CN110298500A - A kind of urban transportation track data set creation method based on taxi car data and city road network - Google Patents
A kind of urban transportation track data set creation method based on taxi car data and city road network Download PDFInfo
- Publication number
- CN110298500A CN110298500A CN201910532080.8A CN201910532080A CN110298500A CN 110298500 A CN110298500 A CN 110298500A CN 201910532080 A CN201910532080 A CN 201910532080A CN 110298500 A CN110298500 A CN 110298500A
- Authority
- CN
- China
- Prior art keywords
- region
- vehicle
- road
- data
- city
- 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
- 238000000034 method Methods 0.000 title claims abstract description 46
- 238000010200 validation analysis Methods 0.000 claims abstract description 15
- 238000002360 preparation method Methods 0.000 claims abstract description 9
- 230000011218 segmentation Effects 0.000 claims abstract description 9
- 230000001133 acceleration Effects 0.000 claims description 28
- 238000009826 distribution Methods 0.000 claims description 25
- 230000006870 function Effects 0.000 claims description 23
- 238000004458 analytical method Methods 0.000 claims description 21
- 238000004364 calculation method Methods 0.000 claims description 15
- 239000011159 matrix material Substances 0.000 claims description 15
- 230000005484 gravity Effects 0.000 claims description 14
- 238000013480 data collection Methods 0.000 claims description 12
- 238000007667 floating Methods 0.000 claims description 11
- 238000011160 research Methods 0.000 claims description 10
- 238000004088 simulation Methods 0.000 claims description 10
- 230000002159 abnormal effect Effects 0.000 claims description 7
- 238000001514 detection method Methods 0.000 claims description 6
- 238000013316 zoning Methods 0.000 claims description 6
- 206010039203 Road traffic accident Diseases 0.000 claims description 4
- 238000004422 calculation algorithm Methods 0.000 claims description 4
- 230000008569 process Effects 0.000 claims description 4
- 238000012942 design verification Methods 0.000 claims description 3
- 230000033001 locomotion Effects 0.000 claims description 3
- 238000013139 quantization Methods 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 2
- 238000000605 extraction Methods 0.000 claims description 2
- 238000003064 k means clustering Methods 0.000 claims description 2
- 238000012986 modification Methods 0.000 claims description 2
- 230000004048 modification Effects 0.000 claims description 2
- 238000007781 pre-processing Methods 0.000 claims description 2
- 235000013399 edible fruits Nutrition 0.000 claims 1
- 238000012795 verification Methods 0.000 abstract description 2
- SAZUGELZHZOXHB-UHFFFAOYSA-N acecarbromal Chemical compound CCC(Br)(CC)C(=O)NC(=O)NC(C)=O SAZUGELZHZOXHB-UHFFFAOYSA-N 0.000 abstract 1
- 239000008186 active pharmaceutical agent Substances 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 238000013508 migration Methods 0.000 description 3
- 230000005012 migration Effects 0.000 description 3
- 230000006855 networking Effects 0.000 description 3
- 241001269238 Data Species 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 230000008676 import Effects 0.000 description 2
- 239000000203 mixture Substances 0.000 description 2
- 241000894007 species Species 0.000 description 2
- 241000208340 Araliaceae Species 0.000 description 1
- 235000005035 Panax pseudoginseng ssp. pseudoginseng Nutrition 0.000 description 1
- 235000003140 Panax quinquefolius Nutrition 0.000 description 1
- 238000013459 approach Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000006243 chemical reaction Methods 0.000 description 1
- 238000004140 cleaning Methods 0.000 description 1
- 239000012141 concentrate Substances 0.000 description 1
- 230000001186 cumulative effect Effects 0.000 description 1
- 238000005520 cutting process Methods 0.000 description 1
- 238000013524 data verification Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 230000018109 developmental process Effects 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 235000008434 ginseng Nutrition 0.000 description 1
- 238000009434 installation Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000001788 irregular Effects 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000012913 prioritisation Methods 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000000630 rising effect Effects 0.000 description 1
- 239000000523 sample Substances 0.000 description 1
- 238000000638 solvent extraction Methods 0.000 description 1
- 238000003860 storage Methods 0.000 description 1
- 230000029305 taxis Effects 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
- G06Q10/043—Optimisation of two dimensional placement, e.g. cutting of clothes or wood
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/10—Services
- G06Q50/26—Government or public services
Landscapes
- Business, Economics & Management (AREA)
- Human Resources & Organizations (AREA)
- Engineering & Computer Science (AREA)
- Strategic Management (AREA)
- Tourism & Hospitality (AREA)
- Economics (AREA)
- General Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Marketing (AREA)
- Theoretical Computer Science (AREA)
- Development Economics (AREA)
- Physics & Mathematics (AREA)
- Educational Administration (AREA)
- Health & Medical Sciences (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Game Theory and Decision Science (AREA)
- Entrepreneurship & Innovation (AREA)
- Operations Research (AREA)
- Quality & Reliability (AREA)
- Traffic Control Systems (AREA)
Abstract
A kind of urban transportation track data set creation method based on taxi car data and city road network, belongs to traffic and transport field.The model is made of preparation layer, generation layer and validation layer, generates the track data of private car based on taxi car data and city road network information for the first time, while proposing the verification method of complete set to verify to the data set accuracy of generation.The present invention mentions the city function limited region dividing method based on neighbouring lane segmentation method, and city is divided into different functional areas.It proposes regional population's weight opportunity model simultaneously, can more accurately portray population moving mode.To verify authenticity and accuracy of the invention, the authenticity and accuracy of the data set that the model based on both macro and micro angle generates come authentication model are proposed respectively in validation layer.By the verifying of validation layer, the track data for the private car that the present invention is generated according to GPS data from taxi and the city vehicle network information, authenticity and accuracy with higher.
Description
Technical field
The invention belongs to traffic and transport field, it is related in traffic big data field generating society based on floating vehicle track data
The method of meeting vehicle track data collection.
Background technique
Internet of Things is the important component of generation information technology, is the important stage of information age development.With more
Carry out more vehicle accesses Internet of Things (Internet of Things), car networking (Internet of Vehicles) technology
Universal and application is in the ascendant.Mankind's mobile trajectory data is the basis of vehicle communication in car networking, is able to reflect space-time environment
Under individual Move Mode and social Transport.Pass through public vehicles (such as private car) and floating vehicle (such as taxi, public affairs
Hand over vehicle etc.) track data, complicated and diversified vehicle social networks (VSNs) can be constructed.The track data collection of floating car,
It can be obtained from internet easily.However, due to the protection and some relevant government policies of individual privacy, it is general to grind
Study carefully the almost impossible track data for obtaining public vehicles of personnel, this is so very big that hinder further investigation and the correlation neck of researcher
The progress in domain.
Summary of the invention
The purpose of the present invention generates the shortcoming of research mainly for available data collection, proposes a kind of three layers of RDMP
(Region Division and Mobility Pattern based Vehicle Trajectory Dataset
Generation Model) track of vehicle data set generation model, the model include preparation layer, generation layer and validation layer, for the first time
The track data of private car is generated based on taxi car data and city road network information, while proposing that the verification method of complete set comes
The data set accuracy of generation is verified.
