile acceleration data |
CN105799709A (en) * | 2016-03-10 | 2016-07-27 | 深圳市元征科技股份有限公司 | Method and device for recognizing sudden turn of vehicle |
Cited By (35)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110415510A (en) * | 2018-04-28 | 2019-11-05 | 江苏骏环昇旺科技产业股份有限公司 | A kind of system and judgment method for being judged to drive risk according to traffic violation |
CN110210980A (en) * | 2018-06-15 | 2019-09-06 | 腾讯科技(深圳)有限公司 | A kind of driving behavior appraisal procedure, device and storage medium |
CN110758398A (en) * | 2018-07-10 | 2020-02-07 | 阿里巴巴集团控股有限公司 | Driving risk detection method and device |
CN109118055A (en) * | 2018-07-19 | 2019-01-01 | 众安信息技术服务有限公司 | A kind of driving behavior methods of marking and device |
CN109118055B (en) * | 2018-07-19 | 2021-12-21 | 众安信息技术服务有限公司 | Driving behavior scoring method and device |
CN109840612A (en) * | 2018-07-24 | 2019-06-04 | 上海赢科信息技术有限公司 | User's driving behavior analysis method and system |
CN109094482A (en) * | 2018-07-25 | 2018-12-28 | 哈尔滨工业大学 | Vehicle operation risk assessment information acquisition system, risk evaluating system and method |
CN111105110A (en) * | 2018-10-25 | 2020-05-05 | 北京嘀嘀无限科技发展有限公司 | Driving risk determination method, device, medium and computing equipment |
CN109583776A (en) * | 2018-12-05 | 2019-04-05 | 百度在线网络技术(北京)有限公司 | A kind of vehicle body-sensing evaluating method, device, electronic equipment, medium and vehicle |
CN109872415A (en) * | 2018-12-28 | 2019-06-11 | 北京理工大学 | A kind of vehicle speed estimation method neural network based and system |
CN109872415B (en) * | 2018-12-28 | 2021-02-02 | 北京理工大学 | Vehicle speed estimation method and system based on neural network |
CN111382883A (en) * | 2018-12-29 | 2020-07-07 | 北京嘀嘀无限科技发展有限公司 | Network appointment vehicle order-receiving authority control method, device and equipment |
CN111599164B (en) * | 2019-02-21 | 2021-09-24 | 北京嘀嘀无限科技发展有限公司 | Driving abnormity identification method and system |
CN110751586A (en) * | 2019-02-21 | 2020-02-04 | 北京嘀嘀无限科技发展有限公司 | Order travel abnormity identification method and system |
WO2020169052A1 (en) * | 2019-02-21 | 2020-08-27 | Beijing Didi Infinity Technology And Development Co., Ltd. | Systems and methods for driving condition identification |
CN111599164A (en) * | 2019-02-21 | 2020-08-28 | 北京嘀嘀无限科技发展有限公司 | Driving abnormity identification method and system |
CN110197588A (en) * | 2019-06-03 | 2019-09-03 | 长安大学 | A kind of truck driving behavior appraisal procedure and device based on GPS track data |
EP3998591A4 (en) * | 2019-07-08 | 2023-07-12 | Hitachi High-Tech Corporation | Accident risk diagnosis method, accident risk diagnosis device, and accident risk diagnosis system |
CN114072863A (en) * | 2019-07-08 | 2022-02-18 | 株式会社日立高新技术 | Accident risk diagnosis method, accident risk diagnosis device, and accident risk diagnosis system |
CN110634288B (en) * | 2019-08-30 | 2022-06-21 | 上海电科智能系统股份有限公司 | Multi-dimensional urban traffic abnormal event identification method based on ternary Gaussian mixture model |
CN110634288A (en) * | 2019-08-30 | 2019-12-31 | 上海电科智能系统股份有限公司 | Multi-dimensional urban traffic abnormal event identification method based on ternary Gaussian mixture model |
CN110949398B (en) * | 2019-11-28 | 2021-09-03 | 同济大学 | Method for detecting abnormal driving behavior of first-vehicle drivers in vehicle formation driving |
CN110949398A (en) * | 2019-11-28 | 2020-04-03 | 同济大学 | Method for detecting abnormal driving behavior of first-vehicle drivers in vehicle formation driving |
CN111131617A (en) * | 2019-12-28 | 2020-05-08 | 长安大学 | Driving behavior analysis and feedback method based on smart phone |
CN111178452A (en) * | 2020-01-02 | 2020-05-19 | 平安科技(深圳)有限公司 | Driving risk identification method, electronic device and readable storage medium |
CN111259935A (en) * | 2020-01-09 | 2020-06-09 | 斑马网络技术有限公司 | Vehicle accident recognition method, device, equipment and storage medium |
US11312374B2 (en) | 2020-01-31 | 2022-04-26 | International Business Machines Corporation | Prediction of accident risk based on anomaly detection |
