CN108280415A - Driving behavior recognition methods based on intelligent mobile terminal - Google Patents
Driving behavior recognition methods based on intelligent mobile terminal Download PDFInfo
- Publication number
- CN108280415A CN108280415A CN201810045607.XA CN201810045607A CN108280415A CN 108280415 A CN108280415 A CN 108280415A CN 201810045607 A CN201810045607 A CN 201810045607A CN 108280415 A CN108280415 A CN 108280415A
- Authority
- CN
- China
- Prior art keywords
- driving behavior
- data
- vehicle
- mobile terminal
- intelligent mobile
- 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
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/50—Context or environment of the image
- G06V20/59—Context or environment of the image inside of a vehicle, e.g. relating to seat occupancy, driver state or inner lighting conditions
- G06V20/597—Recognising the driver's state or behaviour, e.g. attention or drowsiness
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
- G06F18/232—Non-hierarchical techniques
- G06F18/2321—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions
- G06F18/23213—Non-hierarchical techniques using statistics or function optimisation, e.g. modelling of probability density functions with fixed number of clusters, e.g. K-means clustering
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- Probability & Statistics with Applications (AREA)
- Multimedia (AREA)
- Traffic Control Systems (AREA)
- Navigation (AREA)
Abstract
The driving behavior recognition methods based on intelligent mobile terminal that the invention discloses a kind of, including：S1, vehicle reset condition data are acquired and screened by intelligent mobile terminal, include the acceleration and angular speed information of three axis；S2, vehicle original motion status data is pre-processed；S3, driving behavior multi-feature vector is obtained from pretreated data using principal component analytical method；S4, clustering is carried out to driving behavior multi-feature vector using k means clustering algorithms, obtains preferable clustering number mesh；S5, according to preferable clustering number mesh after, and multi-feature vector is clustered using FCM algorithms, obtains the final cluster structure of deblurring；S6, acquisition real-time vehicle status data, and the driving behavior of vehicle is identified according to final cluster result.The present invention realizes the fine cluster of driving behavior data, and vehicle drive behavioural characteristic is effectively gathered for three class behavior of turning, speed change and lane change.
Description
Technical field
The present invention relates to vehicle assistant drive technical field more particularly to a kind of driving behaviors based on intelligent mobile terminal
Recognition methods and system.
Background technology
With increasing rapidly for city vehicle quantity, driver's bad steering behavior become one of traffic problems it is important because
Element.The traffic jam caused by bad steering behavior and traffic accident problem are advantageously accounted for the research of vehicle drive behavior.
View-based access control model image is broadly divided into for the research of driving behavior at present and based on both means of Fusion.
For the driving behavior analysis of view-based access control model image, light intensity in practical application, camera angle it is uncertain
The interference of property and ambient enviroment can all impact recognition effect.Vehicle-state, phase are perceived by Multi-sensor fusion
It is smaller that situation is disturbed for；But install portable sensor process complexity additional to vehicle, cost is higher.
K-means algorithms and FCM algorithms are clustering methods important in unsupervised machine learning, can explore driving
The hidden attribute of behavioural characteristic data, to realize that driving behavior is classified.K-means clusters belong to hard clustering algorithm, sample point
It can only belong to or be not belonging to certain one kind, therefore classifying quality has limitation；FCM algorithms belong to soft clustering algorithm, using degree of membership
Function representation sample point belongs to certain a kind of degree, and classification results are finer, but algorithm increases calculation amount, speed relative to
K-means algorithms are declined.
Invention content
The technical problem to be solved in the present invention is, for car steering behavior classification and identification problem, to provide one kind and be based on
The driving behavior recognition methods of intelligent mobile terminal.
The technical solution adopted by the present invention to solve the technical problems is：
The present invention provides a kind of driving behavior recognition methods based on intelligent mobile terminal, and this approach includes the following steps：
S1, acquire and screen vehicle reset condition data by intelligent mobile terminal, including the acceleration of three axis and angle
Velocity information；
S2, vehicle original motion status data is pre-processed；
S3, driving behavior multi-feature vector is obtained from pretreated data using principal component analytical method；
S4, clustering is carried out to driving behavior multi-feature vector using k-means clustering algorithms, is most preferably clustered
Number；
S5, multi-feature vector is clustered according to preferable clustering number mesh, and using FCM algorithms, obtains deblurring
Final cluster result；S6, acquisition real-time vehicle status data, and the driving behavior of vehicle is known according to final cluster result
Not.
Further, the intelligent mobile terminal collection vehicle data in this method of the invention are specially：
Utilize the 3-axis acceleration sensor and three-axis gyroscope acquisition vehicle original motion shape built in intelligent mobile terminal
The value of state data, sensor output describes motion state of the equipment relative to local Coordinate System.When equipment plane and vehicle are flat
Face is parallel, and when equipment Y direction is consistent with vehicle forward direction, and the acceleration in three axis direction of equipment can be retouched accurately
State the motion state of vehicle.By comparing normally travel, acceleration and the sensor parameters change turned right under these three driving behaviors
Change, finds the X of acceleration transducer, data are most sensitive to behavioural characteristic about the z axis for Y-axis and gyroscope.
In view of the behaviors such as turning and lane change have similitude in very short time interval, can extract in intervals
The temporal signatures of sensing data.
