CN109635852B - User portrait construction and clustering method based on multi-dimensional attributes - Google Patents
User portrait construction and clustering method based on multi-dimensional attributes Download PDFInfo
- Publication number
- CN109635852B CN109635852B CN201811414495.7A CN201811414495A CN109635852B CN 109635852 B CN109635852 B CN 109635852B CN 201811414495 A CN201811414495 A CN 201811414495A CN 109635852 B CN109635852 B CN 109635852B
- Authority
- CN
- China
- Prior art keywords
- driving
- driver
- value
- vehicle speed
- average
- 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
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
Landscapes
- Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Traffic Control Systems (AREA)
Abstract
The invention discloses a user portrait construction and clustering method based on multi-dimensional attributes, which comprises the following steps: preprocessing the original data and extracting driving behavior characteristics of a driver; the method comprises the steps of screening features by using a maximum correlation minimum redundancy feature selection method based on mutual information, gathering driver groups with similar driving behaviors into a class through the similarity of the driving behaviors of the drivers and a network structure analysis technology, carrying out user portrayal on the driving preference of the drivers from different dimensionalities by adopting a radar map, and finally analyzing factors influencing the driving behaviors. The invention can effectively reflect the driving behavior preference of the driver, can lead the driver to know the driving habit which needs to be improved, and can lead the public transport company to train the driver in a targeted way.
Description
Technical Field
The invention relates to the technical field of driver behavior detection, in particular to a user portrait construction and clustering method based on multi-dimensional attributes.
Background
In recent years, with the rapid increase of national economy, urban public transport plays a great role in blockage relieving and smoothness keeping, and is more and more popular with governments and the public. In urban traffic environment, the driving behaviors of the bus driver, such as rapid acceleration, rapid braking and driving, seriously affect the driving safety and the comfort level of passengers. Therefore, the driving behavior of the bus driver is standardized, the service quality and the safety of the bus are ensured, and the importance is placed. At present, an effective method for detecting the behavior of a driver in a bus is not available.
Disclosure of Invention
Aiming at the defects of the prior art, the invention aims to provide a user portrait construction and clustering method based on multi-dimensional attributes.
In order to achieve the purpose, the invention adopts the following technical scheme:
a user portrait construction and clustering method based on multi-dimensional attributes comprises the following steps:
s1, acquiring corresponding fields of description information in complete driving data of each trip in the bus, and sorting the data of each field into csv documents;
s2, primarily extracting driving behavior characteristics related to driving behaviors from the csv document obtained in the step S1;
s3, according to the driving behavior characteristics extracted in the step S2, the characteristics with low distinguishing degree are deleted according to the characteristic probability distribution condition, and more representative characteristics are obtained through screening;
s4, sorting the features screened in the step S3 by using a maximum correlation minimum redundancy feature selection method based on mutual information, setting a screening threshold value and deleting the features of which the average mutual information value is smaller than the screening threshold value; determining the weight of the remaining features after screening according to the size of the average mutual information value, namely setting the average mutual information value of each feature as the weight value of the feature;
s5, weighting and normalizing the remaining characteristics after being screened in the step S4 by using the weight value of each characteristic obtained in the step S4;
s6, measuring the similarity of the driving behavior characteristics weighted and normalized in the step S5 among the driving data of each trip; setting a similarity threshold, and establishing a connecting edge when the similarity of the driving behavior characteristics among the driving data of each trip is greater than the similarity threshold;
s7, dividing the driving behaviors of each trip into different classes according to the similarity of the driving behavior characteristics among the driving data of each trip by adopting a Fast Unfolding algorithm;
s8, comprehensively analyzing the driving habits and preferences of the driver from three dimensions of driving unevenness, braking preference and vehicle speed preference in the form of a radar map:
the driving instability dimension comprises three characteristics of a vehicle speed standard deviation, an acceleration standard deviation and an accelerator pedal percentage standard deviation; the brake preference includes two characteristics of electronic brake use probability and service brake use probability; the vehicle speed preference comprises four characteristics of a vehicle speed average value, a vehicle speed median, an acceleration absolute value average value and an accelerator pedal percentage average value; when the driver is comprehensively analyzed in three dimensions, firstly, the values of all the characteristics are normalized, then the values of the characteristics contained in each dimension are summed and normalized, finally, the scores of the driver in the three dimensions are obtained, and the scores are visually displayed in a radar map form.
Further, in step S1, the driving data of the bus is recorded by using a CAN bus onboard instrument.
Further, in step S2, the driving behavior characteristics include a median of vehicle speed, a standard deviation of vehicle speed, a mean of vehicle speed, a percentage standard deviation of accelerator pedal, a percentage mean of accelerator pedal, a sliding probability in neutral state, a standard deviation of acceleration, a usage probability of electronic brake, a usage probability of foot brake, a probability of pulling a hand brake during driving, and an absolute value of acceleration greater than 2m/S2Probability, average acceleration value, mode of vehicle speed, median of accelerator pedal percentage, mode of accelerator pedal percentage, electronic brake recording times, foot brake recording times, and absolute acceleration value greater than 2m/s2The number of times of recording and the number of times of idling are in proportion.
Further, in step S6, a cosine similarity is calculated for the weighted and normalized driving behavior characteristics between the driving data of each trip, and the cosine similarity is compared with a similarity threshold.
Further, in step S7, the method further includes calculating an average classification accuracy of Fast Unfolding algorithm clusters according to the following formula:
wherein p iscTo average classification accuracy, niIs the total number of laps traveled by driver i,
is the CiThe number of driving trips of the driver i in the class,
the maximum number of runs of driver i in each class, and m is the total number of drivers.
Further, the method also comprises the following steps:
s9, in the behavior portrait of the driver, the type of the driver and the type of the vehicle have cross influence, the type of the vehicle and the type of the driver are mutually independent and are not related under different characteristics, and the chi-square value is checked to find out the characteristics which are obviously influenced by the cross of the driver and the vehicles of different types.
Still further, the method comprises the following steps:
s10, taking oil consumption as a dependent variable and the driving behavior characteristics of the driver as independent variables, constructing a multiple linear regression model, and analyzing which driving behavior characteristics of the driver can obviously influence the change of the oil consumption by checking the fitting regression effect.
The invention has the beneficial effects that:
the user behavior characteristics are obtained through the similarity of each driving behavior of the driver and a network structure analysis technology, the driver group is divided according to the user behavior characteristics, the driver groups with different behavior characteristic categories are obtained, the driver portrait is established, the dangerous driving behavior or the bad driving habit of the driver is found in time, and therefore the driver is trained in a targeted mode.
Drawings
FIG. 1 is a schematic flow chart of a method in an embodiment of the present invention.
Detailed Description
The present invention will be further described with reference to the accompanying drawings, and it should be noted that the present embodiment is based on the technical solution, and the detailed implementation and the specific operation process are provided, but the protection scope of the present invention is not limited to the present embodiment.
As shown in fig. 1, the embodiment provides a user portrait construction and clustering method based on multidimensional attributes, which includes the following steps:
s1, acquiring corresponding fields of description information in complete driving data of each trip in the bus, and sorting the data of each field into csv documents;
specifically, the driving data of the bus CAN be recorded by adopting a CAN bus vehicle-mounted instrument.
The actual driving data includes data of driving in many sections, and in the method of the present embodiment, a complete driving data record (i.e., driving data from a starting point to an end point) of each trip is required, so in the present embodiment, a trip threshold is set according to a distance of the whole bus to screen the driving data of each trip, and finally, complete driving data of each trip is obtained.
S2, primarily extracting driving behavior characteristics related to driving behaviors from the csv document obtained in the step S1;
in the embodiment, the driving behavior characteristics comprise a vehicle speed median, a vehicle speed standard deviation, a vehicle speed average value, an accelerator pedal percentage standard deviation, an accelerator pedal percentage average value, a sliding probability in a neutral gear state, an acceleration standard deviation, an electronic brake use probability, a foot brake use probability, a hand brake pull probability in a driving process, and an acceleration absolute value greater than 2m/s2Probability, average acceleration value, mode of vehicle speed, median of accelerator pedal percentage, mode of accelerator pedal percentage, electronic brakeThe absolute value of the vehicle recording times, the foot brake recording times and the acceleration is more than 2m/s2The recording times and the idling times of (1) are in proportion;
and S3, according to the driving behavior characteristics extracted in the step S2, deleting the characteristics with low distinguishing degree according to the characteristic probability distribution condition, and screening to obtain more representative characteristics.
Specifically, the more representative features can be obtained by drawing a feature probability distribution map and analyzing the feature probability distribution condition, deleting the features with the discrimination degree smaller than a set threshold value according to the specific distribution condition of the feature probability, and screening;
s4, sorting the features screened in the step S3 by using a maximum correlation minimum redundancy feature selection method (UmRMR model) based on mutual information, setting a screening threshold value and deleting the features of which the average mutual information value is smaller than the screening threshold value; determining the weight of the remaining features after screening according to the size of the average mutual information value, namely setting the average mutual information value of each feature as the weight value of the feature;
s5, weighting and normalizing the remaining features screened in the step S4, namely weighting the features by the weight of each feature in the step S4, wherein the calculation formula of the normalization is
Wherein xiWeight value of feature i, sum (x)i) The sum of the weight values of all features, x*Is a characteristic xiA normalized value;
s6, measuring the similarity of the driving behavior characteristics weighted and normalized in the step S5 among the driving data of each trip; setting a similarity threshold, and establishing a connecting edge when the similarity of the driving behavior characteristics among the driving data of each trip is greater than the similarity threshold;
specifically, cosine similarity is calculated for the weighted and normalized driving behavior characteristics among the driving data of each trip, and the cosine similarity is compared with a similarity threshold.
S7, dividing the driving behaviors of each trip into different classes (network communities) according to the similarity of the driving behavior characteristics among the driving data of each trip by adopting a Fast Unfolding algorithm;
the clustering principle of the Fast Unfolding algorithm is that each node in the network is regarded as an independent community, adjacent nodes are merged slowly, if the modularity of the whole network is improved after merging, merging is carried out, otherwise, the merging is cancelled; the above steps are circulated until the modularity of the network cannot be improved; then, each community is used as a node, and the merging algorithm is carried out on each community until the modularity of the whole network cannot be improved. In this embodiment, the driving data of each trip is specifically regarded as a node, the similarity between different drivers is regarded as a continuous edge, and the Fast Unfolding algorithm is used for clustering.
In this embodiment, the average classification accuracy of Fast Unfolding algorithm clustering in step S7 is also calculated according to the following formula:
wherein p iscTo average classification accuracy, niIs the total number of laps traveled by driver i,
is the CiThe number of driving trips of the driver i in the class,
the maximum number of runs of driver i in each class, and m is the total number of drivers.
In each category, the driving behavior characteristics of the driving data of each trip are taken as classification objects, namely, for a driver, if the driving behavior characteristics of the driver have stability, each trip of the driver is classified into the same category. However, in the case where there is a change in the driver's behavior, it is caused to be classified into another class. Therefore, in the method of the embodiment, an average classification accuracy index of behaviors is defined, namely if the behaviors of the driver can be effectively classified into the same class in each driving record, the classification accuracy is the highest. And averaging the classification accuracy of all drivers to obtain the average classification accuracy.
And S8, comprehensively analyzing the driving habits and the preference of the driver from three dimensions of driving unevenness, braking preference and vehicle speed preference in the form of a radar map.
Specifically, the driving instability dimension comprises three characteristics of a vehicle speed standard deviation, an acceleration standard deviation and an accelerator pedal percentage standard deviation; the brake preference includes two characteristics of electronic brake use probability and service brake use probability; the vehicle speed preference comprises four characteristics of a vehicle speed average value, a vehicle speed median, an acceleration absolute value average value and an accelerator pedal percentage average value. When the driver is comprehensively analyzed in three dimensions, firstly, the values of all the characteristics are normalized, then the values of the characteristics contained in each dimension are summed and normalized, finally, the scores of the driver in the three dimensions are obtained, and the scores are visually displayed in a radar map form.
S9, in the behavior portrait of the driver, the type of the driver and the type of the vehicle have cross influence, the type of the vehicle and the type of the driver are mutually independent and are not related under different characteristics, and the chi-square value is checked to find out the characteristics which are obviously influenced by the cross of the driver and the vehicles of different types.
S10, taking oil consumption as a dependent variable and the driving behavior characteristics of the driver as independent variables, constructing a multiple linear regression model, and analyzing which driving behavior characteristics of the driver can obviously influence the change of the oil consumption by checking the fitting regression effect.
In the embodiment, the vehicle types are divided into oil consumption vehicles, electricity consumption vehicles and oil-electricity hybrid vehicles for distribution analysis, and specific characteristics of different vehicle types are found out to obviously influence oil consumption.
In this embodiment, in step S1, after the obtained csv document is sorted, data cleaning is further performed on the csv document, where the data cleaning includes data for filling missing data and detecting a logical error; specifically, missing data can be filled in by using hot deck padding.
Since the raw data has a high collection density and the traveling data does not change significantly in a very short time, the missing values are filled with data having similar missing values in this embodiment. And (3) deleting the data of the track sections, such as the motor rotating speed reaching 16000r/min, for the detected abnormal data, such as the speed of the track sections reaching 120km/h, and replacing and filling the abnormal data with null values nearby.
In this embodiment, in step S2, the vehicle speed mode, the median accelerator pedal percentage, the mode accelerator pedal percentage, the electronic brake record times, the foot brake record times, and the absolute value of acceleration greater than 2m/S may be deleted according to the service requirement and the actual data condition2The number of times of recording and the number of times of idling of (c) account for these 7 features.
Various corresponding changes and modifications can be made by those skilled in the art based on the above technical solutions and concepts, and all such changes and modifications should be included in the protection scope of the present invention.
Claims (7)
1. A user portrait construction and clustering method based on multi-dimensional attributes is characterized by comprising the following steps:
s1, acquiring corresponding fields of description information in complete driving data of each trip in the bus, and sorting the data of each field into csv documents;
s2, primarily extracting driving behavior characteristics related to driving behaviors from the csv document obtained in the step S1;
s3, according to the driving behavior characteristics extracted in the step S2, the characteristics with low distinguishing degree are deleted according to the characteristic probability distribution condition, and more representative characteristics are obtained through screening;
s4, sorting the features screened in the step S3 by using a maximum correlation minimum redundancy feature selection method based on mutual information, setting a screening threshold value and deleting the features of which the average mutual information value is smaller than the screening threshold value; determining the weight of the remaining features after screening according to the size of the average mutual information value, namely setting the average mutual information value of each feature as the weight value of the feature;
s5, weighting and normalizing the remaining characteristics after being screened in the step S4 by using the weight value of each characteristic obtained in the step S4;
s6, measuring the similarity of the driving behavior characteristics weighted and normalized in the step S5 among the driving data of each trip; setting a similarity threshold, and establishing a connecting edge when the similarity of the driving behavior characteristics among the driving data of each trip is greater than the similarity threshold;
s7, dividing the driving behaviors of each trip into different classes according to the similarity of the driving behavior characteristics among the driving data of each trip by adopting a Fast Unfolding algorithm;
s8, comprehensively analyzing the driving habits and preferences of the driver from three dimensions of driving unevenness, braking preference and vehicle speed preference in the form of a radar map:
the driving instability dimension comprises three characteristics of a vehicle speed standard deviation, an acceleration standard deviation and an accelerator pedal percentage standard deviation; the brake preference includes two characteristics of electronic brake use probability and service brake use probability; the vehicle speed preference comprises four characteristics of a vehicle speed average value, a vehicle speed median, an acceleration absolute value average value and an accelerator pedal percentage average value; when the driver is comprehensively analyzed in three dimensions, firstly, the values of all the characteristics are normalized, then the values of the characteristics contained in each dimension are summed and normalized, finally, the scores of the driver in the three dimensions are obtained, and the scores are visually displayed in a radar map form.
2. The method for constructing and clustering user profiles based on multidimensional attributes as recited in claim 1, wherein in step S1, a CAN bus onboard instrument is used to record driving data of the bus.
3. The method for constructing and clustering user portraits based on multidimensional attributes as claimed in claim 1, wherein in step S2, the driving behavior characteristics comprise vehicle speed median, vehicle speed standard deviation, vehicle speed average, accelerator pedal percentage standard deviation, accelerator pedal percentage average, and neutral gear stateThe sliding probability, the acceleration standard deviation, the electronic brake use probability, the service brake use probability, the handle brake use probability in the driving process and the acceleration absolute value are more than 2m/s2Probability, average acceleration value, mode of vehicle speed, median of accelerator pedal percentage, mode of accelerator pedal percentage, electronic brake recording times, foot brake recording times, and absolute acceleration value greater than 2m/s2The number of times of recording and the number of times of idling are in proportion.
4. The method for constructing and clustering user portraits according to claim 1, wherein in step S6, cosine similarity is calculated for the weighted and normalized driving behavior features of the driving data of each pass, and the cosine similarity is compared with a similarity threshold.
5. The method for constructing and clustering user portraits based on multidimensional attribute as claimed in claim 1, wherein in step S7, the method further comprises calculating the average classification accuracy of Fast Unfolding algorithm clustering according to the following formula:
wherein p iscTo average classification accuracy, niIs the total number of laps traveled by driver i,
is the CiThe number of driving trips of the driver i in the class,
the maximum number of runs of driver i in each class, and m is the total number of drivers.
6. The method for building and clustering user portraits based on multidimensional attribute as claimed in claim 1, further comprising the steps of:
s9, in the behavior portrait of the driver, the type of the driver and the type of the vehicle have cross influence, the type of the vehicle and the type of the driver are mutually independent and are not related under different characteristics, and the chi-square value is checked to find out the characteristics which are obviously influenced by the cross of the driver and the vehicles of different types.
7. The method of claim 6, further comprising the steps of:
s10, taking oil consumption as a dependent variable and the driving behavior characteristics of the driver as independent variables, constructing a multiple linear regression model, and analyzing which driving behavior characteristics of the driver can obviously influence the change of the oil consumption by checking the fitting regression effect.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811414495.7A CN109635852B (en) | 2018-11-26 | 2018-11-26 | User portrait construction and clustering method based on multi-dimensional attributes |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811414495.7A CN109635852B (en) | 2018-11-26 | 2018-11-26 | User portrait construction and clustering method based on multi-dimensional attributes |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109635852A CN109635852A (en) | 2019-04-16 |
| CN109635852B true CN109635852B (en) | 2021-03-23 |
Family
ID=66069138
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811414495.7A Active CN109635852B (en) | 2018-11-26 | 2018-11-26 | User portrait construction and clustering method based on multi-dimensional attributes |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109635852B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110097287B (en) * | 2019-05-07 | 2021-06-22 | 宏图物流股份有限公司 | Group portrait method for logistic drivers |
| CN110837587B (en) * | 2019-09-30 | 2023-05-23 | 北京水滴科技集团有限公司 | Data matching method and system based on machine learning |
| CN110807052B (en) * | 2019-11-05 | 2022-08-02 | 佳都科技集团股份有限公司 | User group classification method, device, equipment and storage medium |
| CN111582563B (en) * | 2020-04-24 | 2023-04-25 | 中山大学 | Short-term prediction method, system, device and storage medium for individual travel time |
| CN112396118A (en) * | 2020-11-25 | 2021-02-23 | 浙江天行健智能科技有限公司 | GM-HMM-based driver acceleration intention modeling method |
| CN113033651A (en) * | 2021-03-22 | 2021-06-25 | 上海仙塔智能科技有限公司 | Method and device for improving driving safety |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107480936A (en) * | 2017-07-11 | 2017-12-15 | 杭州电子科技大学 | A kind of talent ability depicting method based on scientific achievement data |
| CN107729951A (en) * | 2017-11-14 | 2018-02-23 | 吉林大学 | A kind of driving behavior analytical equipment and method for considering road and environmental characteristic |
| CN108229304A (en) * | 2017-11-17 | 2018-06-29 | 清华大学 | A kind of driving behavior recognition methods based on Clustering of systematization |
| CN108875795A (en) * | 2018-05-28 | 2018-11-23 | 哈尔滨工程大学 | A kind of feature selecting algorithm based on Relief and mutual information |
-
2018
- 2018-11-26 CN CN201811414495.7A patent/CN109635852B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107480936A (en) * | 2017-07-11 | 2017-12-15 | 杭州电子科技大学 | A kind of talent ability depicting method based on scientific achievement data |
| CN107729951A (en) * | 2017-11-14 | 2018-02-23 | 吉林大学 | A kind of driving behavior analytical equipment and method for considering road and environmental characteristic |
| CN108229304A (en) * | 2017-11-17 | 2018-06-29 | 清华大学 | A kind of driving behavior recognition methods based on Clustering of systematization |
| CN108875795A (en) * | 2018-05-28 | 2018-11-23 | 哈尔滨工程大学 | A kind of feature selecting algorithm based on Relief and mutual information |
Non-Patent Citations (3)
| Title |
|---|
| A Two-Step Segmentation Algorithm for Behavioral Clustering of Naturalistic Driving Styles;Bryan Higgs 等;《Proceedings of the 16th International IEEE Annual Conference on Intelligent Transportation Systems (ITSC 2013)》;20140130;全文 * |
| 基于车联网数据挖掘的营运车辆驾驶速度行为聚类研究;孙川 等;《交通运输系统工程与信息》;20151231;全文 * |
| 集团专车驾驶行为分析研究;徐国功;《中国优秀硕士学位论文全文数据库 信息科技辑》;20170815;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109635852A (en) | 2019-04-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109635852B (en) | User portrait construction and clustering method based on multi-dimensional attributes | |
| CN109033332B (en) | Driving behavior analysis method, medium and system | |
| CN109117869B (en) | User portrayal method, medium and system | |
| Chen et al. | A graphical modeling method for individual driving behavior and its application in driving safety analysis using GPS data | |
| DE112009002674B4 (en) | navigation device | |
| CN110197588A (en) | A kind of truck driving behavior appraisal procedure and device based on GPS track data | |
| CN111461185A (en) | Driving behavior analysis method based on improved K-means | |
| CN112201038A (en) | Road network risk assessment method based on risk of bad driving behavior of single vehicle | |
| CN111688713B (en) | Driving behavior analysis method and device | |
| CN110929939B (en) | Landslide hazard susceptibility spatial prediction method based on clustering-information coupling model | |
| CN109408557B (en) | Traffic accident cause analysis method based on multiple correspondences and K-means clustering | |
| CN110304068B (en) | Method, device, equipment and storage medium for collecting automobile driving environment information | |
| CN110084534B (en) | Driving risk factor quantification method based on driving behavior portrait | |
| CN108769104B (en) | Road condition analysis and early warning method based on vehicle-mounted diagnosis system data | |
| CN109886724B (en) | Robust resident travel track identification method | |
| CN110562261B (en) | Method for detecting risk level of driver based on Markov model | |
| CN114299742B (en) | Speed limit information dynamic identification and update recommendation method for expressway | |
| CN109598931A (en) | Group based on traffic safety risk divides and difference analysis method and system | |
| CN113581188A (en) | Commercial vehicle driver driving style identification method based on Internet of vehicles data | |
| CN113095387B (en) | Road risk identification method based on networking vehicle-mounted ADAS | |
| CN114048798A (en) | Automobile driving condition construction method based on improved noise reduction self-encoder | |
| CN109849926B (en) | Method and system for distinguishing whether taxi is handed to others for driving | |
| Van Hinsbergh et al. | Vehicle point of interest detection using in-car data | |
| CN116092037A (en) | Vehicle type identification method integrating track space-semantic features | |
| CN114169426A (en) | Beidou position data-based highway traffic potential safety hazard investigation method |
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 |