CN111680561B - Driving behavior habit analysis system - Google Patents
Driving behavior habit analysis system Download PDFInfo
- Publication number
- CN111680561B CN111680561B CN202010384266.6A CN202010384266A CN111680561B CN 111680561 B CN111680561 B CN 111680561B CN 202010384266 A CN202010384266 A CN 202010384266A CN 111680561 B CN111680561 B CN 111680561B
- Authority
- CN
- China
- Prior art keywords
- driver
- driving
- module
- driving behavior
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V40/00—Recognition of biometric, human-related or animal-related patterns in image or video data
- G06V40/20—Movements or behaviour, e.g. gesture recognition
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V20/00—Scenes; Scene-specific elements
- G06V20/40—Scenes; Scene-specific elements in video content
- G06V20/41—Higher-level, semantic clustering, classification or understanding of video scenes, e.g. detection, labelling or Markovian modelling of sport events or news items
-
- 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
- G06V20/52—Surveillance or monitoring of activities, e.g. for recognising suspicious objects
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/01—Protocols
- H04L67/12—Protocols specially adapted for proprietary or special-purpose networking environments, e.g. medical networks, sensor networks, networks in vehicles or remote metering networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L67/00—Network arrangements or protocols for supporting network services or applications
- H04L67/50—Network services
- H04L67/55—Push-based network services
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
- H04W4/30—Services specially adapted for particular environments, situations or purposes
- H04W4/40—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P]
- H04W4/44—Services specially adapted for particular environments, situations or purposes for vehicles, e.g. vehicle-to-pedestrians [V2P] for communication between vehicles and infrastructures, e.g. vehicle-to-cloud [V2C] or vehicle-to-home [V2H]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Multimedia (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Software Systems (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Computational Linguistics (AREA)
- Psychiatry (AREA)
- Social Psychology (AREA)
- Human Computer Interaction (AREA)
- Computing Systems (AREA)
- Medical Informatics (AREA)
- Traffic Control Systems (AREA)
Abstract
A driving behavior habit analysis system relates to the technical field of automobiles. The driving behavior habit analysis system comprises an image acquisition module, a data processing module, a data uploading module, a data analysis module and an information pushing module. The image acquisition module can acquire the driving behavior information of the driver in real time. The data processing module can receive the driving behavior information, record the occurrence frequency of abnormal behaviors and store data according to corresponding categories. The data uploading module can upload the stored data to the data analysis module to determine the quality grade. And finally, pushing the driving behavior statistical information to a single user by an information pushing module. According to the method, the mode of pushing the driving behavior information is adopted to help the user to know the bad driving behavior habit of the user and push different contents and types of pushtext for the user according to different analysis results, so that the good driving habit of the user is helped to be developed, driving habits are improved, user experience is improved, and the traffic accident occurrence rate is reduced.
Description
Technical Field
The invention belongs to the technical field of advanced auxiliary driving systems of automobiles, and relates to a driving behavior habit analysis system.
Background
The rapid development of the traffic road brings great convenience to the work and life of people, and promotes the social development and the prosperity and stability. But the frequent occurrence of traffic accidents also brings great troubles and disasters to people. The abnormal driving posture of the driver needs to be paid sufficient attention as one of the main causes of traffic accidents.
The working modes of the driver monitoring system equipped for the automobile in the current market are real-time detection and early warning prompt. Chinese patent application No. CN108229345A, the patent name is "a driver detection system", inventor Lu Minggang, liu Wei, qiu Congyu, etc. discloses a driver detection system, the system can detect the operation intention and driving state of the driver, and give out early warning prompt in real time; chinese patent application No. CN109910901a, entitled "an intelligent driver-assisted system with driver behavior analysis and monitoring function", inventor Chen Yue, et al disclose a driver driving behavior assisted monitoring system, which can capture the driver's actions, collect the driving state of the vehicle, and give feedback to control and remind the driver in time, so as to prevent accidents. The two patents can monitor abnormal information of a driver or a vehicle and help the driver to be aware of possible dangers by adopting an early warning prompt mode, but the serious influence of frequent early warning prompts of a system on driving experience is not considered, and the driver is very easy to have a dysphoric mood. In addition, the bad driving habits of the driver cannot be changed from the root through single early warning prompt, and the traffic accident occurrence rate cannot be reduced fundamentally.