Further, the original motion status data in this method of the invention, which pre-processes, is specially：
The interference noise in collected data is filtered using moving average filtering algorithm.
Further, principal component analytical method is specially in this method of the invention：
Signature analysis is carried out to pretreated data first, the mean value, standard deviation, middle position of every column data is calculated
Then number, maximum value, minimum value and range add the related coefficient between every two row, 21 dimensions of composition in three column datas
The various features of travel condition of vehicle can be described according to vector.For multigroup vehicle data of acquisition, can be described as more than one
The matrix that row 21 arranges：
Wherein the value of p is 21.Before principal component analysis need that every column data is normalized.
Principal component analysis, which is divided into, to be calculated correlation matrix, calculates characteristic root, determines principal component number and calculate 4 step of principal component.
By principal component analysis, original motion state matrix dimension reduces, to greatly reduce the complexity of algorithm.
Further, the k-means algorithm clustering methods in this method of the invention are specially：
Clustering is carried out to given driving behavior vector data sample set using k-means clustering algorithms, most
It is divided into k classes eventually, cluster centre is updated to solve the optimal solution of object function by iteration optimization, is as follows：
1) k sample is randomly selected from X as initial cluster center, gives iterations and threshold value；
2) each data sample is calculated to the distance of cluster centre, is referred to apart from nearest one kind；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error.
K is incremented by since 2, and according to DBI index DsIObtain preferable clustering number mesh K.
Further, the FCM algorithm clustering methods in this method of the invention are specially：
Cluster centre and subordinated-degree matrix are changed to solve the optimal value of object function by iteration optimization, and algorithm specifically walks
It is rapid as follows：
1) K sample is randomly selected as initial cluster center, gives iterations, threshold value and fuzzy parameter；
2) it calculates and updates degree of membership；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error；
5) class where finally taking its degree of membership maximum value to each sample is classified as its certainty, realizes fuzzy clustering
As a result deblurring.
Above-mentioned technical proposal is connect, driving behavior is divided into turning behavior, speed change behavior and lane change behavior by final cluster result；
Wherein turning behavior is divided into violent, mild and moderate three kinds of turning behaviors；
Speed change behavior is divided into three kinds of violent, moderate and mild speed change behaviors；
Lane change behavior is divided into no lane change behavior and has lane change behavior, has lane change behavior to be divided into violent, mild and moderate three
Kind.
The driving behavior identifying system based on intelligent mobile terminal that the present invention also provides a kind of, including：
Reset condition acquisition module, for vehicle reset condition data to be acquired and screened by intelligent mobile terminal, including
The acceleration and angular speed information of three axis；
Preprocessing module, for being pre-processed to vehicle original motion status data；
Principal component analysis module, it is comprehensive for obtaining driving behavior from pretreated data using principal component analytical method
Close feature vector；
K-means cluster arithmetic modules, for being carried out to driving behavior multi-feature vector using k-means clustering algorithms
Clustering obtains preferable clustering number mesh；
FCM algoritic modules, for being clustered to multi-feature vector according to preferable clustering number mesh, and using FCM algorithms,
Obtain the final cluster structure of deblurring；
Identification module, for the real-time vehicle status data to acquisition, according to final cluster result to the driving row of vehicle
To be identified.
The present invention also provides a kind of computer readable storage medium, it is stored in the storage medium and is based on intelligence for executing
The computer program of the driving behavior recognition methods of energy mobile terminal.
Beneficial effects of the present invention：The present invention utilizes what is acquired from intelligent mobile terminal sensor can characterize vehicle traveling
The data of state extract the feature that can characterize vehicle drive behavior after being filtered to data.Pass through principal component analysis
Feature Dimension Reduction is carried out to vehicle drive behavioural characteristic, finally realizes the classification to vehicle drive behavior, and achieve good
Clustering Effect.
Description of the drawings
Present invention will be further explained below with reference to the attached drawings and examples, in attached drawing：
Fig. 1 is the main-process stream schematic diagram of the embodiment of the present invention；
Fig. 2 is the equipment of the embodiment of the present invention coordinate system and global coordinate system schematic diagram；
Fig. 3 is that the acceleration transducer X-axis of the embodiment of the present invention is original with filtering data contrast schematic diagram.
Fig. 4 is the principal component analysis result schematic diagram of the embodiment of the present invention.
Fig. 5 is the embodiment of the present invention based on Z1、Z2The FCM Clustering Effect schematic diagrames for index of turning.
Specific implementation mode
In order to make the purpose , technical scheme and advantage of the present invention be clearer, with reference to the accompanying drawings and embodiments, right
The present invention is further elaborated.It should be appreciated that described herein, specific examples are only used to explain the present invention, not
For limiting the present invention.
As shown in Figure 1, the driving behavior recognition methods based on intelligent mobile terminal of the embodiment of the present invention, including following step
Suddenly：
S1, acquire and screen vehicle reset condition data by intelligent mobile terminal, including the acceleration of three axis and angle
Velocity information；
S2, vehicle original motion status data is pre-processed；
S3, driving behavior multi-feature vector is obtained from pretreated data using principal component analytical method；
S4, clustering is carried out to driving behavior multi-feature vector using k-means clustering algorithms, is most preferably clustered
Number；
S5, multi-feature vector is clustered according to preferable clustering number mesh, and using FCM algorithms, obtains deblurring
Final cluster structure；
S6, acquisition real-time vehicle status data, and classified to the driving behavior of vehicle according to final cluster result.