CN112396096A (en) * | 2020-11-07 | 2021-02-23 | 山东派蒙机电技术有限公司 | SVM-based method, system and medium for establishing driving risk identification model |
CN114670752A (en) * | 2020-12-25 | 2022-06-28 | 西安墨科通讯科技有限公司 | Vehicle behavior data analysis and alarm method and safety device |
CN114670752B (en) * | 2020-12-25 | 2024-02-13 | 西安墨科通讯科技有限公司 | Method and safety device based on vehicle behavior data analysis and alarm |
CN112613786B (en) * | 2020-12-30 | 2022-05-06 | 武汉理工大学 | Individualized and differentiated driving risk evaluation method based on driving safety event |
CN112613786A (en) * | 2020-12-30 | 2021-04-06 | 武汉理工大学 | Individualized and differentiated driving risk evaluation method based on driving safety event |
CN114842571A (en) * | 2021-02-02 | 2022-08-02 | 深圳市易流科技股份有限公司 | Method and device for determining driving behavior data |
CN114707915A (en) * | 2022-06-06 | 2022-07-05 | 苏交科集团股份有限公司 | Vehicle transportation risk assessment system and method |
CN114707915B (en) * | 2022-06-06 | 2022-08-12 | 苏交科集团股份有限公司 | Vehicle transportation risk assessment system and method |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN107909678A (en) | One kind driving risk evaluating method and system | |
US11407410B2 (en) | Method and system for estimating an accident risk of an autonomous vehicle | |
JP7155266B2 (en) | Systems and methods for racing data analysis using telemetry data and wearable sensor data | |
CN110197588A (en) | A kind of truck driving behavior appraisal procedure and device based on GPS track data | |
CN111761409A (en) | Multi-sensor numerical control machine tool cutter wear monitoring method based on deep learning | |
CN109118055A (en) | A kind of driving behavior methods of marking and device | |
CN110263870B (en) | Motion state identification method and device, intelligent terminal and storage medium | |
CN111552754B (en) | Ship track similarity measurement method and system | |
CN104269028A (en) | Fatigue driving detection method and system | |
CN110705582B (en) | Ship bridge-crossing behavior feature extraction method based on improved K-Means clustering algorithm | |
JP2019112049A (en) | Method for recognizing driving style of driver of land vehicle and corresponding device | |
CN108021888A (en) | A kind of fall detection method | |
CN105975757A (en) | Urgent speed reduction behavior recognition method based on vehicle driving data | |
EP3578433A1 (en) | Method for estimating an accident risk of an autonomous vehicle | |
CN110263836B (en) | Bad driving state identification method based on multi-feature convolutional neural network | |
CN117173913A (en) | Traffic control method and system based on traffic flow analysis at different time periods | |
RU2018145712A (en) | SYSTEMS AND METHODS FOR GEOPOSITIONING AND ASSESSMENT IN SCORES FOR VEHICLE DATA IN COMMUNITIES | |
CN112967801A (en) | PAI value processing method, PAI value processing device, PAI value processing equipment and storage medium | |
EP3382570A1 (en) | Method for characterizing driving events of a vehicle based on an accelerometer sensor | |
CN109117719B (en) | Driving posture recognition method based on local deformable component model fusion characteristics | |
CN106067036A (en) | Set up different automobile types based on artificial intelligence's unsupervised learning K means method and divide the long-range loss assessment system of part and method | |
CN106056140A (en) | System and method for establishing working condition division remote damage assessment of different vehicle types based on artificial intelligence supervised learning linear regression method | |
CN112329714A (en) | GM-HMM-based driver high-speed driving distraction identification modeling method | |
CN117268424B (en) | Multi-sensor fusion automatic driving hunting method and device | |
CN116821398B (en) | Data set acquisition method for road defect recognition model training |
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 | ||
AD01 | Patent right deemed abandoned | ||
AD01 | Patent right deemed abandoned |
Effective date of abandoning: 20210817 |