Therefore, in terms of product practicability, the current products cannot satisfy consumers; on the social significance level, the effect of reducing the traffic accident rate is not obvious.
Disclosure of Invention
The invention aims to solve the problems that an existing driver monitoring system is poor in practicability and user experience, and bad driving habits of a driver cannot be changed fundamentally, and provides a driver driving habit analysis system. Namely, a data uploading module, a data analyzing module and a pushing module are added on the basis of the original driver monitoring system. The invention reduces the influence of frequent early warning prompt of the existing driver monitoring system on the driving experience, increases the user experience, helps the driver to further know and improve the driving habit of the driver, and fundamentally reduces the occurrence rate of traffic accidents.
In order to achieve the purpose, the invention adopts the technical scheme that:
the invention provides a driving behavior habit analysis system, which is used for helping a driver to develop good driving habits and improve driving habits. The existing driver monitoring system component mainly comprises an image acquisition module and a data processing module.
The image acquisition module is used for acquiring the driving behavior information of a driver, and the driving behavior information mainly comprises: driver facial information, driver hand motion, driver body posture, 3D object information.
The data processing module is used for receiving and classifying the driving behaviors of the driver, recording the occurrence times, storing data in a classified mode according to the types of the abnormal driving behaviors, and specifically realizing the following functions: whether the driver is in a fatigue driving state can be judged through the facial information and the body posture of the driver; whether the driver smokes, makes a call or drinks can be judged according to the hand action of the driver, the body posture of the driver and the 3D object information; whether the driver fastens the safety belt or not can be judged according to the 3D object information; whether the driver is in a distracted driving state or not can be judged according to the facial information of the driver; whether the driver is in an irregular driving posture or not can be judged according to the body posture of the driver; the driver driving behavior information can be received and classified, the occurrence frequency of abnormal driving behaviors is recorded, and the analyzed data is stored in the local database in a classified mode. The abnormal driving behavior includes: smoking, making a call, drinking, unbelted, tired driving, distracted driving, and non-normative driving postures that occur during driving.
And the data uploading module is used for uploading the stored abnormal driving behavior classification data.
And the data analysis module is used for determining the excellent level of the driving habits of the driver. The determination of the excellent level of the driving habits of the driver is calculated according to the following method:
1) First, a full sample set D = ((t, x) of driver driving behavior statistics collected over a fixed period of time for a vehicle equipped with a driver monitoring system is obtained 1 ),(t,x 2 ),…,(t,x m ))Wherein x is 1 ,x 2 ,…,x m The times of abnormal driving behaviors of different individual drivers in the set are respectively represented, t is a fixed time period, and the unit can be set to year, month and week.
2) According to the specific sample size in the complete sample set D, setting the appropriate minimum sample point number (MinPts) and eps (epsilon neighborhood) distance and defining three types of data points:
core point: if sample x i The epsilon field of (c) contains at least the minimum number of sample points (MinPts) samples, i.e., N ε (X i ) Not less than MinPts, sample point x is called i Is the core point.
Boundary points are as follows: if sample x i Is smaller than MinPts but is in the neighborhood of other core points, sample point x is said to be i Are boundary points.
Noise points: i.e. points that are neither core points nor boundary points.
3) Initializing a core object set omega = phi, initializing a cluster number k =0, initializing an unvisited sample set Γ = D, and clustering C = phi.
4) Randomly selecting a sample point (t, x) from the complete sample set D obtained in the step 1) i ) Then, all points with the distance less than or equal to the eps (epsilon neighborhood) distance are found and the category of the points is judged. If this point is a core point it is assigned a new cluster label. The manner of calculating the eps distance between two points uses the euclidean distance:
wherein, the coordinates of any two points in the A and B planes are respectively (a) 1 ,a 2 ) And (b) 1 ,b 2 )。
5) All neighbors of the point within eps are visited. If they have not been assigned a cluster, then the new cluster labels created in step 3) are assigned to them. If they are core samples, then their neighbors are visited in turn, and so on. The cluster is gradually increased until there are no more core samples within the eps distance of the cluster.