Intelligent mobile terminal collection vehicle data in this method are specially：
Utilize the 3-axis acceleration sensor and three-axis gyroscope acquisition vehicle original motion shape built in intelligent mobile terminal
The value of state data, sensor output describes motion state of the equipment relative to local Coordinate System.When equipment plane and vehicle are flat
Face is parallel, and when equipment Y direction is consistent with vehicle forward direction, and the acceleration in three axis direction of equipment can be retouched accurately
State the motion state of vehicle.By comparing normally travel, acceleration and the sensor parameters change turned right under these three driving behaviors
Change, finds the X of acceleration transducer, data are most sensitive to behavioural characteristic about the z axis for Y-axis and gyroscope.
In view of the behaviors such as turning and lane change have similitude in very short time interval, can extract in intervals
The temporal signatures of sensing data.
Original motion status data in this method, which pre-processes, is specially：
The interference noise in collected data is filtered using moving average filtering algorithm.
Principal component analytical method in this method is specially：
Signature analysis is carried out to pretreated data first, the mean value, standard deviation, middle position of every column data is calculated
Then number, maximum value, minimum value and range add the related coefficient between every two row, 21 dimensions of composition in three column datas
The various features of travel condition of vehicle can be described according to vector.For multigroup vehicle data of acquisition, a multirow can be described as
The matrix of 21 row：
Wherein the value of p is 21.It before principal component analysis needs that every column data is normalized, principal component analysis side
Method can be realized describes original matrix characteristic with relatively low dimension, substantially reduces the complexity of algorithm.It is related that principal component analysis is divided into calculating
Matrix calculates characteristic root, determines principal component number and calculates 4 step of principal component：
1) the correlation matrix R=(r of former eigenmatrix are calculatedij)p×p, wherein rijRelated coefficient between being arranged for i, j two；
2) correlation matrix feature root is calculated.Its characteristic root is arranged from big to small, i.e. λ1≥λ2≥…≥λp；
3) principal component number is determined.It is determined and is led according to the principal component contribution rate of accumulative total α (generally taking 85%~95%) of setting
The number m of ingredient；
4) principal component is calculated.Calculate the corresponding unit character vector β of m (m≤p) a characteristic rooti, from gained unit character to
Measure corresponding principal component Zi。
K-means algorithm clustering methods in this method are specially：
Clustering is carried out to given driving behavior vector data sample set using k-means clustering algorithms, most
It is divided into k classes eventually, cluster centre is updated to solve the optimal solution of object function by iteration optimization, is as follows：
1) k sample is randomly selected from X as initial cluster center, gives iterations and threshold value；
2) each data sample is calculated to the distance of cluster centre, is referred to apart from nearest one kind；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error.
K is incremented by since 2, and according to DBI index DsIObtain preferable clustering number mesh K.
FCM algorithm clustering methods in this method are specially：
On the basis of k-means algorithms obtain preferable clustering number mesh K, it is K to specify FCM algorithm cluster centres.Then
Cluster centre and subordinated-degree matrix are changed to solve the optimal value of object function by iteration optimization, and algorithm is as follows：
1) K sample is randomly selected as initial cluster center, gives iterations, threshold value and fuzzy parameter；
2) it calculates and updates degree of membership；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error；
5) class where finally taking its degree of membership maximum value to each sample is classified as its certainty, realizes fuzzy clustering
As a result deblurring.
In another specific embodiment of the present invention：
Using smart mobile phone as intelligent mobile terminal, the APP based on Android platform is obtained in mobile phone accordingly for exploitation
It sets 3-axis acceleration sensor and three-axis gyroscope obtains mobile phone original motion status data.
Fig. 2 gives the direction relations between device coordinate system and global coordinate system.
As shown in Fig. 2, by mobile phone coordinate system and global coordinate system to just, the motion state of mobile phone just reflects the fortune of vehicle
Dynamic state feature.
Acceleration transducer and gyro data are the matrix that n rows 6 arrange, and extraction the 1st, 2,6 column data of matrix is as vehicle
Important original operating status characteristic.The interference noise in collected data is filtered using moving average filtering algorithm.
For data Xj={ x (i), i=1,2 ..., n }, sets sliding window size as M=10, then the filtering output value y at i moment
(i) it is
Y (i)=y (i-1)+(x (i+p)-x (i-q))/M
Wherein, p=(M-1)/2；Q=p+1.
Fig. 3 gives that acceleration transducer X-axis is original to be compared with filtering data.
As shown in figure 3, the smooth effect of moving average filtering is apparent, the noise that vehicle vibration is brought can be effectively filtered out.
As shown in table 1, calculating analysis is carried out to pretreated data, obtains 7 alternative features parameters of every column data.
1 alternative features parameter list of table
21 groups of obtained data form 21 maintenance and operations and move feature vector, and the motion feature vector data of multi collect collectively constitutes
The matrix of 21 row：
Every column data of matrix is normalized, formula is：
Wherein xmin、xmaxThe minimum value and maximum value of each column characteristic value data are indicated respectively.