Technical solution of the present invention:
A kind of urban transportation track data set creation method based on taxi car data and city road network, the city are handed over
Logical track data set creation method is completed in three layers of RDMP track of vehicle data set generation model, described three layers
RDMP track of vehicle data set generation model includes that steps are as follows for preparation layer, generation layer and validation layer:
S1: in preparation layer, pre-processing original taxi track data, deletes useless and abnormal data；Together
When using the city function limited region dividing method based on neighbouring lane segmentation method carry out urban area division, and city road network is carried out
Building, includes the following steps:
S1.1: pretreatment: deleting the useless track data under complete vehicle curb condition, and cleaning is led due to equipment precision and statistical error
The absurd data of cause obtain available taxi track data；
S1.2: the city of research is divided using mankind's motion track and POI point of interest based on neighbouring lane segmentation method
For different regions, the trip of all vehicles is converted into the circulation between different zones, detailed process is as follows:
S1.2.1: it is using the DMR-based probability topic model of unsupervised learning, city is whole as a document,
The different functions that each region is possessed are as theme, by interregional Move Mode as word, by POIs feature vector
As the metadata of document, frequency density v of i-th kind of POIs in the r of regionI, rCalculation formula are as follows:
Wherein, NumiRepresent the quantity of the POIs of the i-th seed type in the r of region, SrRepresent the area of region r, the POI of region r
Feature vector is denoted as xr=(v1, r, v2, r..., vI, r..., vF, r, 1), the metadata of region r is represented, wherein F is POI in r
Species number, last vector 1 is default feature；The theme distribution of region r is K dimensional vector θr=(θR, 1, θR, 2..., θR, e...,
θR, K), θR, eRepresent ratio shared by theme e in the r of region；
S1.2.2: k-means clustering algorithm is used, theme distribution obtained in S1.2.1 is clustered, each travelling
Beginning and end coordinate information be input in Density Estimator KDE model, the functional strength in quantization function region；Equipped with n
A region, by Density Estimator model KDE model, using Density Estimator amount come the functional strength of zoning r:
Wherein, dI, rThe distance of region i to region r is represented, R represents bandwidth, and KF () represents gaussian kernel function；Carry out
After the estimation of complete functional strength, the region after dividing, the actual functional capability in the city Lai Fanying are annotated, and definition region r number belongs to
Property is Ka；
S1.2.3: region clustering
Using neighbouring lane segmentation method, strategic road is considered as line segment, and grid regards node as, calculates each grid to road
Euclidean distance, and record the nearest road number K of each grid distancel, to each of rasterizing map after calculating
Grid is clustered, by KaAnd KlIt is worth node all the same as a cluster；
The judgment principle of strategic road are as follows: be that 0.001*0.001 divides net according to longitude and latitude by city map rasterizing
Lattice go out whole day average traffic 100,000 or more road extractions as strategic road.
S1.3: according to practical study region, to the original document downloaded from the website Street Map Open carry out cutting and
Layering, meanwhile, the road in manual modification real world, which is gone on a journey, to be limited, and updates newest traffic condition, to the road in city
Network is constructed；
S2: in generation layer:
S2.1: according to the staticaccelerator scale of floating vehicle and public vehicles in different function area, each region is calculated
The quantity of public vehicles, calculation formula are as follows:
Wherein, SAiFor the private car trip sum of region i；SGjRepresent the taxi in each grid j for including in the i of region
Vehicle quantity, NiIndicate the lattice number divided in the i of region, αiRepresent the ratio of private car and taxi in the i of region, SRwIt indicates
The quantity of taxi in the i of region in grid j in road w, njRepresent the number of road in grid j；
S2.2: it proposes regional population's weight chance RPWO model, calculates the origin and destination OD matrix of public vehicles, including following
Step:
S2.2.1: setting a land j to the attraction of starting area i, with using the center of gravity of region j as the center of circle, to rise
The center of gravity and the center of gravity distance R of land j of point region iijFor the total number of people Q in the border circular areas of radiusjiIt is inversely proportional；
S2.2.2: it is for center of gravityRegion, calculation method is as follows:
Wherein, L indicates the grid number in the region；xlAnd ylRespectively indicate the opposite longitude and relative altitude of grid in region；
S2.2.3: according to the center of gravity in beginning and end region, the attraction to land j to starting area i is calculated
Aji, calculation method is as follows:
Wherein, QjiIt indicates with RijFor radius, it is total as the private car in the border circular areas in the center of circle using the center of gravity of land j
Number；NsIt is the number in region included in circle, βrRepresent the ratio that region r is in border circular areas；The gross area of region r is
Sr, the area that region r is in border circular areas isPrIt is the private car number of region r, AjiRepresent land j relative to
The attraction of starting area i, ojThe private car number of land j is represented, M represents the private car total number in entire city；
S2.2.4: the attraction based on each region, the volume of traffic of the zoning i to region j:
Wherein, SAiTo go out number of lines from the private car of region i, it is calculated by S2.1, n is the region divided in city
Number；
S2.3: simulation track data are generated using SUMO emulation tool in conjunction with OD matrix, steps are as follows:
S2.3.1: classified according to the region divided to urban road；Road network file includes that every road connects
The longitude and latitude of contact, calculates which region road belongs to using longitude and latitude, and each region Zhong Bao is written in road network file
The road ID contained；
S2.3.2: it using the OD2TRIPS plug-in unit in SUMO tool, imports O/D matrix and is broken down into single unit vehicle row
Journey inputs the road column for including in the O/D matrix, road network file, region of S2.2 generation according to the concrete condition and data in city
Table, and be arranged and generate period, the trip proportion of each period, generation type of vehicle parameter, generate a series of vehicle travels letters
The xml document of breath, each travel information include vehicle ID, departure time, departure place ID and destination ID；
S2.3.3: the path locus information in order to generate private vehicle, it is defeated using DUAROUTER plug-in unit in SUMO tool
The trip information that approach online article part and OD2TRIPS are generated, is arranged simulation time section, shortest path calculation method, generates simulation
Vehicle route, including vehicle ID, the travel time, road by situation, ultimately produce vehicle driving trace information；
S2.3.4: to generate the information per second for generating vehicle in specified time interval, including longitude and latitude position, traveling angle
Degree, instantaneous velocity, place road number, using the Trace File Generation plug-in unit in SUMO tool,
Corresponding information is inputted in DUAROUTER function, while writing corresponding configuration file, sets △ t minutes for time interval, and
Generate vehicle tracking file in different time periods；
S3: in validation layer, the design verification model in terms of both macro and micro two, verifying generate the accuracy of data respectively
And authenticity, comprising the following steps:
S3.1: in macromodel, by the track data of generation, analysis comparison is carried out with actual traffic situation, including hand over
Through-current capacity comparison, Travel Range comparison, traffic condition comparison；
S3.2: in microcosmic verifying model, the true of data is generated from acceleration and the angle analysis of relative distance and evaluation
Reality, a kind of method for devising quantitative detection track data, steps are as follows:
S3.2.1: generated track data acceleration and its gradient value J are verified in acceleration analysise[m/s3] accuracy,
When result precision in the reasonable scope, then prove the data set generated inherent reasonability with higher, using three indexs come
Carry out acceleration gradient analysis: JeGreater than ± 3m/s3J in the track data percentage of threshold value, data seteIn maximum value, data set
JeMinimum value；
S3.2.2: consistency analysis verifies the authenticity for generating data set from the distance between vehicle, when in data set
In the presence of abnormal small distance interval, then the accuracy of the data set is doubtful；
S3.2.3: vehicle to the distance between calculation formula it is as follows:
The vehicle spacing in T moment vehicle v1 and v2 is represented,WithMean
Road starting point end is arrived a little respectivelyWithRoad alignment length, also
It is the projection of the road alignment of actual vehicle position；
S3.2.4: vehicle is set on road always along straight-line travelling, directly from the corresponding projection coordinate of vehicle geographical coordinate
Calculate the distance between vehicle:
WhenAt least when dropping to 5m or less in a flash, two vehicles can collide, and lead to traffic accident, because
This, passes through detectionWhen ratio shared in total data set, carry out the accuracy of validation data set.
Beneficial effects of the present invention: the present invention realizes private savings by three layers of track of vehicle data set generation model of building
The generation and verifying of vehicle mobile trajectory data.The present invention combines main roads network and trip track data, proposes one
The new city function limited region dividing method of kind.The flow pattern for having analysed in depth city vehicle simultaneously proposes interregional population power
Weight opportunity model, it is contemplated that the connection between the functional area of different cities depicts intercity vehicle migration model well
And rule.In addition, the present invention constructs verifying model from two angles of both macro and micro, the data set generation side of proposition is demonstrated
The private car track data that method generates has accuracy and use value.It is generated the present invention provides private car track data collection
A kind of new method can provide strong data and support for car networking and traffic study work.
Detailed description of the invention
Fig. 1 is RDMP model proposed by the invention, which is three layers of track of vehicle data set generation model, by preparing
Layer, generation layer and validation layer composition.