6) Another point is selected that has not been visited and the same process is repeated.
7) After clustering is finished, a final cluster sample set Y = { c is obtained 1 ,c 2 ,…,c n In which c is 1 ,c 2 ,…,c n Respectively, representing different cluster samples.
8) Repeating the steps 1-7, and maintaining 5 elements in the final cluster sample set Y, wherein the elements correspond to five danger levels: excellent, good, standard, dangerous and high-risk. The specific range of each hazard class is determined by observing the distribution of the clustered samples.
The information pushing module is configured to be capable of periodically pushing the abnormal driving behavior statistical information of the driver to a single user and pushing text with different contents and types to the user according to different analysis statistical results. The pushing mode can be carried out according to the following method:
1) Setting all content sets to be recommended in the information pushing module as T = { T = { T } 1 ,t 2 ,…,t n H, where t 1 ,t 2 ,…,t n Respectively representing different recommended contents. All occurring in the recommended content set of words (dictionary) is V = { V = { V } 1 ,v 2 ,…,v n In which v is 1 ,v 2 ,…,v n The dictionaries corresponding to different recommended contents are respectively represented, that is, the total number of all the recommended contents is N, and N different words are contained in the recommended contents.
2) Representing jth recommended content as T by using vector j ＝(ω 1j ,ω 2j ,…,ω nj ) Wherein ω is 1j Denotes the first word v 1 In the article j, a larger value indicates more importance.
3) Computing ω using word frequency-inverse document frequency (TF-IDF) k,j The value of (c). The TF-IDF corresponding to the kth in the dictionary in the jth recommendation is:
wherein, TF (t) k ,d j ) Indicates the number of times of occurrence of the k-th word in the recommended content j, n k Indicating the number of recommended contents including the k-th word in all the articles. The final weight of the kth word in the recommended content j is:
4) Combining the risk level set Y obtained by clustering to obtain a recommended content data set Z = { (r) 1 ,y i ),(r 2 ,y i ),…,(r n ,y i ) In which r is i e.T is a feature vector of the recommended content, y i ∈Y＝{c 1 ,c 2 ,…,c n I =1,2, …, N, a single risk level may correspond to multiple recommended contents.
5) Recommending recommended contents under corresponding danger levels for users according to given danger levels of drivers, wherein the priority of the recommended contents is determined by word frequency-inverse document frequency, and omega k,j The larger the value of (d), the more advanced the ranking of the recommended content.
Furthermore, the image acquisition module is configured to be a double-camera imaging mode, and the camera adopts an infrared camera or a color camera to obtain complete 3D information of a driver and surrounding objects and improve the recognition rate of abnormal driving behaviors.
Further, the data processing module is configured to process more than or equal to 30 frames of video stream data, and the video stream data comes from the image acquisition module.
Further, the data uploading module configures a 5G high-speed network to ensure timeliness and integrity of data uploading within a specific time period.
The beneficial effects of the invention are as follows: the influence of frequent early warning prompt of the existing driver monitoring system on driving experience can be effectively reduced, the driving experience of a user is improved, a driver is effectively helped to know and improve driving habits of the driver, and the traffic accident rate is reduced from the root.
The above objects, advantages and features of the present invention will become more apparent to those skilled in the art from the following detailed description of specific embodiments thereof, taken in conjunction with the accompanying drawings.
Drawings
FIG. 1 is a schematic flow diagram of a driver driving habit analysis system designed according to a specific embodiment of the invention.
Fig. 2 is a basic architecture of the present invention, and the specific architecture composition sequence is: the system comprises an image acquisition module 1, a data processing module 2, a data uploading module 3, a data analysis module 4 and an information pushing module 5.
Detailed Description
A specific embodiment of the present invention will now be described in detail by way of example and not by way of limitation, with reference to the accompanying drawings.