Principal component analysis, which is divided into, to be calculated correlation matrix, calculates characteristic root, determine principal component number, calculate 3 step of principal component：
1) the correlation matrix R=(r of former eigenmatrix are calculatedij)p×pCalculation formula is：
Wherein rijRelated coefficient between being arranged for i, j two, x is column vector.
2) correlation matrix feature root is calculated.Calculate the solution of det (R- λ E)=0, the as characteristic root of R, by its characteristic root from
Minispread, i.e. λ are arrived greatly1≥λ2≥…≥λp>0。
3) principal component number is determined.It is determined and is led according to the principal component contribution rate of accumulative total α (generally taking 85%~95%) of setting
The number m of ingredient, calculation formula are：
4) principal component is calculated.Calculate the corresponding unit character vector β of m (m≤p) a characteristic rootiRespectively：
Corresponding principal component Z is obtained by gained unit character vectori, calculation formula is：
Zi=β1iX1+β2iX2+…+βpiXp, i=1,2 ... m
Fig. 4 gives data principal component analysis result.
As shown in figure 4, the 6th principal component is its inflection point, broken line is more precipitous before inflection point, and broken line becomes gentle after inflection point,
Choose preceding 6 principal components.
As shown in table 2, characteristic value, contribution rate and the contribution rate of accumulative total of preceding 6 principal components of correlation matrix are obtained.
2 principal component analysis table of table
Clustering is carried out to given driving behavior vector data sample set using k-means clustering algorithms, most
It is divided into k classes eventually, object function is：
Wherein, μiFor cluster CiCluster centre, calculation formula is：
D (x, μi) indicate data sample x to cluster C where itiCluster centre μiMinkowski distances, calculation formula
For：
Cluster centre is updated to solve the optimal solution of object function by iteration optimization, is as follows：
1) k sample is randomly selected from X as initial cluster center, gives iterations and threshold value；
2) each data sample is calculated to the distance of cluster centre, is referred to apart from nearest one kind；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error.
K is incremented by since 2, and according to DBI index DsIPreferable clustering number mesh K is obtained, calculation formula is：
The object function of FCM algorithm clustering methods is：
Wherein, uij, m be degree of membership and fuzzy parameter, take m values be 2.
On the basis of k-means obtains preferable clustering number purpose, cluster centre and degree of membership square are changed by iteration optimization
Battle array solves the optimal value of object function, and algorithm is as follows：
1) K sample is randomly selected as initial cluster center, gives iterations, threshold value and fuzzy parameter；
2) it calculates and updates degree of membership；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error.
5) class where finally taking its degree of membership maximum value to each sample is classified as its certainty, realizes fuzzy clustering
As a result deblurring.
Fig. 5 gives based on Z1、Z2The FCM Clustering Effects for index of turning.
As shown in figure 5, two indexs of triangle class cluster all very littles in left side, show that vehicle does not have apparent turning behavior,
And 3 class clusters on the right side of it then have different degrees of turning behavior, can be divided into these three violent, mild and moderate turnings
Behavior.
As shown in table 3, speed change index is gathered for 3 classes, wherein violent speed change behavior accounting 35%, moderate and mild speed change
Behavior accounting 65%.
Table 3 is based on Z3And Z4Speed change driving behavior cluster
As shown in table 4, lane change index is gathered for 4 classes, and most of situation does not have lane change behavior to vehicle in the process of moving；
And in the lane change behavior of generation, it can be divided into according to lane change index violent, mild and moderate.
Table 4 is based on Z5Lane change driving behavior cluster
By acquiring the data that acceleration and gyro sensor obtain on mobile phone in real time, 3+3+ may finally be classified as
One kind in 4=10 classes.I.e. automobile be straight trip speed change, or turning or lane change, then degree how about.
But it does not fully achieve also at this stage, it is studying at present as a result, it is possible to by carrying out analysis realization to initial data
Vehicle behavior is classified, that is, it is feasible to do so.And it was found that carry out principal component analysis after, 6 principal components and vehicle that extract
There are some relationships for behavior.For example, only considering principal component Z1And Z2Be assured that vehicle whether turning and degree how；
Only consider principal component Z3And Z4It may determine that its speed change situation；Consider principal component Z5It may determine that lane change situation.
That is, present invention finds 6 principal components, there are some relationships with vehicle behavior, and respectively come consider turn
Three aspects of curved, speed change and lane change, achievement of classifying in obtaining 10, shown in table 5 specific as follows：
10 kinds of behaviors cluster of the turning of table 5, speed change and lane change
| Turning | Speed change | Lane change |
| Acutely | Acutely | Acutely |
| It is moderate | It is moderate | It is moderate |
| Mildly | Mildly | Mildly |
| Nothing |
It should be understood that for those of ordinary skills, it can be modified or changed according to the above description,
And all these modifications and variations should all belong to the protection domain of appended claims of the present invention.
Claims (9)
1. a kind of driving behavior recognition methods based on intelligent mobile terminal, which is characterized in that this approach includes the following steps：
S1, vehicle reset condition data are acquired and screened by intelligent mobile terminal, include the acceleration and angular speed of three axis
Information；
S2, vehicle original motion status data is pre-processed；
S3, driving behavior multi-feature vector is obtained from pretreated data using principal component analytical method；
S4, clustering is carried out to driving behavior multi-feature vector using k-means clustering algorithms, obtains preferable clustering number
Mesh；
S5, multi-feature vector is clustered according to preferable clustering number mesh, and using FCM algorithms, obtains the final of deblurring
Cluster result；
S6, acquisition real-time vehicle status data, and the driving behavior of vehicle is identified according to final cluster result.