Fig. 2 be it is proposed that ARS region partitioning method.By taking Beijing as an example, city can be divided into 153 regions.
Fig. 3 is that private savings car data is compared with true data in present invention Beijing five rings generated, in main roads
On traffie distribution accounting figure.
Fig. 4 is the travel time accounting of private savings car data and true data in present invention Beijing five rings generated
Figure.Wherein, (a) is 1 point to 12 points of travel time accounting；(b) it is 13 points to 24 points of travel time accounting.
Fig. 5 is the trip distance accounting figure of private savings car data and true data in invention Beijing five rings generated.
Fig. 6 is between the simulaed path generated using Baidu map API navigation Service and present invention data set generated
Comparison.As the result is shown when within trip route is 5km with 5-10km, the two distribution is essentially identical.Wherein, (a) is represented
Walking along the street diameter is (b) to indicate that trip route is 5-10km within 5km.
Fig. 7 is the comparison of real data set and generation data set in terms of acceleration.Wherein, (a) is that real data set exists
The vehicle acceleration distribution map of 7:00-7:15；It (b) is vehicle acceleration distribution map of the real data set in 7:15-7:30；(c)
To generate data set in the vehicle acceleration distribution map of 7:00-7:15；(d) add to generate vehicle of the data set in 7:15-7:30
Velocity contour.
Fig. 8 is the distribution map of real data set and generation data set in terms of relative distance.
Specific embodiment
It to make the objectives, technical solutions, and advantages of the present invention clearer, below will be to specific embodiment party of the invention
Formula is described in further detail.
Present example has used the taxi track data collection in Beijing in November, 2012 (wherein comprising about 2.7 ten thousand
Taxi is more than 10,000,000,000 GPS record), it provides and public vehicles track data collection is generated based on floating vehicle track data
Method.The track of vehicle data set generation model (RDMP) that core of the invention is three layers, structure is as shown in Figure 1.The model
It is made of preparation layer, generation layer and validation layer, private car can be generated based on GPS data from taxi and the city vehicle network information
Track data.This method comprises:
S1 is the embodiment of the preparation layer of RDMP model in the present invention, and in this layer, we first carry out data
Pretreatment, obtains available taxi track data.By preparation layer, lay the foundation for next work.
S1.1: data set pretreatment.In instances, we used Pekinese, China's in November, 2012 to hire out wheel paths number
According to wherein being recorded comprising the GPS that about 2.7 ten thousand taxis are more than 10,000,000,000.The location information of taxi is by GPS device with every
11 seconds 1 time frequency updates.Raw data file is stored in the text document of storage time name.There are two types of go out for taxi
Row mode, cargo mode and empty wagons mode.Only under cargo mode, the driving trace of taxi can be just it is regular,
There can be similar trip mode with private car.On the contrary, taxi is planless traveling under empty wagons mode, this is right
Our research is nonsensical.Therefore, when taxi does not have passenger, we delete useless track data.In addition,
Due to equipment precision and statistical error, some data are obviously absurd.For example, some route are too long or too short, can not obtain
Valuable analysis is as a result, the wrong data of this part is also required to be removed.Then we handle the data after simplifying,
To obtain vehicle travel each time.We extract identical vehicle ID in one file, and by the travel time to they into
Row sequence.To obtain the driving trace of single taxi.It is noted that our research emphasis is in the five rings of Beijing
Vehicle Move Mode (With).Therefore, we eliminate
Track of vehicle data of the longitude and latitude not within the scope of five rings.
The ratio of taxi and private car on 1 Beijing's main traffic main line of table
S1.2: region division.We divide Pekinese region, and the trip of vehicle is reduced between different zones
Trip.Neighbouring lane segmentation method (ARS) proposed by the present invention is based on the trip track and POIs (geographic interest for utilizing people
Point), in conjunction with road network main traffic main line, carry out region division, the specific steps are as follows:
S1.2.1: using the function of identifying single region based on the method for topic model, one has multiple functions
Region contains the document of multiple themes just as one.The city to be studied is whole as a document, each region institute
The different functions of possessing are as theme.Meanwhile by interregional Move Mode as word, by POIs feature vector when composition
The metadata of shelves.In the present invention, using the DMR-based topic model in unsupervised learning, it is based on Latent
Dirichlet Allocation (LDA) and Dirichlet Multinomial Regression (DMR).By POIs feature to
Amount and Move Mode combine, and the function in region is comprehensively probed into terms of two.It is different in region for each region r
The number of the POIs of classification is can to count acquisition.Frequency density v of i-th kind of POIs in the r of regionI, rCalculation formula are as follows:
Wherein, NumiRepresent the quantity of the POIs of the i-th seed type in the r of region, SrRepresent the area of region r.In addition, for
For the r of region, its POI feature vector can be denoted as xr=(v1, r, v2, r..., vI, r..., vF, r, 1), represent the member of region r
Data, F are the species numbers of POI in r, and last vector 1 is default feature.After carrying out parameter Estimation using DMR model, for
For the r of region, theme distribution is K dimensional vector θr=(θR, 1, θR, 2..., θR, e..., θR, K), θR, eRepresent theme e in the r of region
Shared ratio.
S1.2.2: Classic Clustering Algorithms --- the k-means algorithm in unsupervised machine learning, the functional areas Lai Jinhang are used
Cluster.It is determined by experiment, for Beijing, when final cluster value is set as 9, the result of cluster is most ideal.In order to quantify
The range of pouplarity and region in functional areas, estimates the functional strength of each functional areas.The welcome journey of one functional areas
Degree has potential association with the volume of traffic, this, which represents mankind's Move Mode, means the functional strength in region.By rising for each travelling
Point and functional strength of the terminal point coordinate information input into Density Estimator KDE model, in quantization function region.Assuming that there is n
Region, by KDE model, using Density Estimator amount come the functional strength of zoning s:
Wherein, dI, rThe distance of region i to region r is represented, R represents bandwidth, and KF () represents gaussian kernel function.Carry out
After the estimation of complete functional strength, we annotate the region after dividing, the actual functional capability in the city Lai Fanying, and define its number
Attribute is Ka.Then, we carry out area marking from the aspect of four.Firstly, we according to the POIs feature in each region to
The average frequency density of amount is ranked up frequency of the type POIs in functional areas, and installation contains the POIs simultaneously
The frequency sizes of all functional areas be ranked up.Second, we count Move Mode most frequent in each functional areas.Third,
We utilize functional strength, probe into most representative POIs in each functional kernel, accordingly progress region annotation, and the 4th, I
According to the actual situation, carry out handmarking, such as scenic spots and historical sites.
S1.2.3: first by functional areas rasterizing, according to longitude and latitude 0.001 it is range in map projection, is divided into big
Measure grid.The relative length and width of each grid are defined as 1, there is a fixed ID.If comprising more in a grid
The grid is then grouped into the maximum functional area of occupied area by a different zones.For Beijing, proposition one based on neighbouring
The functional areas depth division methods of lane segmentation method.According to the information of vehicle flowrate on road of government statistics, 18 main cities are filtered out
City's road is numbered according to Arabic numerals 1-18.Every strategic road is marked in the map of rasterizing, and
It is regarded as line segment.Meanwhile we approximately regard each grid as node.Then, each node (grid) is calculated to line segment
The Euclidean distance of (road), and record the nearest road number Kl of each grid distance.It is each after distance calculating terminates
A grid node all has two attribute values, functional areas number Ka and nearest road number Kl where being respectively.To rasterizing
Each grid in figure is clustered, Ka and kl value node all the same is regarded a cluster.Finally, having obtained 153 thin
The functional area divided.It is same in our division methods since taxi is different with the ratio of private car between different roads
Region not only city function having the same, and the ratio of taxi private car is also identical.
