CN109410586A - A kind of Traffic State Detection Method based on multivariate data fusion - Google Patents
A kind of Traffic State Detection Method based on multivariate data fusion Download PDFInfo
- Publication number
- CN109410586A CN109410586A CN201811524358.9A CN201811524358A CN109410586A CN 109410586 A CN109410586 A CN 109410586A CN 201811524358 A CN201811524358 A CN 201811524358A CN 109410586 A CN109410586 A CN 109410586A
- Authority
- CN
- China
- Prior art keywords
- gps
- track
- section
- data
- taxi
- 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
- G08—SIGNALLING
- G08G—TRAFFIC CONTROL SYSTEMS
- G08G1/00—Traffic control systems for road vehicles
- G08G1/01—Detecting movement of traffic to be counted or controlled
- G08G1/0104—Measuring and analyzing of parameters relative to traffic conditions
- G08G1/0125—Traffic data processing
- G08G1/0133—Traffic data processing for classifying traffic situation
Abstract
The invention discloses a kind of Traffic State Detection Methods based on multivariate data fusion, first, GPS track segmentation is carried out by user to GPS data from taxi and cellphone GPS data respectively using clustering algorithm, obtains each segmentation track taxi GPS and cellphone GPS segmentation track；Then, using taxi GPS segmentation track as training sample, building identification model, the class taxi GPS extracted in cellphone GPS segmentation track is segmented track；Recycle map-matching algorithm that the segmentation track taxi GPS and class taxi GPS segmentation track are matched on the section in urban road network；The matching result for recycling two kinds of track calculates separately the Vehicle Speed of respective stretch；Finally the section Vehicle Speed calculated separately by two kinds of path matching result is merged using evidence theory, estimates the Vehicle Speed in the section.The present invention can accurately detect urban road traffic state information.
Description
Technical field
The invention belongs to technical field of transportation, and in particular to a kind of traffic condition detection side based on multivariate data fusion
Method.
Background technique
With the continuous development of Chinese Urbanization, urban road is increasingly complicated, administrative department to the management of urban transportation and
Control difficulty also continue to increase, therefore urban highway traffic situation is measured in real time and extensively, accurately evaluation have
Highly important meaning.Traditional traffic condition detection is mainly based upon artificial traffic study, carries out manually to road section
Vehicle speed measuring etc., in recent years, with the rapid development of sensing technology and mechanics of communication, the traffic of various traffic sensing equipment records
Condition information is widely used in the research of traffic administration and control aspect, such as circuit detector data records what vehicle passed through
The position of quantity, the snapshot picture of video monitoring data record road different moments, Floating Car GPS data record different moments vehicle
Confidence breath etc..Most cities taxi is equipped with GPS positioning device at present, and GPS data from taxi is widely used in traffic
Research field.It largely include the maps APP users such as private car, bus with the rise of various map APP softwares and universal
Location information can effectively be obtained extensively, it is a kind of new to be provided using traffic big data the research of urban traffic conditions
Direction.But existing method still has following problem:
1) though GPS data from taxi can effectively record the location information of taxi, taxi tends to travel on city
Major trunk roads, the regions such as the lower road of city level and suburb go out to lease coverage rate it is still lower, to urban highway traffic
The complete detection of situation has some limitations.Its recording frequency is lower simultaneously, to the accuracy of detected traffic condition
There are certain limitations.
2) traffic user information of the cellphone GPS data record including various modes of transportation, including bus, hire out
Vehicle, private car, bicycle, pedestrian etc., and it is motor-driven primarily directed to bus, taxi etc. to the research of urban traffic conditions
Vehicle, the information such as bicycle, pedestrian that cellphone GPS data are recorded can cause serious interference and shadow to traffic condition testing result
It rings.
3) traffic behavior the case where there may be inconsistent or even contradictions of two kinds of GPS datas perception, a kind of method is needed
This inconsistency is measured, to merged on this basis to obtain accurate traffic behavior.
Summary of the invention
Technical problem solved by the invention is in view of the deficiencies of the prior art, to provide one kind and melted based on multivariate data
The Traffic State Detection Method of conjunction, fusion GPS data from taxi and cellphone GPS data, can accurately detect in urban road network
The Vehicle Speed in each section, i.e. traffic state information.
Technical solution provided by the present invention are as follows:
A kind of Traffic State Detection Method based on multivariate data fusion, comprising the following steps:
Step 1: constructing urban road network according to urban road information；
Step 2: obtaining the GPS data from taxi and cellphone GPS data under passenger carrying status；
Step 3: pressing user to GPS data from taxi and cellphone GPS data (continuous path point) respectively using clustering algorithm
GPS track segmentation is carried out, each segmentation track taxi GPS and cellphone GPS segmentation track are obtained；It can first will be under passenger carrying status
GPS data from taxi and cellphone GPS data recycle clustering algorithm to the GPS number after sequence by user's progress time-sequencing
According to progress GPS track segmentation；
Step 4: using taxi GPS segmentation track as training sample, building identification model extracts cellphone GPS and is segmented track
In class taxi (motor vehicle) GPS be segmented track；
Step 5: the segmentation track taxi GPS and class taxi GPS segmentation track are matched to using map-matching algorithm
On section in urban road network；
Step 6: utilizing of two kinds of track (taxi GPS is segmented track and class taxi GPS is segmented track)
With as a result, calculating separately the Vehicle Speed density fonction of respective stretch；
Step 7: for the section with two kinds of path matching result, using evidence theory to by two types
The section Vehicle Speed density fonction that calculates separately of path matching result merged, obtain to the road
The final estimation of section traffic behavior (Vehicle Speed)；
For there was only the section of a type of path matching result, will be calculated by the matching result of the track of single type
The obtained section Vehicle Speed density fonction is as the final estimation to the road section traffic volume state.
Further, in the step 2, GPS data from taxi is that low frequency records data, has one within average every 25 seconds
A user location records point, records the longitude and latitude and temporal information of user location.Mobile phone is obtained by various map APP software records
GPS data, cellphone GPS data are high frequency recording data, and general every 1 second user location records point, same to record user position
The longitude and latitude and temporal information set.
Further, in the step 3, the clustering algorithm uses DBSCAN clustering algorithm.DBSCAN clustering algorithm is
A kind of density-based algorithms are divided into different clusters by density to data set, and each cluster is classified as one kind.Data set is close
Degree is measured by the different attribute of data, and different attribute is different dimensions, then by European between calculating different data point
Distance measures the tightness degree between any two data point.Data set is divided into different clusters by DBSCAN clustering algorithm, is had
Two key parameters, the minimum neighborhood sample number MinPts of neighborhood distance ∈ and cluster inner core point, neighborhood distance ∈ are
The threshold value of Euclidean distance between above-mentioned two data point, MinPts are cluster inner core neighborhood of a point sample number threshold value (certain neighborhood of a point
It is core point that sample number, which is greater than MinPts), the ∈-neighborhood at any point contains at least one core point in a cluster.This
In invention, the dimension of data includes longitude, latitude and time, carries out standard scores processing to each data dimension respectively, and calculate warp
Euclidean distance after standard scores processing between each data point, determines the minimum neighborhood sample number of neighborhood distance ∈ and core point
MinPts is gathered the GPS data from taxi of each user and cellphone GPS data respectively for different clusters by DBSCAN cluster,
Each cluster represents one kind, one as studied section GPS track (travel locus).
Further, the neighborhood distance parameter ∈ in DBSCAN clustering algorithm is determined using the 4th neighborhood distance method, had
Body method are as follows: the 1) GPS data from taxi of each user is considered as a data set, calculated in each data set between any two points
The fourth-largest Euclidean distance value of each data set, i.e. the 4th neighborhood distance are found in Euclidean distance and from big to small arrangement；2) make the 4th
The probability distribution graph of neighborhood distance simultaneously determines the corresponding distance value of its inflection point as neighborhood distance parameter ∈.
Further, the Euclidean distance calculation formula in a certain data set between any two position record point is as follows:
Wherein: x '1With x '2Respectively indicate the longitude that latter two position record point is handled by standard scores, y '1With y '2Point
The latitude value of latter two position record point, t ' Biao Shi not be handled by standard scores1With t '2It respectively indicates after standard scores are handled
The time value of two position record points；Dimension data x any for any user, standard scores handle formula are as follows:
Wherein x ' is x by standard scores treated data,It is averaged for all initial data of the relative users dimensions
Value, σ are the standard deviation of all initial data of the relative users dimension.
Further, in the step 4, using isolated forest model as the identification model；Isolated forest model
(isolation forest) is a kind of unsupervised outlier detection method, is that a variety of statistical natures based on sample construct
Abnormality detection model.In the present invention, using the segmentation track taxi GPS as normal sample, binary tree, system are constructed as training set
Counting feature includes speed, acceleration, corner, path length and time of taxi vehicle driving etc.；Cellphone GPS is segmented track
It is tested as test sample by binary tree, thus Exception Filter track, the class extracted in cellphone GPS segmentation track goes out
It hires a car (motor vehicle) track, as the detection sample to road traffic condition.Specifically includes the following steps:
Taxi GPS obtained in step 3 is segmented track as training sample, with cellphone GPS stepped rail by step 4.1
Mark is as test sample；
Step 4.2 is based on training sample, constructs more binary trees (iTree) of isolated forest model；
Step 4.3, for each test sample, its path length in each binary tree is calculated, to obtain
Its abnormal score；
Step 4.4, the threshold value for determining abnormal score, the test sample that abnormal score is less than the threshold value regard as normal sample
This, i.e., the test sample is that class taxi GPS is segmented track.
Further, in the step 5, map-matching algorithm uses S-T map-matching algorithm.Based on space-time analysis
S-T map-matching algorithm is the GPS track matching algorithm based on space-time analysis, which considers the maximum of the time and space
Probability is by path matching to path.
Further, in the step 6, according to the matching result of the track of a certain type, the vehicle of respective stretch is calculated
The method of travel speed density fonction are as follows:
For certain a road section, obtain the track set of all the type for being matched to the section, by track gather in it is each
Track calculates separately to obtain a velocity amplitude, as a speed observation sample, forms the speed observation sample collection in the section；
The Vehicle Speed density fonction poss (v) in the section is calculated according to speed observation sample collection:
Wherein, v is speed, and n is the speed observation sample number that speed observation sample is concentrated, viIt is observed for i-th of speed
Sample；ρKIt (z) is kernel function, using gaussian kernel function:
Wherein, h indicates smoothing parameter, and the h the big, and then density function is more smooth, and Silverman rule can be used and determine h's
Value, i.e.,D is sample dimension (sample is speed, dimension 1 in the present invention), it can be seen that when
Sample number n is smaller, and the value of h is bigger, and vice versa, this is because being often not enough to provide when sample number is few enough
Distributed intelligence, biggish h value can embody this inaccurate estimation.
Further, the method for a velocity amplitude being obtained by a certain trajectory calculation are as follows: by this path matching to section
On true driving path length divided by the hourage of this section of track, obtain corresponding speed value.
Further, in the step 7, for a certain specific section, using evidence theory to by two kinds of track
The section Vehicle Speed density fonction that calculates separately of matching result merged, obtain handing over the section
The specific steps of logical state finally estimated are as follows:
Step 7.1 sets the road calculated separately by the segmentation track taxi GPS and class taxi GPS segmentation track
The Vehicle Speed density fonction of section is posst(v) and possm(v)；Respectively to posst(v) and possm(v) it carries out
Normalized:
posstn(v) and possmn(v) poss is respectively indicatedt(v) and possm(v) value after normalizing, posstn(v)∈
[0,1], possmn(v) [0,1] ∈；
Step 7.2 calculates fusion results:
possF(v)=posstn(v)*possmn(v)；
Step 7.3 calculates syncretizing effect index；
In command speed section [v1, v2] in, using the likelihood function in evidence theoryAs syncretizing effect index, the effect of fusion is measured, wherein a and norm is intermediate becomes
Amount,
PlsDempster([v1, v2]) value range be [0,1], be 1 when syncretizing effect it is best, two groups of data noncontradictories,
Conversely, contradiction is bigger；
Step 7.4, given threshold, if fusion index is lower than threshold value, corresponding fusion results are not used, and combine and go through
History data and the data in upstream and downstream section, using posst(v) and possm(v) in the data of historical data and upstream and downstream section
Final estimation of more close one as the road section traffic volume state；Otherwise using fusion results as the road section traffic volume state
Final estimation.
Further, command speed section [v1, v2] it is taken as [v15th, v85th], wherein v15thAnd v85thRespectively by matching
15 tantiles and 85 tantiles in all velocity amplitudes being calculated to the taxi GPS segmentation track in the section；If by
Being assigned to the number of all velocity amplitudes that the taxi GPS segmentation track in the section is calculated is N, by N number of velocity amplitude from it is small to
It is ranked up greatly；If 0.15N/0.85N is integer, ranking is 15 tantiles/85 quartiles in the velocity amplitude of 0.15N/0.85N
Value；If 0.15N/0.85N is not integer, the average value for two velocity amplitudes for taking ranking adjacent with 0.15N/0.85N is 15 points
Place value/85 tantiles.
The utility model has the advantages that
The present invention provides a kind of Traffic State Detection Methods based on multivariate data fusion, pass through GPS data from taxi
And the traffic behavior that cellphone GPS data (map APP data) obtain urban road is merged, data sample can greatly be increased,
The road conditions in the regions such as the city subsidiary road less for taxi and suburb also can be obtained effectively, can greatly improve data
Route coverage.Meanwhile cellphone GPS data sample data volume is big, recording frequency is high, can effectively reduce detection error, improves
The accuracy and validity of block status estimation.
The present invention merges the advantages of GPS data from taxi and cellphone GPS data, due to GPS data from taxi and cellphone GPS
Data are all easier to obtain in real time, using its two kinds of data can effectively large-scale real-time perception urban road traffic state,
Important reference value, mobile phone are provided to the real time monitoring of urban transportation and management for traffic and each administration of communications department
The application of GPS data also provides a kind of new approach to the research of urban transportation using traffic data for researchers
The present invention carries out GPS track using GPS data from taxi and cellphone GPS data of the DBSCAN clustering algorithm to user
Segmentation, using vehicle GPS segmentation track as research object, based on the taxi GPS track structure for being able to reflect taxi vehicle travelling state
It builds isolated forest model (isolation forest), extracts class taxi (motor vehicle) GPS track in cellphone GPS data,
To supplement road vehicle coverage information.For taxi GPS track and extracted class taxi GPS track, utilize
S-T map-matching algorithm carries out map match；Using the matching result of taxi GPS track and class taxi GPS track, respectively
Calculate the Vehicle Speed distribution of respective stretch；And by evidence theory to by taxi GPS track and class taxi GPS rail
The Vehicle Speed distribution that mark obtains section is merged, to obtain to section Vehicle Speed, i.e. traffic behavior
Final estimation.
This method merges two class GPS datas by introducing evidence theory, and result has after ensuring fusion
Effect property.
Detailed description of the invention
Fig. 1 is a kind of flow chart of Traffic State Detection Method based on multivariate data fusion of the present invention；
Fig. 2 is the evidence theory fusion process schematic of GPS data from taxi and cellphone GPS data；
Fig. 3 is the taxi wheel paths of GPS data from taxi and the motor-driven wheel paths in cellphone GPS data via S-T
Figure matching primitives section speed simultaneously carries out the section speed that two class track datas obtain by evidence theory fused early high
Peak (8:00~8:10) city road speed syncretizing effect schematic diagram.
Specific embodiment
The invention proposes a kind of Traffic State Detection Methods based on multivariate data fusion, as shown in Figure 1, and with Shenzhen
For city's road network, based on Shenzhen's GPS data from taxi and certain map APP software records to cellphone GPS data carried out reality
Border application.Time-sequencing is carried out to the data record of each user first, and clusters the stepped rail for obtaining user by DBSCAN
Then mark constructs training pattern to taxi GPS track using isolated forest model (isolation forest), and to mobile phone
GPS track carry out class taxi (motor vehicle) trajectory extraction, finally using S-T map-matching algorithm to taxi GPS track with
Map match is carried out by the motor-driven wheel paths of cellphone GPS trajectory extraction, obtains the vehicle driving in two class track different moments sections
VELOCITY DISTRIBUTION is fitted it and is merged using the section speed that evidence theory obtains two class tracks.
Shenzhen's road network is that Shenzhen simplifies version road network, altogether includes 21115 sections and 73415 intersections.
The GPS data from taxi is collected Shenzhen's taxi on September 15, to September totally one week on the 22nd 2016
GPS records data under taxi passenger carrying status, during which 13731 taxis are recorded in average every 25 seconds data record points altogether
Automobile-used family, and screen 13509 taxi passenger carrying status users.
The cellphone GPS data be the Residents in Shenzhen that arrives of certain map APP software records equally September in 2016 15 days extremely
September 22 days trip GPS datas.Recording frequency is that a GPS per second records point, and about 900,000 map is during which recorded altogether
APP user.
The neighborhood distance of the DBSCAN cluster calculates three longitude for considering GPS record data, latitude and time dimensions,
Standard scores processing is carried out to each dimension data first；Dimension data any for any user, standard scores handle formula are as follows:
Wherein: x is a certain initial data without standard scores processing of the user dimension, and x ' is that x is handled by standard scores
Data afterwards,For the average value of all initial data of the user dimension, σ is all initial data of the user dimension
Standard deviation；
Then the Euclidean distance between standard scores treated data point is calculated, and determines neighborhood distance parameter ∈；
The calculation of Euclidean distance is as follows:
Wherein: x '1With x '2Respectively indicate the longitude of the 2 GPS track points after standard scores are handled, y '1With y '2Respectively
Indicate the latitude value of the 2 GPS track points after standard scores are handled, t '1With t '2Respectively indicate two GPS after standard scores are handled
The time value of tracing point.
Neighborhood distance parameter ∈ determined by a kind of typical 4th neighborhood distance method, method particularly includes: 1) it will every
GPS data from taxi of a user after sequence is considered as a data set, calculates European between any two points in each data set
The fourth-largest Euclidean distance value of each data set, i.e. the 4th neighborhood distance are found in distance and from big to small arrangement；2) make the 4th neighborhood
The probability distribution graph of distance simultaneously determines the corresponding distance value of its inflection point as neighborhood distance parameter ∈；According to this method, this implementation
The final value for determining ∈ of example is 0.1.
Then determine that the minimum neighborhood sample number MinPts of core point in DBSCAN algorithm is 4 (the general values of MinPts),
And the segmentation track that cluster obtains user is carried out to the GPS record data that data are concentrated.
The isolated forest model (isolation forest) is a kind of unsupervised outlier detection model, training sample
This is segmented track for the taxi GPS after DBSCAN is clustered, and test sample is the cellphone GPS point after DBSCAN is clustered
Section track, the randomly drawing sample from training sample (as unit of user trajectory section) extract the various statistical nature structures of sample
Binary tree is made, training pattern is constructed, and test test sample, calculates the abnormal score of test sample, abnormal score is got over
Height illustrates that a possibility that sample is abnormal is higher.The outlier threshold for determining isolated forest test model is less than abnormal score different
The test sample of normal threshold value, it is believed that it has no significant difference with wheel paths are hired out, and determines it as class taxi (motor vehicle) GPS
It is segmented track.Selected data statistical characteristics include the average speed of track, middle bit rate, maximum speed, path length, put down
Equal acceleration, stop rate, average corner, corner transition rate, speed-shifting rate, terminus linear distance, the path length of track
And whether track occurs at wagon flow peak period (early 7:00-10:00, late 17:00-20:00).
The S-T map-matching algorithm considers the combination of data track point and road network, considers time and space
Maximum probability is by path matching to path, so that passage path length and trajectory time calculate the Vehicle Speed in path, and
Correspond to respective stretch.Specific algorithm are as follows:
1) for one section of track (p1→p2→…→pn), find the track candidate point of each tracing point(i is this section of rail
I-th of tracing point in mark, j are j-th of track candidate point of i-th of tracing point), wherein track candidate point position is track
Point makees the intersection point position of vertical line to section, and the length of vertical line need to be less than 35 meters；
2) track candidate point is calculatedObservation probability
In formula:For track candidate pointWith tracing point piBetween straight line geographic distance.
3) previous track candidate point is calculatedTo the latter track candidate pointTransition probability
In formula: di-1→ i is straight line geographic distance of (i-1)-th tracing point to i-th of tracing point, w(i-1, t) → (i, s)It is
The section candidate point of i-1 tracing pointTo the section candidate point of i-th of tracing pointShortest path in road network away from
From.
4) observation probability and transition probability are combined, track candidate point is obtainedTo track candidate pointCombined chance
For every section of track, selecting maximum a string of tracks candidate point of combined chance, (each track selects a track to wait
Reconnaissance) true driving trace on section is corresponded to as the track.
5) this section of track is calculated according to the hourage corresponding true driving path on section of this section of track
Route speed v1→n, and speed is corresponded into section.
In formula: w1→nThe true driving path length on section, Δ t are corresponded to for this section of track1→nFor the trip of this section of track
The row time.
It is a dynamic coverage problem in itself since all kinds of vehicles equipped with GPS device travel on section.So for
For a certain road, in real-time application, speed observation sample number is indefinite, while the prior information of VELOCITY DISTRIBUTION is often
It is uncertain, therefore the present invention estimates speed density using nonparametric technique-KERNEL FUNCTION METHOD, then is translated into possibility
Function.
For a certain specific section, after map match above, all tracks set for being matched to the section is obtained,
By track gather in each track calculate separately to obtain a velocity amplitude, as a speed observation sample, form the road
The speed observation sample collection of section；The speed in the section is calculated according to speed observation sample collection:
Wherein, v is speed, viFor i-th of speed observation sample on the section, n is that the speed on the section observes sample
This number；ρK(z) it is kernel function, the present invention is selected in the common gaussian kernel function of engineering field:
Wherein h indicates smoothing parameter, and the h the big, and then density function is more smooth, and Silverman rule can be used and determine taking for h
Value, i.e.,D is sample dimension (sample is speed, dimension 1 in the present invention), it can be seen that work as sample
This number n is smaller, and the value of h is bigger, and vice versa, this is because being often not enough to provide enough point when sample number is few
Cloth information, biggish h value can embody this inaccurate estimation.
Evidence theory is that one kind can be to the effective ways that uncertain data is merged, including support function and likelihood letter
Number supports function and verisimilitude function by calculating, the lower limit and the upper limit of the new data after available two kinds of data fusions, simultaneously
Fused effect can be measured.Melted in the present invention using the section speed that evidence theory obtains two class tracks
Specific step is as follows for conjunction:
Step 1: setting and the section that track and class taxi GPS segmentation track calculate separately is segmented by taxi GPS
Vehicle Speed density fonction be posst(v) and possm(v)；Respectively to posst(v) and possm(v) returned
One change processing:
posstn(v) and possmn(v) poss is respectively indicatedt(v) and possm(v) value after normalizing, posstn(v),
possmn(v) [0,1] ∈；[0,120] v ∈ in the present embodiment；
Step 2: calculate fusion results:
possF(v)=posstn(v)*possmn(v)；
Step 3: calculating syncretizing effect index；
In command speed section [v1, v2] in, using the likelihood function in evidence theoryAs syncretizing effect index, the effect of fusion is measured, wherein a and norm is intermediate becomes
Amount,
PlsDempster([v1, v2]) value range be [0,1], be 1 when syncretizing effect it is best, two groups of data noncontradictories,
Conversely, contradiction is bigger；
Step 4: given threshold (the present embodiment is taken as 0.5), if fusion index is lower than threshold value, corresponding fusion results are not
Use is given, and combines the data in historical data and upstream and downstream section, using posst(v) and possm(v) in historical data and on
A more close final estimation as the road section traffic volume state of the data of downstream road section；Otherwise using fusion results as this
The final estimation of road section traffic volume state.
GPS data from taxi and the evidence theory fusion process of cellphone GPS data are as shown in Figure 2.
For the machine extracted in the taxi wheel paths and cellphone GPS data of GPS data from taxi by isolating forest model
In the city road VELOCITY DISTRIBUTION result calculated after S-T map match, (morning peak is as a result, morning as shown in Figure 3 for motor-car track
Morning 8:00~8:10), it can be seen that cellphone GPS data can effectively supplement the road traffic shape that GPS data from taxi is lacked
State information.
Claims (10)
1. a kind of Traffic State Detection Method based on multivariate data fusion, which comprises the following steps:
Step 1: constructing urban road network according to urban road information；
Step 2: obtaining the GPS data from taxi and cellphone GPS data under passenger carrying status；
Step 3: being carried out respectively to GPS data from taxi and cellphone GPS data (continuous path point) by user using clustering algorithm
GPS track segmentation obtains each segmentation track taxi GPS and cellphone GPS segmentation track；
Step 4: using taxi GPS segmentation track as training sample, building identification model is extracted in cellphone GPS segmentation track
Class taxi GPS is segmented track；
Step 5: the segmentation track taxi GPS and class taxi GPS segmentation track are matched to city using map-matching algorithm
On section in road network；
Step 6: calculating separately the Vehicle Speed density point of respective stretch using the matching result of two kinds of track
Cloth function；
Step 7: for the section with two kinds of path matching result, using evidence theory to by two kinds of rail
The section Vehicle Speed density fonction that mark matching result calculates separately is merged, and is obtained to the section
Vehicle Speed, the i.e. final estimation of traffic behavior；
For there was only the section of a type of path matching result, will be calculated by the matching result of the track of single type
The section Vehicle Speed density fonction as the final estimation to the road section traffic volume state.
2. the Traffic State Detection Method according to claim 1 based on multivariate data fusion, which is characterized in that the step
In rapid two, each position record point in GPS data from taxi and cellphone GPS data includes longitude, latitude and the time three dimensions
The information of degree.
3. the Traffic State Detection Method according to claim 2 based on multivariate data fusion, which is characterized in that the step
In rapid three, the clustering algorithm uses DBSCAN clustering algorithm.
4. the Traffic State Detection Method according to claim 3 based on multivariate data fusion, which is characterized in that using the
Four neighborhood distance methods determine the neighborhood distance parameter ∈ in DBSCAN clustering algorithm, method particularly includes: 1) by each user
GPS data from taxi is considered as a data set, calculates the Euclidean distance in each data set between any two points and arranges from big to small
Find the fourth-largest Euclidean distance value of each data set, i.e. the 4th neighborhood distance；2) make the probability distribution graph of the 4th neighborhood distance simultaneously
Determine the corresponding distance value of its inflection point as neighborhood distance parameter ∈.
5. the Traffic State Detection Method according to claim 4 based on multivariate data fusion, which is characterized in that a certain number
It is as follows according to the Euclidean distance calculation formula for concentrating any two position to record between point:
Wherein: x '1With x '2Respectively indicate the longitude that latter two position record point is handled by standard scores, y '1With y '2Table respectively
Show the latitude value that latter two position record point is handled by standard scores, t '1With t '2It respectively indicates and handles latter two by standard scores
The time value of position record point；Dimension data x any for any user, standard scores handle formula are as follows:
Wherein x ' is x by standard scores treated data,For the average value of all initial data of the relative users dimensions, σ
For the standard deviation of all initial data of the relative users dimensions.
6. the Traffic State Detection Method according to claim 1 based on multivariate data fusion, which is characterized in that the step
In rapid four, using isolated forest model as the identification model；Specifically includes the following steps:
Taxi GPS obtained in step 3 is segmented track as training sample by step 4.1, is made with cellphone GPS segmentation track
For test sample；
Step 4.2 is based on training sample, constructs more binary trees of isolated forest model；
Step 4.3, for each test sample, its path length in each binary tree is calculated, so that it is different to obtain its
Chang get Fen；
Step 4.4, the threshold value for determining abnormal score, the test sample that abnormal score is less than the threshold value regard as normal sample, i.e.,
The test sample is that class taxi GPS is segmented track.
7. the Traffic State Detection Method according to claim 1 based on multivariate data fusion, which is characterized in that the step
In rapid six, according to the matching result of the track of a certain type, the Vehicle Speed density fonction of respective stretch is calculated
Method are as follows:
For certain a road section, obtain the track set of all the type for being matched to the section, by track gather in each rail
Mark calculates separately to obtain a velocity amplitude, as a speed observation sample, forms the speed observation sample collection in the section；According to
Speed observation sample collection calculates the Vehicle Speed density fonction poss (v) in the section:
Wherein, v is speed, and n is the speed observation sample number that speed observation sample is concentrated, viFor i-th of speed observation sample；
ρKIt (z) is kernel function, using gaussian kernel function:
Wherein, h indicates smoothing parameter, and the h the big, and then density function is more smooth, and the value that Silverman rule determines h can be used,
I.e.D is sample dimension, d=1.
8. the Traffic State Detection Method according to claim 7 based on multivariate data fusion, which is characterized in that by a certain
The method that trajectory calculation obtains a velocity amplitude are as follows: by the true driving path length on this path matching to section divided by
The hourage of this section of track obtains corresponding speed value.
9. the Traffic State Detection Method according to claim 7 based on multivariate data fusion, which is characterized in that the step
In rapid seven, for a certain specific section, the matching result by two kinds of track is calculated separately to obtain using evidence theory
The section Vehicle Speed density fonction merged, obtain to the road section traffic volume state finally estimate it is specific
Step are as follows:
Step 7.1 sets the section calculated separately by the segmentation track taxi GPS and class taxi GPS segmentation track
Vehicle Speed density fonction is posst(v) and possm(v)；Respectively to posst(v) and possm(v) normalizing is carried out
Change processing:
posstn(v) and possmn(v) poss is respectively indicatedt(v) and possm(v) value after normalizing, posstn(v) [0,1] ∈,
possmn(v) [0,1] ∈；
Step 7.2 calculates fusion results:
possF(v)=posstn(v)*possmn(v)；
Step 7.3 calculates syncretizing effect index；
In command speed section [v1, v2] in, using the likelihood function in evidence theoryMake
For syncretizing effect index, the effect of fusion is measured, wherein a and norm is intermediate variable,
Norm=max (possF(v))
PlsDempster([v1, v2]) value range be [0,1], be 1 when syncretizing effect it is best, two groups of data noncontradictories, conversely,
Contradiction is bigger；
Step 7.4, given threshold, if fusion index is lower than threshold value, corresponding fusion results are not used, and combine history number
According to and upstream and downstream section data, using posst(v) and possm(v) in more with historical data and the data in upstream and downstream section
Final estimation of close one as the road section traffic volume state；Otherwise using fusion results as the final of the road section traffic volume state
Estimation.
10. the Traffic State Detection Method according to claim 8 based on multivariate data fusion, which is characterized in that specified
Speed interval [v1, v2] it is taken as [v15th, v85th], wherein v15thAnd v85thRespectively divided by the taxi GPS for being matched to the section
15 tantiles and 85 tantiles in all velocity amplitudes that section trajectory calculation obtains；If by the taxi GPS for being matched to the section
The number for all velocity amplitudes that segmentation track is calculated is N, and N number of velocity amplitude is ranked up from small to large；If 0.15N/
0.85N is integer, then ranking is 15 tantiles/85 tantiles in the velocity amplitude of 0.15N/0.85N；If 0.15N/0.85N is not
It is integer, then the average value for two velocity amplitudes for taking ranking adjacent with 0.15N/0.85N is 15 tantiles/85 tantiles.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811524358.9A CN109410586B (en) | 2018-12-13 | 2018-12-13 | Traffic state detection method based on multi-metadata fusion |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811524358.9A CN109410586B (en) | 2018-12-13 | 2018-12-13 | Traffic state detection method based on multi-metadata fusion |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109410586A true CN109410586A (en) | 2019-03-01 |
| CN109410586B CN109410586B (en) | 2020-06-05 |
Family
ID=65458932
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811524358.9A Active CN109410586B (en) | 2018-12-13 | 2018-12-13 | Traffic state detection method based on multi-metadata fusion |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109410586B (en) |
Cited By (30)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110113716A (en) * | 2019-04-22 | 2019-08-09 | 腾讯大地通途（北京）科技有限公司 | A kind of path state information acquisition methods, device and storage medium |
| CN110264740A (en) * | 2019-06-24 | 2019-09-20 | 长沙理工大学 | The inhuman real-time track detector of traffic machine and detection method based on video |
| CN110285817A (en) * | 2019-07-12 | 2019-09-27 | 东北电力大学 | Complicated road network map matching process based on adaptive D-S evidence theory |
| CN110555544A (en) * | 2019-07-16 | 2019-12-10 | 广东工业大学 | traffic demand estimation method based on GPS navigation data |
| CN110570658A (en) * | 2019-10-23 | 2019-12-13 | 江苏智通交通科技有限公司 | Method for identifying and analyzing abnormal vehicle track at intersection based on hierarchical clustering |
| CN110634299A (en) * | 2019-10-25 | 2019-12-31 | 福州大学 | Urban traffic state fine division and identification method based on multi-source track data |
| CN110751196A (en) * | 2019-10-12 | 2020-02-04 | 东北石油大学 | Oil-like drop attachment identification method in oil-water two-phase flow transparent pipe wall |
| CN110766940A (en) * | 2019-09-24 | 2020-02-07 | 重庆交通大学 | Method for evaluating running condition of road signalized intersection |
| CN111081016A (en) * | 2019-12-18 | 2020-04-28 | 北京航空航天大学 | Urban traffic abnormity identification method based on complex network theory |
| CN111123333A (en) * | 2019-12-30 | 2020-05-08 | 公安部交通管理科学研究所 | Vehicle track positioning method fusing bayonet and GPS data |
| CN111179592A (en) * | 2019-12-31 | 2020-05-19 | 合肥工业大学 | Urban traffic prediction method and system based on spatio-temporal data flow fusion analysis |
| CN111190891A (en) * | 2019-12-27 | 2020-05-22 | 武汉长江通信产业集团股份有限公司 | Multi-semantic track data segmented storage method |
| CN111311910A (en) * | 2020-02-19 | 2020-06-19 | 中南大学 | Abnormal track detection method for multi-level road-level floating vehicle |
| CN111368881A (en) * | 2020-02-19 | 2020-07-03 | 浙江工业大学 | Low-frequency GPS track road network matching method based on multidimensional data fusion analysis |
| CN111739283A (en) * | 2019-10-30 | 2020-10-02 | 腾讯科技（深圳）有限公司 | Road condition calculation method, device, equipment and medium based on clustering |
| CN111862605A (en) * | 2020-07-20 | 2020-10-30 | 腾讯科技（深圳）有限公司 | Road condition detection method and device, electronic equipment and readable storage medium |
| CN111949896A (en) * | 2020-07-30 | 2020-11-17 | 安徽师范大学 | Track anomaly detection method based on track big data |
| CN112014859A (en) * | 2019-05-29 | 2020-12-01 | 安波福技术有限公司 | Estimating a velocity profile |
| CN112116813A (en) * | 2020-11-20 | 2020-12-22 | 深圳市城市交通规划设计研究中心股份有限公司 | State recognition model establishing method, road traffic state recognition method and road traffic state recognition device |
| CN112183589A (en) * | 2020-09-14 | 2021-01-05 | 西北工业大学 | Real-time vehicle K neighbor query method under low sampling rate |
| CN112382095A (en) * | 2020-11-26 | 2021-02-19 | 长沙理工大学 | Urban expressway traffic state estimation method based on multi-source data fusion |
| CN112712729A (en) * | 2019-10-26 | 2021-04-27 | 华为技术有限公司 | Method and system for predicting motion trajectory |
| CN112837393A (en) * | 2019-11-22 | 2021-05-25 | 中国航天系统工程有限公司 | Method and system for generating extra-large city vector road network based on vehicle position data |
| CN113053107A (en) * | 2020-12-23 | 2021-06-29 | 沈阳世纪高通科技有限公司 | Method and device for calculating road congestion and slowness state by combining free flow |
| CN113259900A (en) * | 2021-05-27 | 2021-08-13 | 华砺智行(武汉)科技有限公司 | Distributed multi-source heterogeneous traffic data fusion method and device |
| CN113375685A (en) * | 2021-03-31 | 2021-09-10 | 福建工程学院 | Urban intersection center identification and intersection turning rule extraction method based on sub-track intersection |
| CN113486719A (en) * | 2021-06-08 | 2021-10-08 | 南京邮电大学 | Vehicle destination prediction method, vehicle destination prediction device, computer equipment and storage medium |
| WO2021218314A1 (en) * | 2020-04-27 | 2021-11-04 | 深圳壹账通智能科技有限公司 | Event identification method and apparatus based on position locating, and device and storage medium |
| CN113743767A (en) * | 2021-08-27 | 2021-12-03 | 广东工业大学 | Vehicle dispatching method, system, computer and medium based on time and safety |
| CN117475639A (en) * | 2023-12-26 | 2024-01-30 | 深圳市城市交通规划设计研究中心股份有限公司 | Bridge vehicle space-time distribution identification method integrating vehicle positioning data of different frequencies |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160119321A (en) * | 2015-04-02 | 2016-10-13 | 김진영 | Method for taxi call service |
| CN107146404A (en) * | 2016-03-01 | 2017-09-08 | 香港生产力促进局 | The information processing system and its method of a kind of real-time collecting and transmission traffic data |
| CN108335482A (en) * | 2017-01-20 | 2018-07-27 | 亚信蓝涛（江苏）数据科技有限公司 | A kind of urban transportation Situation Awareness method and method for visualizing |
| CN108538054A (en) * | 2018-05-17 | 2018-09-14 | 北京中交汇智数据有限公司 | A kind of method and system obtaining traffic information based on mobile phone signaling data |
| CN108806250A (en) * | 2018-06-08 | 2018-11-13 | 北京航空航天大学 | A kind of area traffic jamming evaluation method based on speed sampling data |
-
2018
- 2018-12-13 CN CN201811524358.9A patent/CN109410586B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20160119321A (en) * | 2015-04-02 | 2016-10-13 | 김진영 | Method for taxi call service |
| CN107146404A (en) * | 2016-03-01 | 2017-09-08 | 香港生产力促进局 | The information processing system and its method of a kind of real-time collecting and transmission traffic data |
| CN108335482A (en) * | 2017-01-20 | 2018-07-27 | 亚信蓝涛（江苏）数据科技有限公司 | A kind of urban transportation Situation Awareness method and method for visualizing |
| CN108538054A (en) * | 2018-05-17 | 2018-09-14 | 北京中交汇智数据有限公司 | A kind of method and system obtaining traffic information based on mobile phone signaling data |
| CN108806250A (en) * | 2018-06-08 | 2018-11-13 | 北京航空航天大学 | A kind of area traffic jamming evaluation method based on speed sampling data |
Cited By (46)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110113716B (en) * | 2019-04-22 | 2021-08-31 | 腾讯大地通途（北京）科技有限公司 | Path state information acquisition method and device and storage medium |
| CN110113716A (en) * | 2019-04-22 | 2019-08-09 | 腾讯大地通途（北京）科技有限公司 | A kind of path state information acquisition methods, device and storage medium |
| CN112014859A (en) * | 2019-05-29 | 2020-12-01 | 安波福技术有限公司 | Estimating a velocity profile |
| CN110264740A (en) * | 2019-06-24 | 2019-09-20 | 长沙理工大学 | The inhuman real-time track detector of traffic machine and detection method based on video |
| CN110285817A (en) * | 2019-07-12 | 2019-09-27 | 东北电力大学 | Complicated road network map matching process based on adaptive D-S evidence theory |
| CN110555544A (en) * | 2019-07-16 | 2019-12-10 | 广东工业大学 | traffic demand estimation method based on GPS navigation data |
| CN110555544B (en) * | 2019-07-16 | 2022-04-08 | 广东工业大学 | Traffic demand estimation method based on GPS navigation data |
| CN110766940A (en) * | 2019-09-24 | 2020-02-07 | 重庆交通大学 | Method for evaluating running condition of road signalized intersection |
| CN110751196A (en) * | 2019-10-12 | 2020-02-04 | 东北石油大学 | Oil-like drop attachment identification method in oil-water two-phase flow transparent pipe wall |
| CN110751196B (en) * | 2019-10-12 | 2020-09-18 | 东北石油大学 | Oil-like drop attachment identification method in oil-water two-phase flow transparent pipe wall |
| CN110570658A (en) * | 2019-10-23 | 2019-12-13 | 江苏智通交通科技有限公司 | Method for identifying and analyzing abnormal vehicle track at intersection based on hierarchical clustering |
| CN110570658B (en) * | 2019-10-23 | 2022-02-01 | 江苏智通交通科技有限公司 | Method for identifying and analyzing abnormal vehicle track at intersection based on hierarchical clustering |
| CN110634299A (en) * | 2019-10-25 | 2019-12-31 | 福州大学 | Urban traffic state fine division and identification method based on multi-source track data |
| CN112712729A (en) * | 2019-10-26 | 2021-04-27 | 华为技术有限公司 | Method and system for predicting motion trajectory |
| CN111739283B (en) * | 2019-10-30 | 2022-05-20 | 腾讯科技（深圳）有限公司 | Road condition calculation method, device, equipment and medium based on clustering |
| CN111739283A (en) * | 2019-10-30 | 2020-10-02 | 腾讯科技（深圳）有限公司 | Road condition calculation method, device, equipment and medium based on clustering |
| CN112837393B (en) * | 2019-11-22 | 2024-04-09 | 中国航天系统工程有限公司 | Method and system for generating oversized city vector road network based on vehicle position data |
| CN112837393A (en) * | 2019-11-22 | 2021-05-25 | 中国航天系统工程有限公司 | Method and system for generating extra-large city vector road network based on vehicle position data |
| CN111081016B (en) * | 2019-12-18 | 2021-07-06 | 北京航空航天大学 | Urban traffic abnormity identification method based on complex network theory |
| CN111081016A (en) * | 2019-12-18 | 2020-04-28 | 北京航空航天大学 | Urban traffic abnormity identification method based on complex network theory |
| CN111190891A (en) * | 2019-12-27 | 2020-05-22 | 武汉长江通信产业集团股份有限公司 | Multi-semantic track data segmented storage method |
| CN111190891B (en) * | 2019-12-27 | 2023-07-25 | 武汉长江通信产业集团股份有限公司 | Multi-semantic track data segment storage method |
| CN111123333B (en) * | 2019-12-30 | 2022-05-03 | 公安部交通管理科学研究所 | Vehicle track positioning method fusing bayonet and GPS data |
| CN111123333A (en) * | 2019-12-30 | 2020-05-08 | 公安部交通管理科学研究所 | Vehicle track positioning method fusing bayonet and GPS data |
| CN111179592A (en) * | 2019-12-31 | 2020-05-19 | 合肥工业大学 | Urban traffic prediction method and system based on spatio-temporal data flow fusion analysis |
| CN111179592B (en) * | 2019-12-31 | 2021-06-11 | 合肥工业大学 | Urban traffic prediction method and system based on spatio-temporal data flow fusion analysis |
| CN111368881A (en) * | 2020-02-19 | 2020-07-03 | 浙江工业大学 | Low-frequency GPS track road network matching method based on multidimensional data fusion analysis |
| CN111311910A (en) * | 2020-02-19 | 2020-06-19 | 中南大学 | Abnormal track detection method for multi-level road-level floating vehicle |
| CN111311910B (en) * | 2020-02-19 | 2020-12-22 | 中南大学 | Abnormal track detection method for multi-level road-level floating vehicle |
| CN111368881B (en) * | 2020-02-19 | 2023-10-03 | 浙江工业大学 | Low-frequency GPS track road network matching method based on multidimensional data fusion analysis |
| WO2021218314A1 (en) * | 2020-04-27 | 2021-11-04 | 深圳壹账通智能科技有限公司 | Event identification method and apparatus based on position locating, and device and storage medium |
| CN111862605A (en) * | 2020-07-20 | 2020-10-30 | 腾讯科技（深圳）有限公司 | Road condition detection method and device, electronic equipment and readable storage medium |
| CN111862605B (en) * | 2020-07-20 | 2022-03-08 | 腾讯科技（深圳）有限公司 | Road condition detection method and device, electronic equipment and readable storage medium |
| CN111949896B (en) * | 2020-07-30 | 2023-09-15 | 安徽师范大学 | Track abnormality detection method based on track big data |
| CN111949896A (en) * | 2020-07-30 | 2020-11-17 | 安徽师范大学 | Track anomaly detection method based on track big data |
| CN112183589A (en) * | 2020-09-14 | 2021-01-05 | 西北工业大学 | Real-time vehicle K neighbor query method under low sampling rate |
| CN112183589B (en) * | 2020-09-14 | 2022-04-22 | 西北工业大学 | Real-time vehicle K neighbor query method under low sampling rate |
| CN112116813A (en) * | 2020-11-20 | 2020-12-22 | 深圳市城市交通规划设计研究中心股份有限公司 | State recognition model establishing method, road traffic state recognition method and road traffic state recognition device |
| CN112382095A (en) * | 2020-11-26 | 2021-02-19 | 长沙理工大学 | Urban expressway traffic state estimation method based on multi-source data fusion |
| CN113053107B (en) * | 2020-12-23 | 2022-08-02 | 沈阳世纪高通科技有限公司 | Method and device for calculating road congestion and slowness state by combining free flow |
| CN113053107A (en) * | 2020-12-23 | 2021-06-29 | 沈阳世纪高通科技有限公司 | Method and device for calculating road congestion and slowness state by combining free flow |
| CN113375685A (en) * | 2021-03-31 | 2021-09-10 | 福建工程学院 | Urban intersection center identification and intersection turning rule extraction method based on sub-track intersection |
| CN113259900A (en) * | 2021-05-27 | 2021-08-13 | 华砺智行(武汉)科技有限公司 | Distributed multi-source heterogeneous traffic data fusion method and device |
| CN113486719A (en) * | 2021-06-08 | 2021-10-08 | 南京邮电大学 | Vehicle destination prediction method, vehicle destination prediction device, computer equipment and storage medium |
| CN113743767A (en) * | 2021-08-27 | 2021-12-03 | 广东工业大学 | Vehicle dispatching method, system, computer and medium based on time and safety |
| CN117475639A (en) * | 2023-12-26 | 2024-01-30 | 深圳市城市交通规划设计研究中心股份有限公司 | Bridge vehicle space-time distribution identification method integrating vehicle positioning data of different frequencies |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109410586B (en) | 2020-06-05 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109410586A (en) | A kind of Traffic State Detection Method based on multivariate data fusion | |
| CN105788273B (en) | The method of urban intersection automatic identification based on low precision space-time trajectory data | |
| CN105957342B (en) | Track grade road plotting method and system based on crowdsourcing space-time big data | |
| CN107766808A (en) | The method and system that Vehicle Object motion track clusters in road network space | |
| CN106781479B (en) | A method of highway operating status is obtained based on mobile phone signaling data in real time | |
| Wang et al. | Automatic intersection and traffic rule detection by mining motor-vehicle GPS trajectories | |
| CN104330089B (en) | A kind of method that map match is carried out using history gps data | |
| Yang et al. | Generating lane-based intersection maps from crowdsourcing big trace data | |
| CN110909788B (en) | Statistical clustering-based road intersection position identification method in track data | |
| Rogers et al. | Mining GPS data to augment road models | |
| CN102607553B (en) | Travel track data-based stroke identification method | |
| CN102521973B (en) | A kind of mobile phone switches the road matching method of location | |
| CN107121690B (en) | A kind of dwell point recognition methods and device based on parameter of doing more physical exercises | |
| CN109000645A (en) | Complex environment target classics track extracting method | |
| CN108961758B (en) | Road junction widening lane detection method based on gradient lifting decision tree | |
| CN105241465B (en) | A kind of method of road renewal | |
| CN108959466A (en) | Taxi hot spot method for visualizing and system based on BCS-DBSCAN | |
| CN109459047A (en) | Vehicle high-precision location data is accurately matched with navigation map and lane change behavior detection method | |
| CN109919347A (en) | Road conditions generation method, relevant apparatus and equipment | |
| CN106710208A (en) | Traffic state acquisition method and device | |
| CN104900057B (en) | A kind of Floating Car map-matching method in the major-minor road of city expressway | |
| CN105740505A (en) | GPS-RTK technology based road space line shape recovery method | |
| CN110413855B (en) | Region entrance and exit dynamic extraction method based on taxi boarding point | |
| CN113903173B (en) | Vehicle track feature extraction method based on directed graph structure and LSTM | |
| Sun et al. | Deep learning-based probability model for traffic information estimation |
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 |