2. the driving behavior recognition methods according to claim 1 based on intelligent mobile terminal, which is characterized in that this method
Middle intelligent mobile terminal collection vehicle data are specially：
Utilize the 3-axis acceleration sensor and three-axis gyroscope acquisition vehicle original motion status number built in intelligent mobile terminal
According to the value of sensor output describes motion state of the equipment relative to local Coordinate System.
3. the driving behavior recognition methods according to claim 1 based on intelligent mobile terminal, which is characterized in that this method
In original motion status data pretreatment be specially：
The interference noise in collected data is filtered using moving average filtering algorithm.
4. the driving behavior recognition methods according to claim 1 based on intelligent mobile terminal, which is characterized in that this method
In principal component analytical method be specially：
Signature analysis is carried out to pretreated data, be calculated the mean value of every column data, standard deviation, median, maximum value,
Then minimum value and range add the related coefficient between every two row, 21 dimension data vector descriptions of composition in three column datas
The various features of travel condition of vehicle, multigroup vehicle data of acquisition form the matrix of a multirow 21 row：
Wherein the value of p is 21；
Every column data of matrix is normalized；
Principal component analysis, which is divided into, to be calculated correlation matrix, calculates characteristic root, determine principal component number, calculate 4 step of principal component, and master is passed through
Constituent analysis, original motion state matrix obtain dimension-reduction treatment, reduce the complexity of algorithm.
5. the driving behavior recognition methods according to claim 1 based on intelligent mobile terminal, which is characterized in that this method
Middle k-means algorithms clustering method is specially：
Clustering is carried out to given driving behavior vector using k-means clustering algorithms, is finally divided into k
Class updates cluster centre to solve the optimal solution of object function by iteration optimization, is as follows：
1) k sample is randomly selected from X as initial cluster center, gives iterations and threshold value；
2) each data sample is calculated to the distance of cluster centre, is referred to apart from nearest one kind；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error.
K is incremented by since 2, and according to DBI index DsIObtain preferable clustering number mesh K.
6. the driving behavior clustering method according to claim 1 based on intelligent mobile terminal, which is characterized in that this method
Middle FCM algorithms clustering method is specially：
Cluster centre and subordinated-degree matrix are changed to solve the optimal value of object function by iteration optimization, and algorithm specific steps are such as
Under：
1) K sample is randomly selected as initial cluster center, gives iterations, threshold value and fuzzy parameter；
2) it calculates and updates degree of membership；
3) cluster centre point is recalculated, cluster centre is updated；
4) repeat the 2) step, be less than threshold value until reaching maximum iteration or square error；
5) class where finally taking its degree of membership maximum value to each sample is classified as its certainty, realizes fuzzy clustering result
Deblurring.
7. the driving behavior recognition methods according to claim 1 based on intelligent mobile terminal, which is characterized in that final poly-
Driving behavior is divided into turning behavior, speed change behavior and lane change behavior by class result；
Wherein turning behavior is divided into violent, mild and moderate three kinds of turning behaviors；
Speed change behavior is divided into three kinds of violent, moderate and mild speed change behaviors；
Lane change behavior is divided into no lane change behavior and has lane change behavior, has lane change behavior to be divided into violent, mild and three kinds moderate.
8. a kind of driving behavior identifying system based on intelligent mobile terminal, which is characterized in that including：
Reset condition acquisition module, for vehicle reset condition data, including three to be acquired and screened by intelligent mobile terminal
The acceleration and angular speed information of axis；
Preprocessing module, for being pre-processed to vehicle original motion status data；
Principal component analysis module, it is special for obtaining driving behavior synthesis from pretreated data using principal component analytical method
Sign vector；
K-means cluster arithmetic modules, for being clustered to driving behavior multi-feature vector using k-means clustering algorithms
It divides, obtains preferable clustering number mesh；
FCM algoritic modules are obtained for being clustered to multi-feature vector according to preferable clustering number mesh, and using FCM algorithms
The final cluster structure of deblurring；
Identification module, for the real-time vehicle status data of acquisition, according to final cluster result to the driving behavior of vehicle into
Row identification.
9. a kind of computer readable storage medium, which is characterized in that be stored in the storage medium for executing such as claim 1
In the driving behavior recognition methods based on intelligent mobile terminal computer program.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810045607.XA CN108280415A (en) | 2018-01-17 | 2018-01-17 | Driving behavior recognition methods based on intelligent mobile terminal |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810045607.XA CN108280415A (en) | 2018-01-17 | 2018-01-17 | Driving behavior recognition methods based on intelligent mobile terminal |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108280415A true CN108280415A (en) | 2018-07-13 |
Family
ID=62803855
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810045607.XA Pending CN108280415A (en) | 2018-01-17 | 2018-01-17 | Driving behavior recognition methods based on intelligent mobile terminal |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108280415A (en) |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109061706A (en) * | 2018-07-17 | 2018-12-21 | 江苏新通达电子科技股份有限公司 | A method of the vehicle drive behavioural analysis based on T-Box and real-time road map datum |
| CN109117869A (en) * | 2018-07-20 | 2019-01-01 | 汉纳森（厦门）数据股份有限公司 | User's portrait method, medium and system |
| CN109345772A (en) * | 2018-10-15 | 2019-02-15 | 深圳市元征科技股份有限公司 | A kind of fatigue driving recognition methods and relevant device |
| CN109408557A (en) * | 2018-09-29 | 2019-03-01 | 东南大学 | A kind of traffic accidents reason analysis method clustered based on multiple correspondence and K-means |
| CN109615018A (en) * | 2018-12-24 | 2019-04-12 | 广东德诚科教有限公司 | User individual behavior evaluation method, apparatus, computer equipment and storage medium |
| CN109708634A (en) * | 2018-12-12 | 2019-05-03 | 平安科技（深圳）有限公司 | Judge automatically method, apparatus, storage medium and the electronic equipment of driving behavior |
| CN109711459A (en) * | 2018-12-24 | 2019-05-03 | 广东德诚科教有限公司 | User individual action estimation method, apparatus, computer equipment and storage medium |
| CN110175557A (en) * | 2019-05-24 | 2019-08-27 | 北京交通大学 | Driver's driving behavior method for identifying and classifying and system based on driving characteristics group |
| CN110321954A (en) * | 2019-07-03 | 2019-10-11 | 中汽研（天津）汽车工程研究院有限公司 | The driving style classification and recognition methods of suitable domestic people and system |
| CN110329271A (en) * | 2019-06-18 | 2019-10-15 | 北京航空航天大学杭州创新研究院 | A kind of multisensor vehicle driving detection system and method based on machine learning |
| CN110544151A (en) * | 2019-08-20 | 2019-12-06 | 北京市天元网络技术股份有限公司 | Method and equipment for determining whether user is online car booking driver |
| CN110682914A (en) * | 2019-09-29 | 2020-01-14 | 皖西学院 | Driving behavior recognition system and method based on wireless perception |
| CN111327755A (en) * | 2018-12-14 | 2020-06-23 | 西安博瑞集信电子科技有限公司 | Intelligent vehicle driving state identification system and method based on mobile phone |
| CN111506627A (en) * | 2020-04-21 | 2020-08-07 | 成都路行通信息技术有限公司 | Target behavior clustering method and system |
| CN111800314A (en) * | 2020-07-06 | 2020-10-20 | 广元量知汇科技有限公司 | Automatic driving fault monitoring system |
| CN111881952A (en) * | 2020-07-14 | 2020-11-03 | 扬州大学 | Driver tendency analysis method based on early warning big data |
| CN112052895A (en) * | 2020-09-03 | 2020-12-08 | 同济大学 | Pure electric vehicle driving style clustering method |
| CN112241748A (en) * | 2019-07-16 | 2021-01-19 | 广州汽车集团股份有限公司 | Data dimension reduction method and device based on multi-source information entropy difference |
| CN112598052A (en) * | 2020-12-21 | 2021-04-02 | 中建八局第二建设有限公司 | Mechanical attitude analysis method and system based on K-Means |
| CN112622914A (en) * | 2020-12-21 | 2021-04-09 | 武汉理工大学 | New energy automobile driving safety state identification system |
| CN113378998A (en) * | 2021-07-12 | 2021-09-10 | 西南石油大学 | Stratum lithology while-drilling identification method based on machine learning |
| CN113569950A (en) * | 2021-07-28 | 2021-10-29 | 大唐环境产业集团股份有限公司 | Power station equipment fault monitoring model generation method, system and device |
| CN113642600A (en) * | 2021-06-29 | 2021-11-12 | 桂林电子科技大学 | Driving behavior feature extraction method based on mRMR algorithm and principal component analysis |
| CN113657194A (en) * | 2021-07-27 | 2021-11-16 | 武汉理工大学 | Vehicle camera image feature extraction and matching method based on improved SURF algorithm |
| CN114387777A (en) * | 2020-10-20 | 2022-04-22 | 腾讯科技（深圳）有限公司 | Road data processing method, device, computer equipment and storage medium |
Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1941859A1 (en) * | 2006-12-18 | 2008-07-09 | ACO Hud Nordic AB | Topical formulations |
| CN102184413A (en) * | 2011-05-16 | 2011-09-14 | 浙江大华技术股份有限公司 | Automatic vehicle body color recognition method of intelligent vehicle monitoring system |
| CN103434512A (en) * | 2013-09-18 | 2013-12-11 | 武汉理工大学 | System and method for detecting lateral driving state based on smart-phone |
| CN103810491A (en) * | 2014-02-19 | 2014-05-21 | 北京工业大学 | Head posture estimation interest point detection method fusing depth and gray scale image characteristic points |
| CN104802737A (en) * | 2015-03-25 | 2015-07-29 | 清华大学 | Mobile phone based vehicle abnormality driving behavior detection method |
| CN105374211A (en) * | 2015-12-09 | 2016-03-02 | 敏驰信息科技（上海）有限公司 | System and method for calculating driving risks and assisting automobile insurance pricing based on multi-source data |
| CN106096626A (en) * | 2016-05-27 | 2016-11-09 | 大连楼兰科技股份有限公司 | The long-range loss assessment system in different automobile types subregion and method is set up based on artificial intelligence's unsupervised learning FuzzyC Means clustering algorithm |
| CN106097709A (en) * | 2016-06-27 | 2016-11-09 | 江苏迪纳数字科技股份有限公司 | Driving behavior recognition methods based on intelligent vehicle mounted terminal |
| CN106203856A (en) * | 2016-07-18 | 2016-12-07 | 交通运输部公路科学研究所 | A kind of Combined Principal Components analysis and the vehicle driving-cycle formulating method of Fuzzy c-means Clustering |
| CN107272654A (en) * | 2017-07-21 | 2017-10-20 | 南京航空航天大学 | A kind of data clusters dimension reduction method for system for flight control computer fault detect |
-
2018
- 2018-01-17 CN CN201810045607.XA patent/CN108280415A/en active Pending
Patent Citations (10)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| EP1941859A1 (en) * | 2006-12-18 | 2008-07-09 | ACO Hud Nordic AB | Topical formulations |
| CN102184413A (en) * | 2011-05-16 | 2011-09-14 | 浙江大华技术股份有限公司 | Automatic vehicle body color recognition method of intelligent vehicle monitoring system |
| CN103434512A (en) * | 2013-09-18 | 2013-12-11 | 武汉理工大学 | System and method for detecting lateral driving state based on smart-phone |
| CN103810491A (en) * | 2014-02-19 | 2014-05-21 | 北京工业大学 | Head posture estimation interest point detection method fusing depth and gray scale image characteristic points |
| CN104802737A (en) * | 2015-03-25 | 2015-07-29 | 清华大学 | Mobile phone based vehicle abnormality driving behavior detection method |
| CN105374211A (en) * | 2015-12-09 | 2016-03-02 | 敏驰信息科技（上海）有限公司 | System and method for calculating driving risks and assisting automobile insurance pricing based on multi-source data |
| CN106096626A (en) * | 2016-05-27 | 2016-11-09 | 大连楼兰科技股份有限公司 | The long-range loss assessment system in different automobile types subregion and method is set up based on artificial intelligence's unsupervised learning FuzzyC Means clustering algorithm |
| CN106097709A (en) * | 2016-06-27 | 2016-11-09 | 江苏迪纳数字科技股份有限公司 | Driving behavior recognition methods based on intelligent vehicle mounted terminal |
| CN106203856A (en) * | 2016-07-18 | 2016-12-07 | 交通运输部公路科学研究所 | A kind of Combined Principal Components analysis and the vehicle driving-cycle formulating method of Fuzzy c-means Clustering |
| CN107272654A (en) * | 2017-07-21 | 2017-10-20 | 南京航空航天大学 | A kind of data clusters dimension reduction method for system for flight control computer fault detect |
Non-Patent Citations (1)
| Title |
|---|
| 肖彪: "基于用户兴趣聚类的协同过滤算法的研究与实现", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (31)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109061706A (en) * | 2018-07-17 | 2018-12-21 | 江苏新通达电子科技股份有限公司 | A method of the vehicle drive behavioural analysis based on T-Box and real-time road map datum |
| CN109117869A (en) * | 2018-07-20 | 2019-01-01 | 汉纳森（厦门）数据股份有限公司 | User's portrait method, medium and system |
| CN109408557A (en) * | 2018-09-29 | 2019-03-01 | 东南大学 | A kind of traffic accidents reason analysis method clustered based on multiple correspondence and K-means |
| CN109408557B (en) * | 2018-09-29 | 2021-09-28 | 东南大学 | Traffic accident cause analysis method based on multiple correspondences and K-means clustering |
| CN109345772A (en) * | 2018-10-15 | 2019-02-15 | 深圳市元征科技股份有限公司 | A kind of fatigue driving recognition methods and relevant device |
| CN109708634A (en) * | 2018-12-12 | 2019-05-03 | 平安科技（深圳）有限公司 | Judge automatically method, apparatus, storage medium and the electronic equipment of driving behavior |
| CN111327755A (en) * | 2018-12-14 | 2020-06-23 | 西安博瑞集信电子科技有限公司 | Intelligent vehicle driving state identification system and method based on mobile phone |
| CN109615018A (en) * | 2018-12-24 | 2019-04-12 | 广东德诚科教有限公司 | User individual behavior evaluation method, apparatus, computer equipment and storage medium |
| CN109711459A (en) * | 2018-12-24 | 2019-05-03 | 广东德诚科教有限公司 | User individual action estimation method, apparatus, computer equipment and storage medium |
| CN109711459B (en) * | 2018-12-24 | 2019-11-15 | 广东德诚科教有限公司 | User individual action estimation method, apparatus, computer equipment and storage medium |
| CN110175557A (en) * | 2019-05-24 | 2019-08-27 | 北京交通大学 | Driver's driving behavior method for identifying and classifying and system based on driving characteristics group |
| CN110329271A (en) * | 2019-06-18 | 2019-10-15 | 北京航空航天大学杭州创新研究院 | A kind of multisensor vehicle driving detection system and method based on machine learning |
| CN110321954A (en) * | 2019-07-03 | 2019-10-11 | 中汽研（天津）汽车工程研究院有限公司 | The driving style classification and recognition methods of suitable domestic people and system |
| CN112241748A (en) * | 2019-07-16 | 2021-01-19 | 广州汽车集团股份有限公司 | Data dimension reduction method and device based on multi-source information entropy difference |
| CN110544151A (en) * | 2019-08-20 | 2019-12-06 | 北京市天元网络技术股份有限公司 | Method and equipment for determining whether user is online car booking driver |
| CN110682914A (en) * | 2019-09-29 | 2020-01-14 | 皖西学院 | Driving behavior recognition system and method based on wireless perception |
| CN111506627A (en) * | 2020-04-21 | 2020-08-07 | 成都路行通信息技术有限公司 | Target behavior clustering method and system |
| CN111800314A (en) * | 2020-07-06 | 2020-10-20 | 广元量知汇科技有限公司 | Automatic driving fault monitoring system |
| CN111881952A (en) * | 2020-07-14 | 2020-11-03 | 扬州大学 | Driver tendency analysis method based on early warning big data |
| CN112052895A (en) * | 2020-09-03 | 2020-12-08 | 同济大学 | Pure electric vehicle driving style clustering method |
| CN114387777A (en) * | 2020-10-20 | 2022-04-22 | 腾讯科技（深圳）有限公司 | Road data processing method, device, computer equipment and storage medium |
| CN112622914A (en) * | 2020-12-21 | 2021-04-09 | 武汉理工大学 | New energy automobile driving safety state identification system |
| CN112598052A (en) * | 2020-12-21 | 2021-04-02 | 中建八局第二建设有限公司 | Mechanical attitude analysis method and system based on K-Means |
| CN113642600A (en) * | 2021-06-29 | 2021-11-12 | 桂林电子科技大学 | Driving behavior feature extraction method based on mRMR algorithm and principal component analysis |
| CN113642600B (en) * | 2021-06-29 | 2024-02-02 | 桂林电子科技大学 | Driving behavior feature extraction method based on mRMR algorithm and principal component analysis |
| CN113378998A (en) * | 2021-07-12 | 2021-09-10 | 西南石油大学 | Stratum lithology while-drilling identification method based on machine learning |
| CN113378998B (en) * | 2021-07-12 | 2022-07-22 | 西南石油大学 | Stratum lithology while-drilling identification method based on machine learning |
| CN113657194A (en) * | 2021-07-27 | 2021-11-16 | 武汉理工大学 | Vehicle camera image feature extraction and matching method based on improved SURF algorithm |
| CN113657194B (en) * | 2021-07-27 | 2023-09-22 | 武汉理工大学 | Vehicle camera image feature extraction matching method based on improved SURF algorithm |
| CN113569950A (en) * | 2021-07-28 | 2021-10-29 | 大唐环境产业集团股份有限公司 | Power station equipment fault monitoring model generation method, system and device |
| CN113569950B (en) * | 2021-07-28 | 2024-05-28 | 大唐环境产业集团股份有限公司 | Power station equipment fault monitoring model generation method, system and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108280415A (en) | Driving behavior recognition methods based on intelligent mobile terminal | |
| Gao et al. | Multivariate time series prediction of lane changing behavior using deep neural network | |
| CN109978893B (en) | Training method, device, equipment and storage medium of image semantic segmentation network | |
| CN107492251B (en) | Driver identity recognition and driving state monitoring method based on machine learning and deep learning | |
| AU2017272288B2 (en) | Method, system, electronic device, and medium for classifying license plates based on deep learning | |
| CN110119714B (en) | Driver fatigue detection method and device based on convolutional neural network | |
| US9664779B2 (en) | Object classification for vehicle radar systems | |
| CN106384100A (en) | Component-based fine vehicle model recognition method | |
| CN110329271A (en) | A kind of multisensor vehicle driving detection system and method based on machine learning | |
| KR20180037192A (en) | Detection of unknown classes and initialization of classifiers for unknown classes | |
| CN109800682B (en) | Driver attribute identification method and related product | |
| WO2018009552A1 (en) | System and method for image analysis | |
| CN108694408B (en) | Driving behavior recognition method based on deep sparse filtering convolutional neural network | |
| CN107657279A (en) | A kind of remote sensing target detection method based on a small amount of sample | |
| Yan et al. | Driving posture recognition by joint application of motion history image and pyramid histogram of oriented gradients | |
| CN108960189A (en) | Image recognition methods, device and electronic equipment again | |
| CN109635784A (en) | Traffic sign recognition method based on improved convolutional neural networks | |
| CN110263836B (en) | Bad driving state identification method based on multi-feature convolutional neural network | |
| Cai et al. | A driving fingerprint map method of driving characteristic representation for driver identification | |
| CN115617217A (en) | Vehicle state display method, device, equipment and readable storage medium | |
| Chu et al. | A review of driving style recognition methods from short-term and long-term perspectives | |
| CN104751197A (en) | Device and method for recognizing faces of drivers during vehicle running on basis of video analysis | |
| CN106777350A (en) | It is a kind of based on bayonet socket data scheming to search drawing method and device | |
| CN110334572A (en) | The fine recognition methods of vehicle under a kind of multi-angle | |
| CN116946183A (en) | Commercial vehicle driving behavior prediction method considering driving capability and vehicle equipment |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20180713 |