The description of S1.3 road network.City road network map datum can exempt from from OpenStreetMap (OSM) or other open source websites
Take downloading.OSM data can be uploaded by any user, so everyone can safeguard and modify map datum.Ours
In research, we have downloaded Pekinese's OSM file, including road, underground, and the information of various building facilities reflects city
Geography information.But due to the characteristic of open source, there may be some mistakes between the data and actual conditions of downloading.In order to build
Vertical accurate simulation road network, we are corrected road topology structure, are allowed to match with real world.We use
Java OpenStreetMap (JOSM) technology modifies to it, this is that a free opening street map geography information is compiled
The tool of collecting.Further, since template of the invention is simulation and generates private car, rather than all vehicles, so I
Delete the redundancies such as railway, pavement.After deleting the related datas such as rds data, by map datum and real world into
Row is compared and is modified.
S2 is the embodiment of the generation layer of RDMP model in the present invention, the rail pre-processed using step 1) preceding layer
Mark data carry out the generation of private car track data collection.The present invention proposes population weight chance (RPWO) model based on region,
Using the vehicle flowrate in each region after division, to predict private car volume of traffic the beginning and the end dot matrix in two regions
(ODmatrix).Later, we use SUMO emulation tool, are the microcosmic of each private car by the matrix conversion between overall region
Trajectory path.
S2.1: the interregional volume of traffic calculates.The volume of traffic of vehicle refers to the vehicle for passing through certain section within a certain period of time
Quantity.The volume of traffic is able to reflect the overall magnitude of traffic flow of road, has important researching value.In the present invention, pass through calculating
The volume of traffic in each region, and interregional population weight opportunity model proposed by the present invention is utilized, it predicts and has divided region
Between the mobile volume of traffic of vehicle.Later according to the volume of traffic between estimation range, the public vehicles data set of simulation is generated.
Beijing is divided into different regions in step 1).The volume of traffic daily in each region is calculated now.Traffic in region
Amount by all roads in the region the volume of traffic it is cumulative form, therefore calculate at the beginning be single road the volume of traffic.Road
On the volume of traffic codetermined by public vehicles and floating vehicle, but the data that we obtain are only floating vehicle (taxi)
Track data.It would therefore be desirable to calculate the volume of traffic of public vehicles (private car) by the volume of traffic of floating vehicle.Often
The ratio of public vehicles and floating vehicle on road is different, according to this proportionate relationship, so that it may easily lead to
Cross the volume of traffic that the existing taxi volume of traffic calculates private car.It is mentioned according to " 2012 annual Traffic In Beijings develop annual report "
The information of confession, we have obtained the different proportion of mobile cart and public vehicles on main traffic road (shown in table 1).Because I
The region that divides be based on these main traffic routes, so we can assume that all roads in the same area have
Floating vehicle identical with main traffic road and public vehicles ratio.We are calculated by the following formula the society in each functional areas
Meeting vehicle vehicle fleet size:
Wherein, SAiFor the private car trip sum of region i；SGjRepresent the taxi in each grid j for including in the i of region
Vehicle quantity, NiIndicate the lattice number divided in the i of region, αiRepresent the ratio of private car and taxi in the i of region, SRwIt indicates
The quantity of taxi in the i of region in grid j in road w, njRepresent the number of road in grid j.It is obtained by above-mentioned two formula
The quantity of each functional areas private car.
S2.2: space-time interaction models.After the volume of traffic for obtaining each region in Beijing, the present invention continues to construct City-scale
Under the mankind go on a journey mode.Although the history for establishing mankind's mobility model is very long, researcher still lacks predicted city shifting
The accuracy method of dynamic model formula, especially in the case where data type is not diversified enough.The invention proposes interregional populations
Weight opportunity model (RPWO), to capture the potential driving factors of mankind's Move Mode on City-scale, the model independent of
Any adjustable parameter.It is noted that the result of study for this model shows that the model is very suitable to the people between city
Mouth flow pattern, but be not suitable between country, this demonstrate the diversity that the mankind under different spaces scale are flowed.With traditional people
Mouth weight opportunity model is different, in the present invention when considering human migrations model, it is contemplated that regional factor has carried out cluster-collection
Group's research.That is, the present invention passes through the technique study of region division migration models of the mankind between different zones.Tool
Steps are as follows for body:
S2.2.1: in RPWO model, if a land j to the attraction of starting area i, and with the weight of region j
The heart is the center of circle, with the center of gravity distance R of the center of gravity of starting area i and land jijFor the total population in the border circular areas of radius
Number QjiIt is inversely proportional；
S2.2.2: it is for center of gravityRegion, calculation method is as follows:
Wherein L indicates the net region number (0.001 longitude and latitude is as division) in the region.xlAnd ylIt respectively indicates in region
The opposite longitude and relative altitude of grid node.
S2.2.3: and then can be calculated to terminal to the attraction of starting point (i.e. according to the center of gravity in beginning and end region
Starting area is to the land volume of traffic):
Wherein, QjiIt indicates with RijFor radius, it is total as the private car in the border circular areas in the center of circle using the center of gravity of land j
Number；NsIt is the number in region included in circle, βrRepresent the ratio that region r is in border circular areas；The gross area of region r is
Sr, the area that region r is in border circular areas isPrIt is the private car number of region r, AjiRepresent land j relative to
The attraction of starting area i, ojThe private car number of land j is represented, M represents the private car total number in entire city；
S2.2.4: the attraction based on each region, the volume of traffic of the zoning i to region j:
Wherein, SAiTo go out number of lines from the private car of region i, it is calculated by S2.1, n is the region divided in city
Number；
S2.3: track emulation and the specific embodiment that data generate are as follows:
S2.3.1: the present invention first classifies to urban road according to the region divided.Road network file includes
The longitude and latitude of every road tie point, we calculate which region road belongs to using longitude and latitude, and write in road network file
Enter the road ID for including in each region.
S2.3.2: in S2.2, urban transportation amount O/D matrix has been obtained, has been counted in conjunction with modified road network file
It is emulated according to collection and generates work.In order to reach this target, we use SUMO tool.Firstly, we use in SUMO tool
OD2TRIPS plug-in unit imports O/D matrix and is broken down into single unit vehicle stroke.Volume of traffic O/D matrix can reflect in city
The trip mode of whole people.In addition, the individual vehicle routing information that matrix generates can reflect individual in terms of microcosmic angle
Motor pattern.In OD2TRIPS plug-in unit, we have input O/D matrix information, road network file, the road column for including in region
Table, and be arranged and generate period, the trip proportion of each period, generation type of vehicle parameter.Then, a series of vehicle rows are generated
The xml document of journey information, each travel information include vehicle ID, departure time, departure place ID and destination ID.
S2.3.3: with the help of OD2TRIPS, departure place and the destination information of each vehicle travel are generated, this is not
It is enough apparent.Therefore, we have used DUAROUTER plug-in unit in SUMO tool, it calculates SUMO using shortest path can
The vehicle route that can be used generates the path locus information of vehicle.The trip information that road network file and OD2TRIPS are generated is inputted,
Simulation time section, shortest path calculation method are set, ultimately produce vehicle driving trace information, including vehicle ID, the travel time,
Road passes through situation.Track of vehicle information can help us to analyze urban highway traffic, region trip mode.We generate
Daily track data collection size be about 3GB.
S2.3.4: other than track of vehicle information, the instantaneous velocity of unit period (per second) interior vehicle, opposite position
Set, longitude and latitude microscopic information it is also critically important for the data verification in the present invention.We use the Trace in SUMO tool
File Generation plug-in unit come be spaced at the appointed time it is interior it is per second generate vehicle information, including longitude and latitude position, traveling
Angle, instantaneous velocity, place road number.We input corresponding information in DUAROUTER function, while writing and matching accordingly
File is set, sets time interval to 15 minutes, and generates vehicle tracking file in different time periods.One day the type generated
The size of data set is about 200GB.
S3 is the embodiment of the validation layer of RDMP model in the present invention.Public vehicles track data is generated in step 2)
Afterwards, our design verification models, verifying generate the accuracy and authenticity of data.In validation layer, the present invention respectively from macroscopic view and
Microcosmic two aspect verifies data.
S3.1: in macromodel, the track data of generation is developed in annual report with Traffic In Beijing in 2012 and is described by we
Actual traffic situation carry out analysis comparison.
S3.1.1: magnitude of traffic flow comparison.Fig. 3 is Beijing main roads magnitude of traffic flow comparison diagram.Used truthful data
It is the Beijing Communication centre of research and development main roads traffic flow data announced in 2012.The result shows that either actual number
According to the data still generated, the whole day flow of West 4th Ring Road and East 4th Ring Road is all in forefront.The magnitude of traffic flow of West 5th Ring Road and South 2nd Ring Road
It is lower, illustrate that traffic burden is lighter.From the point of view of overall comparison result, in addition to South 5th Ring Road and East 5th Ring Road, data that the present invention generates
It coincide with real data preferable.
S3.1.2: Travel Range.Mobile for the mankind to study, travel time distribution and range distribution are two crucial ginsengs
Number.By studying the travel amount of different periods, researcher can propose prioritization scheme of preferably going on a journey, and alleviate condition of road surface,
Improve out line efficiency.In addition, the trip distance distribution of research people can also play important work in roading and trip prediction
With.Therefore, in the present invention, we analyze the distribution of hourage and distance using the track data collection of generation.By with reality
The comparison of border data has rated the accuracy of emulation data.Fig. 4 is that the travel amount of resident trip time is distributed.What participation was compared
Data include generating the official statistics of data, the first half of the year in 2012 and the second half year in 2012.The result shows that track number generated
According to real data trip characteristics having the same.In addition, from figure 7 it can be seen that 7:00-9:00 and 17:00-19:00 are two
A travel surge phase.The volume of traffic of the two periods accounts for about the 50% of daily total wheel traffic.In terms of resident trip range distribution,
Fig. 5 is our analysis result.We are by official's trip distance distribution of the data of generation and the first half of the year in 2012 and the second half year
Data are compared.From the point of view of the distribution of trip distance, trip number is inversely proportional with total distance.With the increase of distance, travelling
Number reduce.In terms of driving, the mankind prefer excursion, and the shortest distance is 0-5 kilometers, account for 40% or more.Consider
To concrete condition, when trip distance is too long, people can consider the factors such as oil consumption, time, select the trip modes such as train, subway
Rather than it drives.
S3.1.3: traffic condition.In terms of Macro-traffic Flow situation, vehicle totality operating range and running time are regular
's.In the present invention, we verify the data set of generation using the navigation Service of Baidu Map APIs.Today, with movement
Equipment using more and more, everyone has the experience using Map Services application program.In these softwares, we are inputted
The position of departure place and destination, and suitable trip mode is selected, such as walking, public transport, private car, the route estimated
Length and travel time.Historical data of these results based on real world is generated including the data that Department of Transportation provides, GPS
Track of vehicle data, and various applications and the electronic equipment of positioning can be sent.It is noted that the map of current mainstream
Service provider provides corresponding api interface for software developer, facilitates us such as Google Map, Baidu Map
Call directly service.Navigation Service can be used for verifying two important indicators in traffic flow:
Path length: the index represents the length of route.By using Baidu Map APIs, we are by the number of generation
It is routed in function according to concentrating the beginning and end coordinate gone on a journey every time to be input to.By navigation Service, corresponding road can be generated
Diameter and stroke length；
Travel time: the index represents the travel time.Similar with the method for path length is obtained, we use navigation clothes
It is engaged in estimate the duration of trip.From fig. 5, it can be seen that be more than 70% vehicle driving distance less than 10 kilometers.Therefore, I
Research emphasis is placed on to the traveling of vehicle within 10 kilometers.Fig. 6 is that the data generated by us and Baidu Maps APIs are pre-
The travel time of survey and the scatter plot of path length.Fig. 6 (a) is vehicle driving situation in 5 kilometer ranges, and Fig. 6 (b) is 5-10 public
In vehicle driving situation in range.It was noticed that within the scope of both path lengths, the present invention travel time generated
The assessment result provided with navigation Service is largely Chong Die.This also reflects the speed distribution of generation closer to very
The speed of real vehicle is distributed.In addition, being compared at a distance from 5~10km, short distance trip of the hourage data of generation within 5km
Closer to truthful data in row, this proves our model more suitable for generating the shorter trip track of distance.
S3.2: in micromodel, whether data set of consideration itself runs counter to reality.Acceleration analysis is carried out first,
Start with from calculating vehicle acceleration and its gradient, whether analysis data set is reliable.Then consistency analysis is carried out, vehicle is randomly selected
It is right, analyze the relative distance between two vehicles.A kind of method that the present invention devises quantitative detection track data.In this part
In, we select workaday morning peak time (7:00) as search time, especially 7:00-7:30.In addition, we are in total
As a comparison using four groups of observation data, the time interval of every group of data is 15 minutes.These track datas are respectively from 7:
00-7:15 raw data set, 7:15-7:30 raw data set, 7:00-7:15 generate data set, 7:15-7:30 generates data
Collection.
S3.2.1: acceleration analysis.Vehicle acceleration is the important component for studying dynamics of vehicle and traffic flow.Cause
This, it is necessary to the accuracy of the generated track data acceleration of verifying.If result precision is in the reasonable scope, prove to generate
Data set inherent reasonability with higher.Verifying accelerates data to have an apparent method, that is, checks it entire
Distribution situation in data set.In data acquisition and estimation procedure, two class problems, i.e., infeasible pole can be clearly found
The irregular shape of value and distribution.Fig. 7 is raw data set and the acceleration profile frequency for generating data set.It can from Fig. 7
Out, the either data that still generate of truthful data, the frequency distribution of vehicle acceleration is all normal distribution, it was demonstrated that stupid invention
The method of proposition has feasibility.
Other than the distribution of acceleration value, acceleration gradient is also used as the important indicator of the quality of data.Acceleration ladder
Spend Je[m/s3] indicate that acceleration changes with time, it is the derivative of acceleration.In the present invention, it is believed that in ± 3m/s3It is left
Right actual acceleration gradient value is the acceptable value in research.Therefore, it is proposed that three indexs carry out acceleration gradient
Analysis:
·JeGreater than ± 3m/s3The track data percentage of threshold value；
J maximum value in data set；
J minimum value in data set.
It analyzes obtained acceleration gradient error statistical result and is shown in Table 2.It can be seen that ± 3m/s3Except JeThe percentage of value
Than being respectively 2.99% (initial data (7:00-7:15)) and 3.28% (initial data (7:15-7:30)), and the data generated
Result it is similar (5.27% and 3.72%).This means that the result of the error statistics of j is fairly good.In lateral comparison, generally
Data set error illustrates that the data set is relatively accurate less than 10%.In addition, the minimum and maximum j value of selected data reaches
Relatively unreasonable value.
2 real trace of table and the acceleration gradient error statistics for generating track
S3.2.2: consistency analysis.Vehicle must keep reasonable distance with other vehicles in the process of moving.Otherwise,
Traffic accident is likely occurred.Such case is also be reflected in track of vehicle data set.The present invention comes from the distance between vehicle
Verifying generates the authenticity of data set.If there are the distance interval that many exceptions are small in data set, the data set it is accurate
Property is doubtful.
When we pay close attention to two vehicles close to each other, vehicle to the distance between can be used to quantitative estimation track number
According to error.In fact, at a time, the distance between a pair of of vehicle can directly be measured from the position of two vehicles at this time
Out.
In our study, it will be assumed that it is the length in the section being spaced between vehicle.Formal definition are as follows:
The vehicle spacing in T moment vehicle v1 and v2 is represented,WithMean
Road starting point end is arrived a little respectivelyWithRoad alignment length, also
It is the projection of the road alignment of actual vehicle position.
In order to simplify calculating process, it will be assumed that vehicle is on road always along straight-line travelling.Then directly from vehicle
The corresponding projection coordinate of reason coordinate calculates the distance between vehicle:
Under normal circumstances, whenAt least when dropping to 5m or less in a flash, two vehicles can collide, and cause
Traffic accident.Therefore, pass through detectionWhen ratio shared in total data set, can be with validation data set
Accuracy.If data are concentrated with a large amount of exceptional values, data set most possibly goes wrong.For consistency analysis, unite below
It counts meaningful:
This is the sum that each selected data concentrates vehicle pair；
The average traffic interval of vehicle pair in data set；
Vehicle number and ratio of the speed lower than 5 meters；
Vehicle is to largest interval.