Currently, driver monitoring systems are already available on high-end branded vehicles. The system has the functions of detecting abnormal driving states such as fatigue driving detection, smoking, drinking, calling, distracted driving, unbuckled safety belts and the like. The direct action mode is generally that when the system detects that the driver is in an abnormal driving state, the system reminds the user to keep a good driving state through a real-time early warning prompting mode. The function mode has the advantages that when the abnormal driving posture occurs to the driver, the system can remind the driver to remove the abnormal driving state in time through one or more of visual alarm, auditory alarm or tactile alarm, and the purpose of preventing traffic accidents is achieved. But the disadvantage is that the frequent early warning prompt easily causes the driver to have a fussy mood to influence the driving experience and cannot radically cause the driver to be aware of the danger of the behaviors so as to improve the driving habit. In order to solve the above problems, the applicant has devised a driver driving habit analysis system.
FIG. 1 is a flow chart of a driver driving habit analysis system designed according to an embodiment of the invention. Fig. 2 shows a system for analyzing the driving behavior of the driver and pushing statistical information, which mainly comprises an image acquisition module 1, a data processing module 2, a data uploading module 3, a data analysis module 4 and an information pushing module 5.
The image acquisition module 1 is used for acquiring driving behavior information of a driver and comprises two cameras and a light supplementing device. Two cameras can all be infrared cameras or all be colored cameras, can also use two kinds of cameras jointly. The light supplement device can adopt an infrared light supplement device or other devices known by those skilled in the art. The double-camera system has the advantages that the double-camera system can enable the image acquisition module to observe an object from two points, and acquire depth information of the object at different visual angles, so that the comprehensiveness and integrity of acquired data are guaranteed as far as possible, and the identification accuracy of the data processing module is improved. The image acquisition module transmits the recorded driving behavior data to the data processing module 2.
The data processing module 2 is configured with a driver posture classifier, can correctly classify different types of abnormal driving behavior data information and record and store the result in a database built in an on-vehicle Electronic Control Unit (ECU).
The types of abnormal driving behavior data in this example include: smoking, drinking, making a call, fatigue driving, distracted driving, not fastening a safety belt. The image acquisition module and the data processing module collect that the user has smoking behavior 87 times, drinking behavior 34 times, making a call 129 times, fatigue driving 3 times, distraction driving 48 times and no safety belt fastening 9 times in a driving time period of one month, namely the driver has abnormal driving behavior 310 times in one month. The cloud database stores other 1000 drivers' abnormal driving behavior samples collected within one month.
The data analysis module 4 divides the driving habit danger level of the user according to a clustering algorithm, and can determine the level by using the following clustering algorithm:
1) The complete sample set of abnormal driving behaviors of 1000 drivers in the database in one month is D = (x) 1 ,x 2 ,…,x 1000 ) The sample metric parameter is (ε, minPts), where x 1 ,x 2 ,…,x 1000 The statistical information of the driving behaviors of 1000 drivers collected in one month by a vehicle equipped with a driver system, and each driver has abnormal driving behaviors for a certain number of times.
2) Initializing a core object set omega = phi, initializing the number of cluster clusters k =0, initializing an unaccessed sample set Γ = D, and partitioning the clusters C = phi.
3) Finding all core objects for the sample set D according to the following steps:
a) By means of distance measurement, find x j Epsilon-neighborhood subsample set N ε (x j )。
b) If the number of the sub-sample set samples satisfies | N ε (x j ) | ≧ MinPts, sample x j Adding a core object sample set: Ω = Ω & { x + j }。
4) If the core object is combined with Ω = Φ, the iteration ends, otherwise go to step (5).
5) Randomly selecting one core object o in a core object set omega, and initializing a current cluster core object queue omega cur = omicron, initialize class number k = k +1, initialize the current cluster sample set C k -o, updating the set of unaccessed samples Γ = Γ - { o }.
6) If the current cluster core object queue omega cur = phi, then the current cluster C is clustered k After the generation is finished, the cluster is divided into C = { C = 1 ,C 2 ,…,C k H, updating a core object set omega = omega-C k And (5) turning to the step (4). Otherwise, updating the core object set omega = omega-C k 。
7) In the current cluster core object queue omega cur Taking out a core object o', and finding out all epsilon neighborhood subsample sets N through a neighborhood distance threshold epsilon ε (o') making Δ = N ε (o') # Γ, update the current cluster sample set C k ＝C k And U delta, updating an unvisited sample set gamma = gamma-delta and updating omega cur ＝Ω cur U (. DELTA.andgate. OMEGA) -proceeds to step (6).
8) Finally, after the iteration is finished, clustering division Y = C = { C ] is obtained 1 ,C 2 ,…,C 5 And obtaining five clustering results corresponding to five danger grades, namely excellent, good, standard, dangerous and high-risk. Wherein the corresponding range of each clustering result is C 1 ∈(0,137),C 2 ∈(138,248),C 3 ∈(249,423),C 4 ∈(424,534),C 5 ∈(535,739)。
9) And referring to the clustering result, carrying out fuzzy processing, and finally determining the frequency range of abnormal driving behaviors corresponding to each grade as follows:
excellent- (0, 100), good- (101-250), standard- (251-400), dangerous- (401-550) and high-risk- (500 +).
The data uploading module is configured with a high-speed 5G transmission network and can upload information in a database built in the ECU to the data analysis module 4. The 5G network is configured to ensure the timeliness and integrity of the uploaded information.
The information pushing module 5 can push recommended contents of different types and contents for a single client according to the driving behavior statistical information of different danger degrees. The recommended content is from a network multimedia library and mainly comprises: short video class, short text class and news class. The following recommendation algorithm may be used for recommending content for the user:
1) Let T = { T) be the set of all recommended content in the multimedia library 1 ,t 2 ,…,t n V = { V } dictionary 1 ,v 2 ,…,v n }, the recommended content vector is T j ＝(ω 1j ,ω 2j ,…,ω nj )
2) Calculating word frequency-inverse document frequency (TF-IDF):
calculating the weight:
and combining the five driving habit danger levels to obtain input data.
3) Combining the risk level Y obtained by clustering to obtain a recommended content data set Z = { (x) 1 ,y i ),(x 2 ,y i ),…,(x n ,y i ) In which x i e.T is the feature vector of the recommended content, y i ∈Y＝{c 1 ,c 2 ,…,c 5 I =1,2,3,4,5.
4) Recommending recommended contents under corresponding danger levels for the user according to given danger levels of the driver, wherein the priority of the recommended contents is determined by word frequency-inverse document frequency, and omega k,j The larger the value of (d), the more advanced the ranking of the recommended content.
5) The driver has 310 times of abnormal driving behaviors in one month, and all c are pushed to the user according to the standard grade of the danger grade range 3 Corresponding content at the level. Push sequence is from omega k,j The value of (d) determines that the larger the value, the earlier the ranking of the recommended content.
The driving habit analyzing system of the driver can continuously monitor all behaviors of the driver in the driving interval. The driving behavior habit statistical data after being evaluated for the period every day, every month and every year and the pushing texts of corresponding types can be pushed to the user through the vehicle machine, the mobile phone, the computer, the tablet and the intelligent wearable product, so that the user can fully know the driving behavior habit problem and the severity existing in the user.
The recommended content mainly comprises: the method comprises the steps of traffic accident short video, traffic accident reason analysis short texts, driving habit formation short texts, home and abroad traffic accident news and the like, and the contents are customized for clients with different danger levels. The method aims to provide a psychological shadow-like effect for the driver, so that the driver can actively keep a good driving state. For example: when a high-risk user subconsciously makes an abnormal driving behavior in the driving process, the serious consequences caused by poor driving habits can be recalled, and the user is worried about the traffic accident, so that the normal driving state is actively recovered.
The driving habit analysis system for the driver adopts the image acquisition module 1 to acquire driving behavior information, the data processing module 2 classifies the data and stores the data in the ECU built-in database, the data is completely and timely uploaded to the cloud server through the data uploading module 3, secondary processing is performed through the data analysis module 4, and finally the driving behavior statistical analysis result and the corresponding recommended content of each grade are pushed to the user through the information pushing module 5. When guaranteeing driver monitoring system normal operating, help the user to know self driving behavior habit, solve the traffic accident problem that causes because of driving the gesture is unusual from the root, solved frequent early warning suggestion and made the user produce the problem of irritability mood, promoted user experience.
The above-mentioned embodiments only represent the embodiments of the present invention, but they should not be understood as the limitation of the scope of the present invention, and it should be noted that those skilled in the art can make several variations and modifications without departing from the spirit of the present invention, and these all fall into the protection scope of the present invention.
Claims (2)
1. A driving behavior habit analysis system is characterized in that a data uploading module, a data analysis module and an information pushing module are added in an existing driver monitoring system component;
the existing driver monitoring system component mainly comprises an image acquisition module and a data processing module; the image acquisition module is used for acquiring driving behavior information of a driver, wherein the driving behavior information mainly comprises: the system comprises driver face information, driver hand actions, driver body postures and 3D object information; the data processing module is used for receiving and classifying the driving behaviors of the driver, recording the occurrence times of abnormal driving behaviors and storing data in a classified manner according to the types of the abnormal driving behaviors; the data uploading module is used for uploading the stored abnormal driving behavior classification data;
the data analysis module is used for determining the excellent level of the driving habits of the driver; the determination of the good level of the driving habits of the driver is calculated according to the following method:
1) First, a full sample set D = ((t, x) of driver driving behavior statistics collected over a fixed period of time for a vehicle equipped with a driver monitoring system is obtained 1 ),(t,x 2 ),…,(t,x m ) Wherein x is 1 ,x 2 ,…,x m Respectively represents abnormal driving behaviors of different driver individuals in the setThe number of times, t is a fixed time period, and the unit can be set as year, month or week;
2) According to the specific sample size in the complete sample set D, setting the appropriate minimum sample point number MinPts and eps distance and defining three types of data points, wherein the eps distance represents an epsilon neighborhood:
core point: if sample x i The epsilon field of (1) contains at least the minimum number of sample points MinPts samples, i.e. N ε (x i ) Not less than MinPts, sample point x is called i Is a core point;
boundary points are as follows: if sample x i Is smaller than MinPts, but is in the neighborhood of other core points, sample point x is called i Is a boundary point;
noise point: i.e. points that are neither core points nor boundary points;
3) Initializing a core object set omega = phi, initializing the number of clustering clusters k =0, initializing an unvisited sample set gamma = D, and clustering C = phi;
4) Randomly selecting a sample point (t, x) from the complete sample set D obtained in the step 1) i ) Then finding all points with the point distance less than or equal to the eps distance and judging the category of the sample point; if the point is a core point, a new cluster label is allocated to the point; the manner of calculating the eps distance between two points uses the euclidean distance:
wherein, the coordinates of any two points in the A and B planes are respectively (a) 1 ,a 2 ) And (b) 1 ,b 2 )；
5) Accessing all neighbors of the point within eps; assigning the new cluster labels created in step 4) to them if they have not already been assigned a cluster; if they are core samples, then their neighbors are visited in turn, and so on; the cluster is gradually increased until there are no more core samples within the eps distance of the cluster;
6) Selecting another point which is not accessed yet, and repeating the same process;
7) After clustering is finished, a final cluster sample set Y = { c is obtained 1 ,c 2 ,…,c 5 In which c is 1 ,c 2 ,…,c 5 Respectively representing different cluster samples;
8) Repeating the steps 1-7, and maintaining 5 elements in the final cluster sample set Y, wherein the elements respectively correspond to five danger levels: excellent, good, standard, dangerous and high-risk; determining the specific range of each danger grade by observing the distribution of the clustered samples;
the information pushing module is configured to periodically push the abnormal driving behavior statistical information of the driver to a single user, and push text with different contents and types to the user according to different analysis statistical results; the pushing mode can be carried out according to the following method:
1) Setting all content sets to be recommended in the information pushing module as T = { T = { T } 1 ,t 2 ,…,t N Where t is 1 ,t 2 ,…,t N Respectively representing different recommended contents; the set of words appearing in all recommended contents is V = { V = } 1 ,v 2 ,…,v N H, wherein v 1 ,v 2 ,…,v N Respectively representing dictionaries corresponding to different recommended contents, namely the total number of all the recommended contents is N, and N different words are contained in the recommended contents;
2) Representing jth recommended content as T by using vector j ＝(ω 1j ,ω 2j ,…,ω Nj ) Wherein ω is 1j Denotes the first word v 1 Weight in article j, ω Nj Denotes the nth word v N The weight in the article j is more important when the weight value is larger;
3) Computing ω using word frequency-inverse document frequency (TF-IDF) kj A value of (d); the TF-IDF corresponding to the kth word in the dictionary in the jth recommendation is:
wherein, TF (t) j ,v k ) Indicates the number of times of occurrence of the k-th word in the recommended content j, n k Representing the number of recommended contents including the kth word in all the articles, wherein N represents the total number of the recommended contents; the final weight of the kth word in the recommended content j is:
wherein N represents the total number of words contained in the recommended content;
4) Combining the risk level set Y obtained by clustering to obtain a recommended content data set Z = { (r) 1 ,y i ),(r 2 ,y i ),…,(r n ,y i ) In which r is i e.T is a feature vector of the recommended content, y i ∈Y＝{c 1 ,c 2 ,…,c 5 The risk level is set, and a single risk level can correspond to a plurality of recommended contents;
5) Recommending recommended contents under corresponding danger levels for users according to given danger levels of drivers and feature vectors of recommended contents corresponding to the danger levels, wherein the priority of the recommended contents is determined by word frequency-inverse document frequency, and omega kj The larger the value of (d), the more advanced the ranking of the recommended content.
2. The driving behavior habit analysis system according to claim 1, wherein the data processing module specifically implements the following functions: whether the driver is in a fatigue driving state can be judged through the facial information and the body posture of the driver; whether the driver smokes, makes a call or drinks can be judged according to the hand motion of the driver, the body posture of the driver and the 3D object information; whether the driver fastens the safety belt or not can be judged according to the 3D object information; whether the driver is in a distracted driving state or not can be judged according to the facial information of the driver; whether the driver is in the non-standard driving posture or not can be judged according to the body posture of the driver; the system can receive and classify the driving behavior information of the driver, record the occurrence frequency of abnormal driving behavior, and store the analyzed data in a local database in a classified manner; the abnormal driving behavior includes: smoking, making a call, drinking, unbelted, tired driving, distracted driving, and non-normative driving postures that occur during driving.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010384266.6A CN111680561B (en) | 2020-05-09 | 2020-05-09 | Driving behavior habit analysis system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010384266.6A CN111680561B (en) | 2020-05-09 | 2020-05-09 | Driving behavior habit analysis system |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111680561A CN111680561A (en) | 2020-09-18 |
| CN111680561B true CN111680561B (en) | 2023-04-14 |
Family
ID=72451799
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010384266.6A Active CN111680561B (en) | 2020-05-09 | 2020-05-09 | Driving behavior habit analysis system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111680561B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112609765A (en) * | 2020-11-18 | 2021-04-06 | 徐州徐工挖掘机械有限公司 | Excavator safety control method and system based on facial recognition |
| CN113053083A (en) * | 2021-03-22 | 2021-06-29 | 重庆长安汽车股份有限公司 | Early warning method and system for dangerous driving vehicle based on V2X |
| CN113119687B (en) * | 2021-05-07 | 2022-12-23 | 戴凯丽 | Automatic control method for working modes of air and air conditioner in automobile |
| CN113232670B (en) * | 2021-06-15 | 2022-06-10 | 杭州链驾科技有限公司 | Driving behavior analysis method based on block chain |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2011107978A (en) * | 2009-11-17 | 2011-06-02 | Fujitsu Ten Ltd | Information processor, in-vehicle device, information processing system, information processing method, and program |
| CN104935659A (en) * | 2015-06-17 | 2015-09-23 | 大连理工大学 | Access algorithm based on service quality perception in vehicle area network |
| CN110516746A (en) * | 2019-08-29 | 2019-11-29 | 吉林大学 | A kind of driver's follow the bus behavior genre classification method based on no label data |
Family Cites Families (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP6558719B2 (en) * | 2017-03-29 | 2019-08-14 | マツダ株式会社 | Vehicle driving support system and vehicle driving support method |
| US11023752B2 (en) * | 2018-08-29 | 2021-06-01 | Here Global B.V. | Method and system for learning about road signs using hierarchical clustering |
-
2020
- 2020-05-09 CN CN202010384266.6A patent/CN111680561B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2011107978A (en) * | 2009-11-17 | 2011-06-02 | Fujitsu Ten Ltd | Information processor, in-vehicle device, information processing system, information processing method, and program |
| CN104935659A (en) * | 2015-06-17 | 2015-09-23 | 大连理工大学 | Access algorithm based on service quality perception in vehicle area network |
| CN110516746A (en) * | 2019-08-29 | 2019-11-29 | 吉林大学 | A kind of driver's follow the bus behavior genre classification method based on no label data |
Non-Patent Citations (2)
| Title |
|---|
| 张林兵.基于多维行为分析的用户聚类方法研究.《电子科技大学学报》.2020,第第49卷卷(第第49卷期),全文. * |
| 苗星至.基于GPS数据的出租车需求热点分析与寻客驾驶方案推荐研究.《中国优秀硕士论文电子期刊网》.2020,26-30. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111680561A (en) | 2020-09-18 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111680561B (en) | Driving behavior habit analysis system | |
| Yi et al. | A machine learning based personalized system for driving state recognition | |
| US20200151479A1 (en) | Method and apparatus for providing driver information via audio and video metadata extraction | |
| US20200057487A1 (en) | Methods and systems for using artificial intelligence to evaluate, correct, and monitor user attentiveness | |
| Schwarz et al. | The detection of drowsiness using a driver monitoring system | |
| US10089692B1 (en) | Risk evaluation based on vehicle operator behavior | |
| Seshadri et al. | Driver cell phone usage detection on strategic highway research program (SHRP2) face view videos | |
| KR20200063193A (en) | Driving management method and system, in-vehicle intelligent system, electronic device, medium | |
| Wali et al. | The relationship between driving volatility in time to collision and crash-injury severity in a naturalistic driving environment | |
| Yan et al. | Driving posture recognition by joint application of motion history image and pyramid histogram of oriented gradients | |
| US10783360B1 (en) | Apparatuses, systems and methods for vehicle operator gesture recognition and transmission of related gesture data | |
| Morando et al. | A model for naturalistic glance behavior around Tesla Autopilot disengagements | |
| CN110991999A (en) | Method and device for improving law enforcement amount cutting efficiency, computer equipment and storage medium | |
| Zhao et al. | Driver posture monitoring in highly automated vehicles using pressure measurement | |
| CN113312958B (en) | Method and device for adjusting dispatch priority based on driver state | |
| Kielty et al. | Neuromorphic driver monitoring systems: A proof-of-concept for yawn detection and seatbelt state detection using an event camera | |
| Shukla et al. | An Efficient Approach of Face Detection and Prediction of Drowsiness Using SVM | |
| Rusmin et al. | Design and implementation of driver drowsiness detection system on digitalized driver system | |
| CN114973214A (en) | Unsafe driving behavior identification method based on face characteristic points | |
| Reed et al. | Upper Extremity Postures and Activities in Naturalistic Driving | |
| Shamsuddin et al. | Eye detection for drowsy driver using artificial neural network | |
| Ahmad et al. | Machine learning approaches for detecting driver drowsiness: a critical review | |
| Shao et al. | Adaptive forward collision warning system for hazmat truck drivers: considering differential driving behavior and risk levels | |
| Hwang et al. | A study on discriminating risky driving using the psychological characteristics and attitudes for providing a personalized driving environment | |
| Miller et al. | Improving Methods to Measure Attentiveness through Driver Monitoring |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |