CN109635852A - A kind of building of user's portrait and clustering method based on multidimensional property - Google Patents
A kind of building of user's portrait and clustering method based on multidimensional property Download PDFInfo
- Publication number
- CN109635852A CN109635852A CN201811414495.7A CN201811414495A CN109635852A CN 109635852 A CN109635852 A CN 109635852A CN 201811414495 A CN201811414495 A CN 201811414495A CN 109635852 A CN109635852 A CN 109635852A
- Authority
- CN
- China
- Prior art keywords
- driver
- driving behavior
- value
- feature
- portrait
- 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.)
- Granted
Links
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 user that the invention discloses a kind of based on multidimensional property draws a portrait building and clustering method, comprising: is pre-processed to initial data and extracts driver's driving behavior;Feature is screened with maximal correlation minimal redundancy feature selection approach based on mutual information, the driver group with similar driving behavior is gathered for one kind by the similitude and Crosslinking Structural technology of driver's driving behavior, user's portrait is carried out from drive preference of the different dimensions to driver using radar map, finally the factor for influencing driving behavior is analyzed.The present invention can be effectively reflected the Behavior preference that driver drives, can allow driver understand oneself need it is improved drive to be accustomed to, can also public transport company be allowed targetedly to give training to driver.
Description
Technical field
The present invention relates to driving behavior detection technique fields, and in particular to a kind of user's portrait structure based on multidimensional property
Build the method with cluster.
Background technique
In recent years, with the rapid growth of national economy, city bus has played great function in slow stifled protect in smooth, more next
More paid attention to and public welcome by government.In urban traffic environment, the driving behavior of bus driver --- it is anxious to add
Speed brings to a halt and steps on the gas, and has seriously affected the safety of driving and the comfort level of passenger.Therefore specification bus driver is driven
Behavior is sailed, guarantees the service quality and safety of bus, becomes the most important thing.A kind of the effective of row is also lacked at present
Driver's bus driver behavioral value method.
Summary of the invention
In view of the deficiencies of the prior art, building and cluster the present invention is intended to provide a kind of user based on multidimensional property draws a portrait
Method, user behavior characteristics are obtained by the similitude and Crosslinking Structural technology of every time driving behavior of driver, and according to
This divides driver group, obtains the driver group of different behavioural characteristic classifications, and establishes driver's portrait, sends out in time
The dangerous driving behavior or bad steering habit of existing driver, to targetedly be giveed training to driver.
To achieve the goals above, the present invention adopts the following technical scheme:
A kind of building of user's portrait and clustering method based on multidimensional property, includes the following steps:
S1, the respective field for obtaining description information in each time complete travelling data in bus, and by each field
Data preparation is at csv document;
S2, driving behavior relevant to driving behavior is tentatively extracted from csv document obtained in step S1;
S3, according to the driving behavior extracted in step S2, discrimination is deleted according to characteristic probability distribution situation
Low feature, screening obtain more representative feature;
S4, the spy that will be screened in step S3 using maximal correlation minimal redundancy feature selection approach based on mutual information
Sign is ranked up, and simultaneously censored mean association relationship is less than the feature for screening threshold value to setting screening threshold value;It will be remained after screening
Under feature determine that weight, i.e., the Average Mutual value of each feature are set as the power of this feature according to the size of Average Mutual value
Weight values;
S5, using obtained in step S4 the weighted value of each feature to after being screened in step S4 remaining feature into
Row is weighted and is normalized;
S6, measurement each time travelling data between the driving behavior after step S5 is weighted and normalized it is similar
Property;Similarity threshold is set, when the similitude of the driving behavior between each time travelling data is greater than the similarity threshold,
The company of foundation side;
S7, using Fast Unfolding algorithm, according to the similar of the driving behavior between every time travelling data
Every time driving behavior is divided into different classes by property;
S8, in the form of radar map from drive uneven stability, brake three dimensions of Preference and speed Preference to driving
Drive habit and the preference progress comprehensive analysis of member:
It drives uneven stability dimension and includes speed standard deviation, acceleration standard deviation and gas pedal Percentage Criterion poor three
Feature;The Preference that brakes includes that electric brake uses probability, service brake to use two features of probability;Speed Preference is flat comprising speed
Four mean value, speed median, acceleration absolute value average value and gas pedal percentage average value features;Under three dimensions
When carrying out comprehensive analysis to driver, the value of all features is normalized first, then by the included feature of each dimension
Value sum and normalize, finally obtain score of the driver under three dimensions, visualization exhibition carried out in the form of radar map
Show.
Further, in step S1, using the travelling data of CAN bus vehicle-mounted instrument record bus.
Further, in step S2, driving behavior includes speed median, speed standard deviation, speed average value, oil
Pedal Percentage Criterion is poor for door, slides probability, acceleration standard deviation, electricity under gas pedal percentage average value, neutral position state
Son brake using probability, service brake using pull the hand brake in probability, driving conditions probability, acceleration absolute value be greater than 2m/s2It is general
Rate, acceleration average value, speed mode, gas pedal percentage median, gas pedal percentage mode, electric brake record
Number, service brake record number, acceleration absolute value are greater than 2m/s2Record number, idling number accounting.
Further, special to the driving behavior after being weighted and normalizing between each time travelling data in step S6
Sign calculates cosine similarity, and cosine similarity is compared with similarity threshold.
It further, further include the average classification for calculating Fast Unfolding algorithm cluster according to the following formula in step S7
Accuracy rate:
Wherein pcFor average classification accuracy, niFor driver i traveling total number,For CiThe traveling of driver i time in class
Number,For the maximum value of traveling number of the driver i in every one kind, m is driver's sum.
Further, further include following steps:
During the behavior of S9, driver are drawn a portrait, there are cross influences between driver and type of vehicle, it is assumed that the type of vehicle
It is mutually indepedent under different characteristic with classification belonging to driver, it is unrelated, it tests, is found out by driver to chi-square value
Feature significant with different types of vehicle cross influence.
Further, further include following steps:
S10, using oil consumption as dependent variable, the driving behavior of driver is independent variable, construct multiple linear regression model,
By checking fitting regression effect, which driving behavior for analyzing driver can significantly affect oil changes.
The beneficial effects of the present invention are:
User behavior characteristics are obtained by the similitude and Crosslinking Structural technology of every time driving behavior of driver, and according to
This divides driver group, obtains the driver group of different behavioural characteristic classifications, and establishes driver's portrait, sends out in time
The dangerous driving behavior or bad steering habit of existing driver, to targetedly be giveed training to driver.
Detailed description of the invention
Fig. 1 is the method flow schematic diagram in the embodiment of the present invention.
Specific embodiment
Below with reference to attached drawing, the invention will be further described, it should be noted that the present embodiment is with this technology side
Premised on case, the detailed implementation method and specific operation process are given, but protection scope of the present invention is not limited to this reality
Apply example.
Building and clustering method as shown in Figure 1, the user that the present embodiment provides a kind of based on multidimensional property draws a portrait, including such as
Lower step:
S1, the respective field for obtaining description information in each time complete travelling data in bus, and by each field
Data preparation is at csv document;
It specifically, can be using the travelling data of CAN bus vehicle-mounted instrument record bus.
It in actual travelling data can include the data of many section travels, and needed in the present embodiment method
It is each time complete travelling data record (i.e. from the travelling data of starting point to the end), so in the present embodiment, according to public transport
The distance of vehicle whole process sets stroke threshold value to screen each time travelling data, finally obtains each time complete travelling data.
S2, driving behavior relevant to driving behavior is tentatively extracted from csv document obtained in step S1;
In the present embodiment, driving behavior includes that speed median, speed standard deviation, speed average value, throttle are stepped on
Plate Percentage Criterion is poor, slide probability, acceleration standard deviation, electronics under gas pedal percentage average value, neutral position state are stopped
Vehicle using probability, service brake using pull the hand brake in probability, driving conditions probability, acceleration absolute value be greater than 2m/s2Probability, plus
Speed average, speed mode, gas pedal percentage median, gas pedal percentage mode, electric brake record number,
Service brake records number, acceleration absolute value is greater than 2m/s2Record number, idling number accounting;
S3, according to the driving behavior extracted in step S2, discrimination is deleted according to characteristic probability distribution situation
Low feature, screening obtain more representative feature.
It specifically, can be by drawing characteristic probability distribution map and analyzing characteristic probability distribution situation, according to characteristic probability
Specific distribution situation delete discrimination be less than given threshold feature, screening obtain more representative feature;
S4, will be in step S3 using maximal correlation minimal redundancy feature selection approach (UmRMR model) based on mutual information
It screens obtained feature to be ranked up, threshold value simultaneously feature of the censored mean association relationship less than the screening threshold value is screened in setting;
Feature remaining after screening is determined that weight, i.e., the Average Mutual value of each feature are set according to the size of Average Mutual value
For the weighted value of this feature;
S5, remaining feature after screening in process step S4 is weighted and is normalized, i.e., with each in step S4
The weighted value of feature carries out tax power to feature, and normalized calculation formula isWherein xiIt is characterized the weight of i
Value, sum (xi) be all features weighted value and, x*It is characterized xiValue after normalization;
S6, measurement each time travelling data between the driving behavior after step S5 is weighted and normalized it is similar
Property;Similarity threshold is set, when the similitude of the driving behavior between each time travelling data is greater than the similarity threshold,
The company of foundation side;
Specifically, cosine phase is calculated to the driving behavior after being weighted and normalizing between each time travelling data
It is compared like degree, and by cosine similarity with similarity threshold.
S7, using Fast Unfolding algorithm, according to the similitude of driving behavior between every time travelling data
Every time driving behavior is divided into different classes (network community);
The principle of Fast Unfolding algorithm cluster is, regards each node in network as an independence first
Corporations, slowly neighbouring node is merged, if the modularity of whole network improves after merging, is just merged, otherwise
Revocation;So circulation, until the modularity of network can not improve;Then each corporations are treated as a node again, to every
A corporations carry out so merging algorithm, until the modularity of whole network can not improve.In the present embodiment, specifically will
Every time travelling data regards a node as, regards the similitude between different drivers as even side, utilizes Fast
Unfolding algorithm is clustered.
In the present embodiment, the average classification for calculating Fast Unfolding algorithm cluster in step S7 according further to following formula is quasi-
True rate:
Wherein pcFor average classification accuracy, niFor driver i traveling total number,For CiThe traveling of driver i time in class
Number,For the maximum value of traveling number of the driver i in every one kind, m is driver's sum.
In each classification, using the driving behavior of every time travelling data as object of classification, i.e., to a driver
For, if his driving behavior has stability, his every time traveling record can all be assigned in same category.But
It is that in the case where changing there are driving behavior, will lead to assign in other class.Therefore in the present embodiment method
In, it defines a kind of behavior and is averaged classification accuracy index, i.e., if every time driving recording driving behavior can be assigned to effectively
In same class, show classification accuracy highest.It is averaged to the classification accuracy of all drivers, it is accurate to obtain average classification
Property.
S8, in the form of radar map from drive uneven stability, brake three dimensions of Preference and speed Preference to driving
Drive habit and the preference progress comprehensive analysis of member.
Specifically, driving uneven stability dimension includes speed standard deviation, acceleration standard deviation and gas pedal percentage mark
Quasi- poor three features;The Preference that brakes includes that electric brake uses probability, service brake to use two features of probability;Speed Preference packet
Four average value containing speed, speed median, acceleration absolute value average value and gas pedal percentage average value features.Three
When carrying out comprehensive analysis to driver under a dimension, the value of all features is normalized first, then by each dimension institute
Value comprising feature is summed and is normalized, and score of the driver under three dimensions is finally obtained, and is carried out in the form of radar map
It visualizes.
During the behavior of S9, driver are drawn a portrait, there are cross influences between driver and type of vehicle, it is assumed that the type of vehicle
It is mutually indepedent under different characteristic with classification belonging to driver, it is unrelated, it tests, is found out by driver to chi-square value
Feature significant with different types of vehicle cross influence.
S10, using oil consumption as dependent variable, the driving behavior of driver is independent variable, construct multiple linear regression model,
By checking fitting regression effect, which driving behavior for analyzing driver can significantly affect oil changes.
In the present embodiment, it is analyzed, is looked for by the way that vehicle is divided into oil consumption vehicle, power consumption vehicle and the mixed motor-car of oil electricity
Which specific characteristic remarkable influences oil consumption in different automobile types out.
In the present embodiment, in step S1, after arranging obtained csv document, it is also necessary to carry out data to csv document
Cleaning, the data cleansing include and detecting the data of logic error to filling up missing data;Specifically, can pass through
Missing data is filled using hot deck enthesis.
Since the acquisition density of initial data is very high, and travelling data change in a very short period of time it is unobvious, so
Missing values are filled using data similar in missing values in the present embodiment.For the abnormal data detected, such as " wheelpath section
Speed reach 120km/h ", delete the data of these orbit segments, such as " motor speed reaches 16000r/min ", it is this exception number
According to null value replacement and nearest polishing.
In the present embodiment, in step S2, it can be needed in conjunction with business and real data situation deletes speed mode, oil
Door pedal percentage median, gas pedal percentage mode, electric brake record number, service brake record number, acceleration are exhausted
2m/s is greater than to value2Record number, idling number accounting this 7 features.
For those skilled in the art, it can be provided various corresponding according to above technical solution and design
Change and modification, and all these change and modification, should be construed as being included within the scope of protection of the claims of the present invention.
Claims (7)
- Building and clustering method 1. a kind of user based on multidimensional property draws a portrait, which comprises the steps of:S1, the respective field for obtaining description information in each time complete travelling data in bus, and by the data of each field It is organized into csv document;S2, driving behavior relevant to driving behavior is tentatively extracted from csv document obtained in step S1;S3, according to the driving behavior extracted in step S2, it is low that discrimination is deleted according to characteristic probability distribution situation Feature, screening obtain more representative feature;S4, using maximal correlation minimal redundancy feature selection approach based on mutual information by the feature screened in step S3 into Row sequence, setting screening threshold value and censored mean association relationship are less than the feature of the screening threshold value;It is remaining after screening Feature determines that weight, i.e., the Average Mutual value of each feature are set as the weight of this feature according to the size of Average Mutual value Value;S5, remaining feature after screening in process step S4 is added using the weighted value for obtaining each feature in step S4 It weighs and normalizes;The similitude of the driving behavior after step S5 is weighted and normalized between each time S6, measurement travelling data; Similarity threshold is set, when the similitude of the driving behavior between each time travelling data is greater than the similarity threshold, is built Li Lianbian;S7, will be every according to the similitude of driving behavior between every time travelling data using Fast Unfolding algorithm The driving behavior plowed is divided into different classes;S8, in the form of radar map from drive uneven stability, brake three dimensions of Preference and speed Preference to driver's Habit of driving and preference carry out comprehensive analysis:Driving uneven stability dimension includes speed standard deviation, acceleration standard deviation and poor three spies of gas pedal Percentage Criterion Sign;The Preference that brakes includes that electric brake uses probability, service brake to use two features of probability;Speed Preference is average comprising speed Value, speed median, acceleration absolute value average value and four features of gas pedal percentage average value;It is right under three dimensions When driver carries out comprehensive analysis, the value of all features is normalized first, then by the included feature of each dimension Value is summed and is normalized, and finally obtains score of the driver under three dimensions, is visualized in the form of radar map.
- Building and clustering method 2. the user according to claim 1 based on multidimensional property draws a portrait, which is characterized in that step In S1, using the travelling data of CAN bus vehicle-mounted instrument record bus.
- Building and clustering method 3. the user according to claim 1 based on multidimensional property draws a portrait, which is characterized in that step In S2, driving behavior includes that speed median, speed standard deviation, speed average value, gas pedal Percentage Criterion be poor, oil Slide probability, acceleration standard deviation, electric brake under door pedal percentage average value, neutral position state are made using probability, service brake With pull the hand brake in probability, driving conditions probability, acceleration absolute value be greater than 2m/s2Probability, acceleration average value, speed it is many Number, service brake record number, accelerates gas pedal percentage median, gas pedal percentage mode, electric brake record number It spends absolute value and is greater than 2m/s2Record number, idling number accounting.
- Building and clustering method 4. the user according to claim 1 based on multidimensional property draws a portrait, which is characterized in that step In S6, cosine similarity is calculated to the driving behavior after being weighted and normalizing between each time travelling data, and will Cosine similarity is compared with similarity threshold.
- Building and clustering method 5. the user according to claim 1 based on multidimensional property draws a portrait, which is characterized in that step Further include the average classification accuracy for calculating Fast Unfolding algorithm cluster according to the following formula in S7:Wherein pcFor average classification accuracy, niFor driver i traveling total number,For CiThe traveling number of driver i in class,For the maximum value of traveling number of the driver i in every one kind, m is driver's sum.
- Building and clustering method 6. the user according to claim 1 based on multidimensional property draws a portrait, which is characterized in that also wrap Include following steps:S9, driver behavior portrait in, there are cross influences between driver and type of vehicle, it is assumed that the type of vehicle with drive Classification belonging to the person of sailing is mutually indepedent under different characteristic, unrelated, tests to chi-square value, finds out by driver and not The significant feature of vehicle cross influence of same type.
- Building and clustering method 7. the user according to claim 6 based on multidimensional property draws a portrait, which is characterized in that also wrap Include following steps:S10, using oil consumption as dependent variable, the driving behavior of driver is independent variable, constructs multiple linear regression model, passes through Check fitting regression effect, which driving behavior for analyzing driver can significantly affect oil changes.
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 true CN109635852A (en) | 2019-04-16 |
CN109635852B 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) |
Cited By (6)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110097287A (en) * | 2019-05-07 | 2019-08-06 | 宏图物流股份有限公司 | A kind of group's portrait method of logistics driver |
CN110807052A (en) * | 2019-11-05 | 2020-02-18 | 佳都新太科技股份有限公司 | User group classification method, device, equipment and storage medium |
CN110837587A (en) * | 2019-09-30 | 2020-02-25 | 北京健康之家科技有限公司 | Data matching method and system based on machine learning |
CN111582563A (en) * | 2020-04-24 | 2020-08-25 | 中山大学 | Individual journey time short-term prediction method, system, device and storage medium |
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 |
---|
BRYAN HIGGS 等: "A Two-Step Segmentation Algorithm for Behavioral Clustering of Naturalistic Driving Styles", 《PROCEEDINGS OF THE 16TH INTERNATIONAL IEEE ANNUAL CONFERENCE ON INTELLIGENT TRANSPORTATION SYSTEMS (ITSC 2013)》 * |
孙川 等: "基于车联网数据挖掘的营运车辆驾驶速度行为聚类研究", 《交通运输***工程与信息》 * |
徐国功: "集团专车驾驶行为分析研究", 《中国优秀硕士学位论文全文数据库 信息科技辑》 * |
Cited By (9)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110097287A (en) * | 2019-05-07 | 2019-08-06 | 宏图物流股份有限公司 | A kind of group's portrait method of logistics driver |
CN110097287B (en) * | 2019-05-07 | 2021-06-22 | 宏图物流股份有限公司 | Group portrait method for logistic drivers |
CN110837587A (en) * | 2019-09-30 | 2020-02-25 | 北京健康之家科技有限公司 | Data matching method and system based on machine learning |
CN110807052A (en) * | 2019-11-05 | 2020-02-18 | 佳都新太科技股份有限公司 | User group classification method, device, equipment and storage medium |
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 |
CN111582563A (en) * | 2020-04-24 | 2020-08-25 | 中山大学 | Individual journey time short-term prediction method, system, device 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 |
Also Published As
Publication number | Publication date |
---|---|
CN109635852B (en) | 2021-03-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109635852A (en) | A kind of building of user's portrait and clustering method based on multidimensional property | |
Toledo et al. | Lane-changing model with explicit target lane choice | |
CN107492251A (en) | It is a kind of to be identified and driving condition supervision method based on the driver identity of machine learning and deep learning | |
CN101751438B (en) | Theme webpage filter system for driving self-adaption semantics | |
CN108492557A (en) | Highway jam level judgment method based on multi-model fusion | |
CN105279966B (en) | A kind of congestion recognition methods of Campus transport | |
CN106023344A (en) | Driving style estimation method based on driving pattern transition probability | |
CN110458244A (en) | A kind of traffic accident Severity forecasting method applied to Regional Road Network | |
CN109033332A (en) | Driving behavior analysis method, medium and system | |
CN106651210A (en) | CAN data-based driver comprehensive quality evaluation method | |
CN105632187B (en) | Low-power consumption curb parking detection method based on geomagnetic sensor network | |
CN107590766A (en) | A kind of method of discrimination of the land used combination form related to road traffic accident risk | |
CN110175713B (en) | Method for predicting trip purpose of time-sharing rental automobile user | |
CN106960189A (en) | A kind of driving intention decision method based on hidden Markov model | |
CN106777169A (en) | A kind of user's trip hobby analysis method based on car networking data | |
CN110194041A (en) | The adaptive bodywork height adjusting method of Multi-source Information Fusion | |
Yahaya et al. | Ensemble-based model selection for imbalanced data to investigate the contributing factors to multiple fatality road crashes in Ghana | |
CN109584561A (en) | Urban highway traffic evaluation of running status method based on floating car data | |
CN116092037B (en) | Vehicle type identification method integrating track space-semantic features | |
Bhuyan et al. | Affinity propagation clustering in defining level of service criteria of urban streets | |
Jain et al. | Analysis of road accidents in India and prediction of accident severity | |
Zhang et al. | Heterogeneity analysis of operating mode distribution for modeling energy consumption of light-duty vehicles | |
Wu et al. | Modeling and quantitative evaluation method of environmental complexity for measuring autonomous capabilities of military unmanned ground vehicles | |
CN114169426A (en) | Beidou position data-based highway traffic potential safety hazard investigation method | |
Sun et al. | Freeway traffic safety state classification method based on multi-parameter fusion clustering |
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 |