The consistency statistical result of analysis is shown in Table 3.We provide that the two cars within 50m can be considered as one group of vehicle
It is right.According to calculated result, it can be seen that the sum of vehicle pair is about between 4000 to 5000.In all selected numbers
According to concentration, the raw data set in 7:00-7:15 and 7:15-7:30 that we use has 24 pairs and 18 pairs abnormal vehicles respectively
It is right, the 0.56% and 0.41% of the difference total vehicle pair of Zhan.For the data set that we generate, in 7:00-7:15 and 7:15-7:30
The two periods, abnormal vehicle account for vehicle to the 1.91% and 1.24% of sum to respectively 94 pairs and 50 pairs respectively.It can be with
Find out, abnormal data is maintained in lesser range in the data set of generation, it was demonstrated that the data of generation authenticity with higher.
In addition, average traffic is spaced about 27m in the data set of generation, it is not much different with actual mean value (about 22m).Fig. 8 reflection
The frequency of occurrences of the vehicle to gap length.Obviously, this four groups of observation data have closely similar distribution, it means that generate
Data set have vehicle Move Mode similar with truthful data, it was demonstrated that modelling effect proposed by the invention is good.
3 real trace of table and consistency (interval index) error statistics for generating track
It is specific embodiments of the present invention and the technical principle used described in above, if conception under this invention institute
The change of work when the spirit that generated function is still covered without departing from specification and attached drawing, should belong to of the invention
Protection scope.
Claims (2)
1. a kind of based on the urban transportation track data set creation method for hiring out car data and city road network, which is characterized in that institute
The urban transportation track data set creation method stated is completed in three layers of RDMP track of vehicle data set generation model, described
Three layers of RDMP track of vehicle data set generation model include that steps are as follows for preparation layer, generation layer and validation layer:
S1: in preparation layer, pre-processing original taxi track data, deletes useless and abnormal data；Make simultaneously
Urban area division is carried out with the city function limited region dividing method based on neighbouring lane segmentation method, and structure is carried out to city road network
It builds, includes the following steps:
S1.1: pretreatment: deleting the useless track data under complete vehicle curb condition, clears up as caused by equipment precision and statistical error
Absurd data obtain available taxi track data；
S1.2: the city of research is divided into not using mankind's motion track and POI point of interest based on neighbouring lane segmentation method
Same region, is converted to the circulation between different zones for the trip of all vehicles, detailed process is as follows:
S1.2.1: it is using the DMR-based probability topic model of unsupervised learning, city is whole as a document, each
The different functions that region is possessed are as theme, by interregional Move Mode as word, by POIs feature vector as
The metadata of document, frequency density v of i-th kind of POIs in the r of regionI, rCalculation formula are as follows:
Wherein, NumiRepresent the quantity of the POIs of the i-th seed type in the r of region, SrRepresent the area of region r, the POI feature of region r
Vector is denoted as xr=(v1, r, v2, r..., vI, r..., vF, r, 1), the metadata of region r is represented, wherein F is the type of POI in r
Number, last vector 1 is default feature；The theme distribution of region r is K dimensional vector θr=(θR, 1, θR, 2..., θR, e...,
θR, K), θR, eRepresent ratio shared by theme e in the r of region；
S1.2.2: using k-means clustering algorithm, theme distribution obtained in S1.2.1 clustered, and each travelling rises
Point and functional strength of the terminal point coordinate information input into Density Estimator KDE model, in quantization function region；Equipped with n area
Domain, by Density Estimator model KDE model, using Density Estimator amount come the functional strength of zoning r:
Wherein, dI, rThe distance of region i to region r is represented, R represents bandwidth, and KF () represents gaussian kernel function；Carrying out function
After the estimation of energy intensity, the region after dividing, the actual functional capability in the city Lai Fanying are annotated, and definition region r number attribute is
Ka；
S1.2.3: region clustering
Using neighbouring lane segmentation method, strategic road is considered as line segment, and grid regards node as, calculate each grid to road Euclidean
Distance, and record the nearest road number K of each grid distancel, to each grid in rasterizing map after calculating
It is clustered, by KaAnd KlIt is worth node all the same as a cluster；
S1.3: according to practical study region, the original document downloaded from the website Street Map Open is cut and is divided
Layer, meanwhile, the road in manual modification real world, which is gone on a journey, to be limited, and updates newest traffic condition, to the road network in city
Network is constructed；
S2: in generation layer:
S2.1: according to the staticaccelerator scale of floating vehicle and public vehicles in different function area, the society in each region is calculated
The quantity of vehicle, calculation formula are as follows:
Wherein, SAiFor the private car trip sum of region i；SGjRepresent the taxi number in each grid j for including in the i of region
Amount, NiIndicate the lattice number divided in the i of region, αiRepresent the ratio of private car and taxi in the i of region, SRwIndicate region
The quantity of taxi in i in grid j in road w, njRepresent the number of road in grid j；
S2.2: it proposes regional population's weight chance RPWO model, calculates the origin and destination OD matrix of public vehicles, including following step
It is rapid:
S2.2.1: setting a land j to the attraction of starting area i, with using the center of gravity of region j as the center of circle, with origin zone
The center of gravity distance R of the center of gravity of domain i and land jijFor the total number of people Q in the border circular areas of radiusjiIt is inversely proportional；
S2.2.2: it is for center of gravityRegion, calculation method is as follows:
Wherein, L indicates the grid number in the region；xlAnd ylRespectively indicate the opposite longitude and relative altitude of grid in region；
S2.2.3: according to the center of gravity in beginning and end region, the attraction A to land j to starting area i is calculatedji, meter
Calculation method is as follows:
Wherein, QjiIt indicates with RijFor radius, using the center of gravity of land j as the private car sum in the border circular areas in the center of circle；Ns
It is the number in region included in circle, βrRepresent the ratio that region r is in border circular areas；The gross area of region r is Sr, area
The area that domain r is in border circular areas isPrIt is the private car number of region r, AjiLand j is represented relative to origin zone
The attraction of domain i, ojThe private car number of land j is represented, M represents the private car total number in entire city；
S2.2.4: the attraction based on each region, the volume of traffic of the zoning i to region j:
Wherein, SAiTo go out number of lines from the private car of region i, it is calculated by S2.1, n is the number of regions divided in city；
S2.3: simulation track data are generated using SUMO emulation tool in conjunction with OD matrix, steps are as follows:
S2.3.1: classified according to the region divided to urban road；Road network file includes every road tie point
Longitude and latitude, calculate which region road belongs to using longitude and latitude, and be written in each region in road network file and include
Road ID；
S2.3.2: using the OD2TRIPS plug-in unit in SUMO tool, importing O/D matrix and be broken down into single unit vehicle stroke,
According to the concrete condition and data in city, the road list for including in the O/D matrix, road network file, region of S2.2 generation is inputted,
And be arranged and generate period, the trip proportion of each period, generation type of vehicle parameter, generate a series of vehicle travel information
Xml document, each travel information include vehicle ID, departure time, departure place ID and destination ID；
S2.3.3: the path locus information in order to generate private vehicle inputs road using DUAROUTER plug-in unit in SUMO tool
The trip information that online article part and OD2TRIPS are generated, is arranged simulation time section, shortest path calculation method, generates the vehicle of simulation
Path, including vehicle ID, the travel time, road by situation, ultimately produce vehicle driving trace information；
S2.3.4: to generate the information per second for generating vehicle in specified time interval, including longitude and latitude position, traveling angle,
Instantaneous velocity, place road number, using the Trace File Generation plug-in unit in SUMO tool, in DUAROUTER letter
Corresponding information is inputted in number, while writing corresponding configuration file, when setting △ t minutes for time interval, and generating different
Between section vehicle tracking file；
S3: in validation layer, the design verification model in terms of both macro and micro two respectively, verifying generates the accuracy of data and true
Reality, comprising the following steps:
S3.1: in macromodel, by the track data of generation, analysis comparison, including traffic flow are carried out with actual traffic situation
Amount comparison, Travel Range comparison, traffic condition comparison；
S3.2: in microcosmic verifying model, the authenticity of data is generated from acceleration and the angle analysis of relative distance and evaluation,
A kind of method for devising quantitative detection track data, steps are as follows:
S3.2.1: generated track data acceleration and its gradient value J are verified in acceleration analysise[m/s3] accuracy, work as knot
Fruit precision in the reasonable scope, is then proved the data set generated inherent reasonability with higher, is carried out using three indexs
Acceleration gradient analysis: JeGreater than ± 3m/s3J in the track data percentage of threshold value, data seteJ in maximum value, data seteMost
Small value；
S3.2.2: consistency analysis verifies the authenticity for generating data set from the distance between vehicle, when existing in data set
Abnormal small distance interval, then the accuracy of the data set is doubtful；
S3.2.3: vehicle to the distance between calculation formula it is as follows:
The vehicle spacing in T moment vehicle v1 and v2 is represented,WithMean that road rises
Point end is arrived a little respectivelyWithRoad alignment length, that is,
The projection of the road alignment of actual vehicle position；
S3.2.4: vehicle is set on road always along straight-line travelling, is directly calculated from the corresponding projection coordinate of vehicle geographical coordinate
The distance between vehicle out:
WhenAt least when dropping to 5m or less in a flash, two vehicles can collide, and lead to traffic accident, therefore, lead to
Cross detectionWhen ratio shared in total data set, carry out the accuracy of validation data set.
2. according to claim 1 a kind of based on the urban transportation track data collection generation for hiring out car data and city road network
Method, which is characterized in that the judgment principle of strategic road are as follows: be 0.001*0.001 according to longitude and latitude by city map rasterizing
Grid division goes out whole day average traffic 100,000 or more road extractions as strategic road.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910532080.8A CN110298500B (en) | 2019-06-19 | 2019-06-19 | Urban traffic track data set generation method based on taxi data and urban road network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910532080.8A CN110298500B (en) | 2019-06-19 | 2019-06-19 | Urban traffic track data set generation method based on taxi data and urban road network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN110298500A true CN110298500A (en) | 2019-10-01 |
| CN110298500B CN110298500B (en) | 2022-11-08 |
Family
ID=68028301
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910532080.8A Expired - Fee Related CN110298500B (en) | 2019-06-19 | 2019-06-19 | Urban traffic track data set generation method based on taxi data and urban road network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN110298500B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111104792A (en) * | 2019-12-13 | 2020-05-05 | 浙江工业大学 | Traffic track data semantic analysis and visualization method based on topic model |
| CN111126713A (en) * | 2019-12-31 | 2020-05-08 | 方正国际软件(北京)有限公司 | Space-time hot spot prediction method and device based on bayonet data and controller |
| CN111401663A (en) * | 2020-04-12 | 2020-07-10 | 广州通达汽车电气股份有限公司 | Method and device for updating public transport space-time OD matrix in real time |
| CN111583641A (en) * | 2020-04-30 | 2020-08-25 | 北京嘀嘀无限科技发展有限公司 | Road congestion analysis method, device, equipment and storage medium |
| CN111651502A (en) * | 2020-06-01 | 2020-09-11 | 中南大学 | City functional area identification method based on multi-subspace model |
| CN111966770A (en) * | 2020-07-21 | 2020-11-20 | 中国地质大学（武汉） | Urban street function identification method and system based on geographic semantic word embedding |
| CN112101132A (en) * | 2020-08-24 | 2020-12-18 | 西北工业大学 | Traffic condition prediction method based on graph embedding model and metric learning |
| CN112559909A (en) * | 2020-12-18 | 2021-03-26 | 浙江工业大学 | Business area discovery method based on GCN embedded spatial clustering model |
| CN112632208A (en) * | 2020-12-25 | 2021-04-09 | 际络科技(上海)有限公司 | Traffic flow trajectory deformation method and device |
| CN112906948A (en) * | 2021-02-02 | 2021-06-04 | 湖南大学 | Private car track big data-based urban area attraction prediction method, equipment and medium |
| CN113111514A (en) * | 2021-04-12 | 2021-07-13 | 清华大学深圳国际研究生院 | Vehicle microscopic driving scene simulation method and computer readable storage medium |
| CN113254565A (en) * | 2021-06-21 | 2021-08-13 | 浙江口碑网络技术有限公司 | Region identification method and device, computer equipment and computer readable storage medium |
| CN113298144A (en) * | 2021-05-24 | 2021-08-24 | 中南大学 | Urban three-generation space identification and situation analysis method based on multi-source data |
| CN113343781A (en) * | 2021-05-17 | 2021-09-03 | 武汉大学 | Urban functional area identification method comprehensively using remote sensing data and taxi track data |
| CN114724362A (en) * | 2022-03-23 | 2022-07-08 | 中交信息技术国家工程实验室有限公司 | Vehicle track data processing method |
| CN115995151A (en) * | 2023-03-06 | 2023-04-21 | 昆明符海姚网络科技有限公司 | Network vehicle-closing abnormal behavior detection method applied to city management |
Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006234569A (en) * | 2005-02-24 | 2006-09-07 | Matsushita Electric Ind Co Ltd | Generation method and device of traffic information, and reproduction method and device |
| US20130060586A1 (en) * | 2011-09-07 | 2013-03-07 | National Tsing Hua University | Dynamic Taxi-Sharing System and Sharing Method Thereof |
| CN103531024A (en) * | 2013-10-28 | 2014-01-22 | 武汉旭云科技有限公司 | Dynamic traffic network urban road feature model and modeling method thereof |
| CN105117595A (en) * | 2015-08-19 | 2015-12-02 | 大连理工大学 | Floating car data based private car travel data integration method |
| GB201711408D0 (en) * | 2016-12-30 | 2017-08-30 | Maxu Tech Inc | Early entry |
| CN107392245A (en) * | 2017-07-19 | 2017-11-24 | 南京信息工程大学 | A kind of taxi trajectory clustering algorithm Tr OPTICS |
| CN107609107A (en) * | 2017-09-13 | 2018-01-19 | 大连理工大学 | A kind of trip co-occurrence phenomenon visual analysis method based on multi-source Urban Data |
| GB201803663D0 (en) * | 2017-01-10 | 2018-04-25 | Beijing Didi Infinity Technology & Dev Co Ltd | No details |
| CN108681795A (en) * | 2018-05-23 | 2018-10-19 | 华南理工大学 | Electric automobile charging load space-time prediction method under constraint of urban traffic network and user travel chain |
| CN109166317A (en) * | 2018-10-29 | 2019-01-08 | 东北林业大学 | Method is determined by the time based on the urban transportation path of state feature |
| CN109272175A (en) * | 2018-11-15 | 2019-01-25 | 山东管理学院 | A kind of data collection system and method based on Urban Migrant network |
| CN109686091A (en) * | 2019-01-17 | 2019-04-26 | 中南大学 | A kind of magnitude of traffic flow based on multisource data fusion fills up algorithm |
-
2019
- 2019-06-19 CN CN201910532080.8A patent/CN110298500B/en not_active Expired - Fee Related
Patent Citations (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2006234569A (en) * | 2005-02-24 | 2006-09-07 | Matsushita Electric Ind Co Ltd | Generation method and device of traffic information, and reproduction method and device |
| US20130060586A1 (en) * | 2011-09-07 | 2013-03-07 | National Tsing Hua University | Dynamic Taxi-Sharing System and Sharing Method Thereof |
| CN103531024A (en) * | 2013-10-28 | 2014-01-22 | 武汉旭云科技有限公司 | Dynamic traffic network urban road feature model and modeling method thereof |
| CN105117595A (en) * | 2015-08-19 | 2015-12-02 | 大连理工大学 | Floating car data based private car travel data integration method |
| GB201711408D0 (en) * | 2016-12-30 | 2017-08-30 | Maxu Tech Inc | Early entry |
| GB201803663D0 (en) * | 2017-01-10 | 2018-04-25 | Beijing Didi Infinity Technology & Dev Co Ltd | No details |
| CN107392245A (en) * | 2017-07-19 | 2017-11-24 | 南京信息工程大学 | A kind of taxi trajectory clustering algorithm Tr OPTICS |
| CN107609107A (en) * | 2017-09-13 | 2018-01-19 | 大连理工大学 | A kind of trip co-occurrence phenomenon visual analysis method based on multi-source Urban Data |
| CN108681795A (en) * | 2018-05-23 | 2018-10-19 | 华南理工大学 | Electric automobile charging load space-time prediction method under constraint of urban traffic network and user travel chain |
| CN109166317A (en) * | 2018-10-29 | 2019-01-08 | 东北林业大学 | Method is determined by the time based on the urban transportation path of state feature |
| CN109272175A (en) * | 2018-11-15 | 2019-01-25 | 山东管理学院 | A kind of data collection system and method based on Urban Migrant network |
| CN109686091A (en) * | 2019-01-17 | 2019-04-26 | 中南大学 | A kind of magnitude of traffic flow based on multisource data fusion fills up algorithm |
Non-Patent Citations (3)
| Title |
|---|
| 李凤岐等: "TWIT:社交网络中局部信任值的双向计算", 《计算机工程与应用》 * |
| 王雷等: "基于出租车轨迹数据的交通异常识别算法", 《科学技术与工程》 * |
| 纪丽娜等: "基于城市交通大数据的车辆类别挖掘及应用分析", 《计算机应用》 * |
Cited By (25)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111104792A (en) * | 2019-12-13 | 2020-05-05 | 浙江工业大学 | Traffic track data semantic analysis and visualization method based on topic model |
| CN111104792B (en) * | 2019-12-13 | 2023-05-23 | 浙江工业大学 | Traffic track data semantic analysis and visualization method based on topic model |
| CN111126713A (en) * | 2019-12-31 | 2020-05-08 | 方正国际软件(北京)有限公司 | Space-time hot spot prediction method and device based on bayonet data and controller |
| CN111126713B (en) * | 2019-12-31 | 2023-05-09 | 方正国际软件(北京)有限公司 | Space-time hot spot prediction method and device based on bayonet data and controller |
| CN111401663A (en) * | 2020-04-12 | 2020-07-10 | 广州通达汽车电气股份有限公司 | Method and device for updating public transport space-time OD matrix in real time |
| CN111583641A (en) * | 2020-04-30 | 2020-08-25 | 北京嘀嘀无限科技发展有限公司 | Road congestion analysis method, device, equipment and storage medium |
| CN111651502A (en) * | 2020-06-01 | 2020-09-11 | 中南大学 | City functional area identification method based on multi-subspace model |
| CN111651502B (en) * | 2020-06-01 | 2021-09-14 | 中南大学 | City functional area identification method based on multi-subspace model |
| CN111966770A (en) * | 2020-07-21 | 2020-11-20 | 中国地质大学（武汉） | Urban street function identification method and system based on geographic semantic word embedding |
| CN112101132A (en) * | 2020-08-24 | 2020-12-18 | 西北工业大学 | Traffic condition prediction method based on graph embedding model and metric learning |
| CN112559909A (en) * | 2020-12-18 | 2021-03-26 | 浙江工业大学 | Business area discovery method based on GCN embedded spatial clustering model |
| CN112632208A (en) * | 2020-12-25 | 2021-04-09 | 际络科技(上海)有限公司 | Traffic flow trajectory deformation method and device |
| CN112906948A (en) * | 2021-02-02 | 2021-06-04 | 湖南大学 | Private car track big data-based urban area attraction prediction method, equipment and medium |
| CN112906948B (en) * | 2021-02-02 | 2023-12-22 | 湖南大学 | Urban area attraction prediction method, device and medium based on private car track big data |
| CN113111514A (en) * | 2021-04-12 | 2021-07-13 | 清华大学深圳国际研究生院 | Vehicle microscopic driving scene simulation method and computer readable storage medium |
| CN113111514B (en) * | 2021-04-12 | 2022-05-27 | 清华大学深圳国际研究生院 | Vehicle microscopic driving scene simulation method and computer readable storage medium |
| CN113343781A (en) * | 2021-05-17 | 2021-09-03 | 武汉大学 | Urban functional area identification method comprehensively using remote sensing data and taxi track data |
| CN113343781B (en) * | 2021-05-17 | 2022-02-01 | 武汉大学 | City functional area identification method using remote sensing data and taxi track data |
| CN113298144A (en) * | 2021-05-24 | 2021-08-24 | 中南大学 | Urban three-generation space identification and situation analysis method based on multi-source data |
| CN113254565B (en) * | 2021-06-21 | 2021-09-28 | 浙江口碑网络技术有限公司 | Region identification method and device, computer equipment and computer readable storage medium |
| CN113254565A (en) * | 2021-06-21 | 2021-08-13 | 浙江口碑网络技术有限公司 | Region identification method and device, computer equipment and computer readable storage medium |
| CN114724362B (en) * | 2022-03-23 | 2022-12-27 | 中交信息技术国家工程实验室有限公司 | Vehicle track data processing method |
| CN114724362A (en) * | 2022-03-23 | 2022-07-08 | 中交信息技术国家工程实验室有限公司 | Vehicle track data processing method |
| CN115995151A (en) * | 2023-03-06 | 2023-04-21 | 昆明符海姚网络科技有限公司 | Network vehicle-closing abnormal behavior detection method applied to city management |
| CN115995151B (en) * | 2023-03-06 | 2023-12-22 | 北京白龙马云行科技有限公司 | Network vehicle-closing abnormal behavior detection method applied to city management |
Also Published As
| Publication number | Publication date |
|---|---|
| CN110298500B (en) | 2022-11-08 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN110298500A (en) | A kind of urban transportation track data set creation method based on taxi car data and city road network | |
| Kong et al. | RMGen: A tri-layer vehicular trajectory data generation model exploring urban region division and mobility pattern | |
| CN103278168B (en) | A kind of paths planning method evaded towards traffic hot spot | |
| US10373494B1 (en) | Method and apparatus for estimating a parking event based on device signal observations | |
| CN110555992B (en) | Taxi driving path information extraction method based on GPS track data | |
| CN110796337B (en) | System for evaluating service accessibility of urban bus stop | |
| CN112036757B (en) | Mobile phone signaling and floating car data-based parking transfer parking lot site selection method | |
| CN104282142B (en) | Bus station arrangement method based on taxi GPS data | |
| CN109520499B (en) | Method for realizing regional real-time isochrones based on vehicle GPS track data | |
| Schneider | Measuring transportation at a human scale: An intercept survey approach to capture pedestrian activity | |
| CN114139251B (en) | Integral layout method for land ports of border regions | |
| CN110413855A (en) | A kind of region entrance Dynamic Extraction method based on taxi drop-off point | |
| Chen et al. | Rail transit ridership: station-area analysis of Boston’s Massachusetts Bay transportation authority | |
| Guo et al. | Mass rapid transit ridership forecast based on direct ridership models: A case study in Wuhan, China | |
| CN109489679A (en) | A kind of arrival time calculation method in guidance path | |
| Dong et al. | Quantitative assessment of urban road network hierarchy planning | |
| Steenberghen et al. | Support study on data collection and analysis of active modes use and infrastructure in Europe | |
| CN112183871B (en) | Urban traffic guidance system based on air index | |
| Zilske et al. | Building a minimal traffic model from mobile phone data | |
| Sun et al. | Measuring the influence of built environment on walking behavior: An accessibility approach | |
| Tomala et al. | TRAVEL TIME MAP-THE CASE STUDY OF WARSAW SUBWAY | |
| Kong et al. | Research on OD Estimation of Public Transit Passenger Flow Based on Multi-source Data | |
| Townsend | Spatial Measurement of Transit Service Frequency in Canada | |
| Hameed et al. | Evaluation the short urban road network by GIS: Case study in Baqubah city | |
| CN114973668B (en) | Urban road traffic weak link identification method based on topological step number analysis |
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 | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20221108 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |