An All-Time-Domain Moving Object Data Model, Location ...faculty.ung.edu/jyang/research_files/papers/wuqy.pdf · To estimate the posi-tion of moving objects at 𝑡time, trajectory

Research ArticleAn All-Time-Domain Moving Object Data Model LocationUpdating Strategy and Position Estimation

Qunyong Wu1 Junyi Huang1 Jianping Luo1 and Jianjun Yang2

1Key Lab of Spatial Data Mining and Information Sharing of MOE Fuzhou University Fuzhou 350002 China2Department of Computer Science University of North Georgia Oakwood GA 30566 USA

Correspondence should be addressed to Qunyong Wu qunyongwuqqcom

Received 20 October 2014 Revised 20 January 2015 Accepted 12 February 2015

Academic Editor Lin Cai

Copyright copy 2015 Qunyong Wu et al This is an open access article distributed under the Creative Commons Attribution Licensewhich permits unrestricted use distribution and reproduction in any medium provided the original work is properly cited

To solve the problems from the existing moving objects data models such as modeling spatiotemporal object continuous actionmultidimensional representation and querying sophisticated spatiotemporal position we firstly established an object-oriented all-time-domain data model for moving objects The model added dynamic attributes into object-oriented model which supportedall-time-domain data storage and query Secondly we proposed a new dynamic threshold location updating strategy The locationupdating threshold was given dynamically in accordance with the velocity accuracy and azimuth positioning information from theGPSThirdly we presented several different position estimationmethods to estimate the historical location and future locationThecubic Hermite interpolation function is used to estimate the historical location Linear extended positioningmethod velocitymeanvalue positioning method and cubic exponential smoothing positioning method were designed to estimate the future locationWefurther implemented the model by abstracting the data types of moving object which was established by PLSQL and extendedOracle Spatial Furthermore the model was tested through the different moving objectsThe experimental results illustrate that thelocation updating frequency can be effectively reduced and thus the position information transmission flow and the data storagewere reduced without affecting the moving objects trajectory precision

1 Introduction

Moving objects database [1] (MOD) is a database whichmanages position-related information of moving objectsRecently the research of moving objects database focusedon moving object representation modeling indexing queryuncertainty dealing and privacy protection [2] The datamodel is the basis for the moving object database Theearly study centers on the spatiotemporal data modelZhao [3] formally defined the types and operations offereddetailed insight into the considerations that went into thedesign exemplified the use of the abstract data types usingSQL and offered a precise and conceptually clean foundationfor implementing a spatiotemporal DBMS extension Aspatiotemporal cubemodel [4] and a simplest time slice snap-shot model [5] were presented and used However the modelcannot fit with the description of spatial changes accom-panying the temporal changes Worboys et al proposed

an object-oriented spatiotemporal data model [6] which wasfurther studied by many other scholars [7ndash10] Tryfona pro-posed spatiotemporal entity relation (STER) data model [1112] which used an extended entity relationship to describe thephenomenon in real world After that many scholars startedto work on the spatiotemporal database model to adapt to theposition management of moving objects Wolfson presentedthe moving object spatial-temporal (MOST) model [13 14]which showed the spatial difference with time by a dynamicattribute However the MOST model failed to describe thewhole spatiotemporal trajectory and only supported currentand future state query of moving objects Jin et al [15] built aspatiotemporal object relationmodel based on object relationdata model and the key was to extend classification systemand operation objectives relation data model by abstractdata types (ADT) Xue et al [16] proposed process-orientedspatiotemporal data model in which studied expressionorganization and storage of continuous changing geographic

Hindawi Publishing CorporationInternational Journal of Distributed Sensor NetworksArticle ID 463749

2 International Journal of Distributed Sensor Networks

entities were given Ding carried out the research on theDynamic Transportation Network Based Moving ObjectsDatabase (DTNMOD) and its real-time dynamic transportnetwork flow analysis method based on the spatiotemporaltrajectory of moving objects [17 18] Yang et al developeda protocol based on prediction that significantly boosts thespeed of message transmission over vehicles [19 20] Fromthe above analysis and comparisons it is indicated thatobject-oriented model and moving object model workedbetter in spatiotemporal object modeling and have theadvantages in spatial query time query and spatiotemporalquery But in the continuous moving modeling the datatransmission and data storage also have some disadvantages

The traditional location update method was time pointselection method including equal-time algorithm equal-distance algorithm and dynamic time point algorithm [2122] Equal-time algorithm was the most simple and wasmore effective to achieve and the communication flow isstable but the adaptability was poorWhen the moving objectis static it also records the position which causes invalidstorage Equal-distance algorithm has dramatically increasedcommunication flow in high-speed motion dynamic timepoint algorithm is capable of reducing the position updatebut this method is only suitable for high positioning accuracyand stably changing positioning parameters

According to the deficiencies of continuousmovingmod-eling multidimensional expression and complex spatiotem-poral queries an object-oriented moving object databasemodel supporting all-time-domain query was proposed bycombining static object model and moving object model Inorder to reduce the amount of moving object data trans-mission and storage a dynamic position updating strategywas studied It is suitable for frequent changes of positionof moving objects This model solved the problem of con-tinuous storage multidimensional expression and complexspatiotemporal queries This paper realized the model by aninstance at last

2 Moving Object Modeling

21 Moving Object Trajectory Segment Model The motionand change of moving object are continuous In the tradi-tionalmoving object data storagemodel the object-relationaldatabase was used to store the information continuouslyobtained from the moving object at a certain time intervalIt not only increases the amount of the data storage butalso increases the difficulties in the history and future spa-tiotemporal queries of themoving objects and in the complexspatiotemporal behavior queries In order to solve theseproblems dynamic properties methods of the MOST modelare proposed and implemented in this paper The discretemodeling ideas are proposed to support dynamic propertiesof the entire trajectory model and all the trajectories ofmoving objects are separated into many small segments asshown in Figure 1

In Figure 1 119905 axis means the time (119909 119910)means the pointinformation of moving objects and 119901119905

119894is the coordinate

information at 119905119894moment 119898119901119905

119894stands for the position

information of moving objects at 119905119894moment Each discrete

x

ytt1 t2 t3 t4 t5

pt4

pt2

pt3

pt1

pt5

mpt2

mpt1

mpt3

mpt4

mpt5

Figure 1 Moving trajectory decomposition of moving object

segment is abstracted as dynamic changes within a certainperiod of time 119875

In order to support the historical trajectories of movingobjects modeling the quadruple (119860 119875 119865 119878) was put for-ward to support the all-time-domain query All trajectoriessegments expressed by the quadruple constitute continuoustrajectories of moving objects

119860 = (119881119872119875119879119894119872119875119879

119894+1)

= (119881 (119901119905119894 V119894 119887119894 119886119894) (119901119905

119894+1 V119894+1 119887119894+1 119886119894+1))

119875 = [119905119894 119905119894+1)

119865 = 119891 (119905)

119878 = 1 2 3

(1)

Here 119860 stands for basic information of moving objectssuch as moving object ID position direction speed accu-racy and other information 119881 indicates the basic attributeinformation of moving objects such as moving object IDname and type 119872119875119879

119894is the position information of the

moving object in the moment 119905119894119872119875119879

119894= (119901119905

119894 V119894 119887119894 119886119894) 119901119905119894

is the coordinate information at 119905119894moment V

119894is the velocity

at 119905119894 119887119894is the azimuth at 119905

119894 and 119886

119894is the accuracy at 119905

119894 In this

paper the dynamic properties of the model are built on thelocation (119872119875119879

119894) change over time 119875

119875 indicates the time the object moves 119875 is not a fixedtime interval but a right open interval which is uncertainThesize of 119875 is related to the updating strategy of moving objectlocation

119865 represents motion function of moving objects 119891(119905)is the motion function of moving objects and is a time-dependent function by which we can estimate the positionof a moving object in the past at present and in the futurewithin a short time Motion function can be divided intomany types according to the actual situation such as linearfunction and curvilinear function

119878 is the state of the moving objects in the period 119875 Thereare mainly three types of the moving object state 119878 pastcurrent and network failure or interruption respectivelyrepresented by 1 2 and 3 When 119878 = 1 the position is usedto store and query the historical trajectory when 119878 = 2 thatis used to store current position and estimate future positionwhen 119878 = 3 the location information will not be stored untilthe terminal receives the location information again Whenthe update data is firstly or lastly received a new moving

International Journal of Distributed Sensor Networks 3

MID MovingPT1 MovingPT2 Period State

PT

Speed

Accurate

Azimuth

Longitude

Latitude

SRID

PeriodMPT StateHistory

Now

None

PT

ST Operation

perationperation

T Operation

S Instant

E Instant

MO Entity

S OS O

Figure 2 The entity model of moving object

object trajectory segment will be created the state of motionis represented by 1 and at the same time the front trackingsegment state is modified The former tracking segment stateis judged according to the time interval 119905 between the currenttime 119905 when it receives data and the previous time 119905

119894+1when

segment 119875 ends When 119905 lt Δ1199051015840 the last trajectory segment

state is changed to 2 and when 119905 ge Δ1199051015840 the last trajectory

fragment state is set to 3

22 Entity Model of Moving Object Trajectory Segment Thetrajectory segment of moving objects is represented as anobject entity The object entity of the trajectory segmentincludes the following moving object identification (MID)the starting point (MovingPT1) the ending point (Mov-ingPT2) the time interval (Period) and the moving state(State) of the trajectory segment as shown in Figure 2 Inthe model the moving trajectory point (MPT) is an abstractentity object including location information (PT) velocityazimuth and accuracy and the location information isalso an abstract spatial object including latitude longitudecoordinates and projection information

The abstraction entity of moving object trajectory seg-ment is MO Entity which inherits the time abstract entity(Period) and themoving point object (MPT) and also extendsthe spatial and temporal relation operation (ST Operation)the MID is a unique identification of the moving objectwhich can be used to indicate the basic information ofthe moving object The moving objects have three movingstates in period 119875 including history now and positioning orcommunication failure

3 Dynamic Threshold Location Updating

31 Position-Based Dynamic Threshold Location UpdatingStrategy There are two kinds of traditional location updatingmethod One is equal-time algorithm and the other is equal-distance algorithm Equal-time algorithm refers to uploadupdate positioning data according to a certain time intervalIn this algorithm the communication flow is stable butthe adaptability is poor Equal-distance algorithm refers toupload update positioning data according to a certain dis-tance intervalWhen themoving object is in high-speed statethe location update frequency and the communication trafficwill be greatly increased

Table 1 Distance Threshold Varying with Accuracy

Accuracy (119886) Distance threshold (Δ119889)[0 10) 10[10 20) 20[20 50) 50[50 100) 100[100 200) 200[200 500) 500[500 1000) 1000[1000infin) 1500

In practical applications the position information suchas velocity position accuracy and azimuth received fromGPS was in constant changing status and the changes arealso irregular Velocity is the main factor to the frequencyof location updating which is the principle of the dynamictime point algorithm [19] When the velocity from twoadjacent points changes slightly the mid-time positions canbe estimated by two adjacent points Thus the positionpoint will not be updated Different positioning accuracydetermines the position changing distance for exampledistance position algorithm if position changing distance119889 gtΔ119889 this point should belong to the location updating pointbut in fact this point may be an error point So we designeda dynamic distance threshold changed by the positioningaccuracy Only when the position change of distance wasgreater than the distance threshold the location point wasupdated The corresponding relationship between accuracyand distance threshold was shown in Table 1 The changingof azimuth mostly affects the trajectory precision of movingobjects shown as in Figure 3 Moving object moves from roadA to road B and then to C and finally enters the D if wedo not consider the impact of orientation on the locationupdating the trajectory of moving object stored in databaseis 1199010 rarr 1199011 rarr 1199012 rarr 1199013 rarr 1199014 while the real trajectorywas 1198910 rarr 1198911 rarr 1198912 rarr 1198913 rarr 1198914 rarr 1198915 rarr 1198916 Inconsideration of the azimuth location update the trajectoryof moving objects stored in the model database was morepractical


p0

p1

p2

p3

p4

f0f1

A

B

D

C

f2

f3

f4

f5

f6

Figure 3 Position updating affected by azimuth

According to the influencing factors of the locationupdate of moving object a new dynamic threshold algo-rithm was designed based on the moving object positioninginformation The algorithm combined with GPS positioninginformation (velocity accuracy and azimuth) to dynamicallydetermine the location update threshold Here is the defini-tion of the location update strategy

Hypothesizing that1198981199011199051is the last position change point

of moving object 1198981199011199052is the location information point of

moving objects lastly acquired the spatial distance between119898119901119905119894and 119898119901119905

2is 119889 azimuth changing is 120572 if 119889 gt Δ119889 and

120572 = 1198872minus 1198871gt Δ120579 and V gt 0 119889 gt 119889max that position is

believed to be with the update conditionHere Δ119889 is the minimum distance update threshold

under the different accuracy and 119889max is the minimum dis-tance update threshold when the distance changes more than119889max no matter whether other conditions satisfied all theupdate location Δ120579 is azimuth threshold and V stands forvelocity V = 0means that the position does not change

32 Design of Location Updating Strategy Flow Accordingto the definition of location update strategy the positionupdating process can be constructed as follows (shown inFigure 4)

(1) real-time acquisition of GPS data packet(2) data packet analysis and instantiation as one position

object location(3) judgment according to the dynamic precision thresh-

old rule on whether the distance change is largerthan the minimum change threshold in the currentprecision condition if not do not update the positionand jump to (1)

(4) judgment on whether the time between the currentposition and last updated position is greater than themaximum time threshold if so that means it meetsthe location update condition then go to (8)

GPS data package

Data analysis

Position data

Whether distance is satisfied under the accuracy

Azimuth threshold

Accuracy threshold rule

Start

End

Position updating

Yes

NoYes

No

No

Yes

No

Yes

t minus t0 gt tmax

Velocity = 0

Figure 4 Position updating flow of moving object

(5) judgment on whether the current positioning pointrsquosvelocity is equal to zero if so do not update locationand jump to (1)

(6) judgment on whether azimuth changing meets theazimuth threshold if not satisfied do not update theposition and jump to (1)

(7) judgment on whether the distance is greater thanthe minimum changing threshold when the aboveconditions are met

(8) meet the conditions of location update upload theposition information to the server

Themoving terminal gets the position point data throughthe GPS module and then the terminal program parses theposition information and judges whether the location infor-mation meets the moving updating strategy requirements Ifthe location information meets the requirements it will besent to the server and update the position point

33 Design of Database Location Store Procedure The data-base server receives mobile terminal location data and judgeswhether the object has movement trajectory in the databaseIf the object trajectory is in the database then find the currentpoint record change the current point as the history pointinsert the new location point and take the position as thecurrent position The definition steps are as follows


(1) Inquire about the location data that client receivedand parse it then process themodel ofmoving objectsdatabase

(2) Query whether the moving object MID is in thehistorical trajectory database if not go to step (7) elsego to step (3)

(3) Query the current location segment of the movingobject denoted as MO1

(4) Calculate the time interval between the upload loca-tion time and the last update time if the time intervalis greater than the threshold of time then go to step(6) else go to (5)

(5) Change the state of MO1 for the history state updatethe end time of the MO1 go to step (7)

(6) Change the state of MO1 to 3 end the updating flow(7) Store the location for the new database records

4 All-Time-Domain Position Estimation

Generally speaking the trajectories of moving objects arecontinuous However the location information is discreteand discontinuous recorded by moving object DatabaseWhen querying and visualizing moving object position andtrajectory at any moment the moving object position needsto be estimated according to the location information storedin the database In order to simplify the study the movingobject trajectory will be defined as a straight line or curve inthe time period 119875 by the proposed dynamic location updatestrategy

41 Estimation of Historical Position The trajectories ofmoving objects are divided into different moving trajectorysegments whose trajectory curve is represented by a dynamicfunction Thus in order to query the moving objects histor-ical trajectory at a certain point in time we need to estimatethe trajectory by the motion function Motion functions areselected under different circumstances and generally there aretwo kinds of motion function one is a linear function andthe other is the curve function A linear function is applied tothe object moving in the fixed road network such as vehiclewhile the curve function is applied to the vehicles movingfreely such as pedestrians In this paper we designed twokinds of estimation function One is the linear path functionand the other is three times Hermite interpolation function

411 Linear Path Function Method To estimate the posi-tion of moving objects at 119905 time trajectory segmentmust be found at the moment of 119879 which includes thestarting point 119898119901119905

1(119901119905(1199091 1199101) V1 119886 119887) and ending point

1198981199011199052(119901119905(1199092 1199102) V2 119886 119887) in the period 119875

119891 (119905) = 1199041 + V sdot Δ119905

= (1199091+V1+ V2

2sdot Δ119905 sin120572 119910

1+V1+ V2

2sdot Δ119905 cos120572)

(2)

t

v

y

x

yx

S(tj)G(tj)

f(tj)

998400x

t998400

x998400

y998400

998400y

Figure 5 Moving trajectory projection decomposition

where 1199041is the starting point V is the average velocity

measured from the starting point through ending pointΔ119905 = 119905 minus 119905

1 1199091is the longitude 119910

1is the latitude and 120572 =

|1198872minus 1198871| is the azimuth angle When using linear functions

to represent the trajectories of moving objects the movingobject trajectory in the period 119875 will be seen as a linearfunction that passes updated location point

412 Three Hermite Interpolation Function Method Thecubic spline function interpolation method was used to esti-mate the curve movement trajectories of the moving objectThe velocity change function 119878V(119905) and the azimuth changefunction 119878

120579(119905)with time were calculatedThen the location of

the moving object is estimated by position function loc(119905) =loc(1199050) + V(119905 minus 119905

0) Here V stands for the velocity at time

119905 but in actual movement the velocity from 1199050to 119905 is

constantly changing and the velocity at time 119905 cannot be usedto represent the middle time velocity so does the azimuthObviously this method has a great error

To solve this problem the three Hermite interpolationfunctionwas establishedThe calculating progress is designedas follows Firstly calculate a three-time spline curve 119891(119905

119895)

in three-dimensional space and decompose it into 119909119905 and 119910119905plane projection to get two three- time spline curves 119878(119905

119895) and

119866(119905119895) as shown in Figure 5 Consider 119878(119905

119895) = 119910119895 119866(119905119895) = 119909119895

Then use 119878(119905119895) and 119866(119905

119895) to estimate the history position at

time 119905Let 119878(119905

119895) = 119910

119895(119895 = 0 1 2 3) be the three-time spine

function of time series of 1199050lt 1199051lt 1199052lt 1199053 1199050 1199051 1199052 1199053

representing the starting time 119875119894minus3 119875119894minus2 119875119894minus1 119875119894 respectively

then

119878 (119905119895) =

3

sum119895=0

[119910119895120572119895 (119905) + 119898119895120573119895 (119905)] (3)

Here119898119895can be solved by cubic Hermite spline boundary

conditions and also by a first derivative from the curve oftwo-end point 120572

119895(119905) and 120573

119895(119905) mean the time function and

can be solved by the Lagrange fundamental polynomials


(interpolation basis function) as shown in the followingformula

120572119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895minus1minus 119905119895

) 119905119895minus1

le 119905 le 119905119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895+1minus 119905119895

) 119905119895le 119905 le 119905

119895+1(119895 = 0 1 2)

120573119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (119905 minus 119905119895) 119905

119895minus1le 119905 le 119905

119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (119905 minus 119905119895) 119905

119895le 119905 le 119905

119895+1(119895 = 0 1 2)

(4)

119866(119905119895) can be solved in the same way and the position 119891(119905)

at any time 119905 can be calculated by 119878(119905119895) 119866(119905119895)

42 Future Position Estimation The moving object datamodel can only store the current and historical locationupdating data The future location prediction of movingobjects in short needs to depend on the correspondingalgorithm Three methods for estimating future location aregiven as follows

421 The Linear Extended Positioning Method Linearextended positioning method means that the moving veloc-ity and direction at 119905+Δ119905moment are consistent with those atupdating moment 119905 In accordance with the current locationupdate points as the future prediction let 119905 be the motionstate updated point of current trajectories and let |V

119905| and 120579

119905

be moving velocity and azimuth at the same time then1003816100381610038161003816V119905+Δ119905

1003816100381610038161003816 =1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(5)

422 Moving Average Method Moving Average Methodmeans that themoving velocity at 119905+Δ119905moment is the averagevalue of velocity at last 119898 updating times and azimuth isconsistent with that at updating moment 119905 Consider

1003816100381610038161003816V119905+Δ1199051003816100381610038161003816 =

1

119898

119898

sum119894=1

1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(6)

Here 119898 is a positive integer and 119898 ge 2 Generallyspeaking the large value of119898 is meaningless Thus the valueof119898 is 2 or 3

423 Cubic Exponential Smoothing Method Generallyspeaking the moving velocity changes are nonlinear withtime The cubic exponential smoothing method can be usedto predict the change trend of velocity with time seriesExponential smoothing method is an iterative process andcan be denoted as 119878(119899)

119905 119896 adjacent location updating points

before current 119905 are usually selected to join the calculationTime sequence is denoted in turn as 119905

1 1199052 1199053 119905

119896 Besides

an initial value 119878(1)0

is needed After multiple period ofsmoothing the effect of 119878(1)

0becomes relatively small so the 119905

4

moment is regarded as the first time point in time sequenceThis time point value is the average of the last 3 time pointsmentioned before that is 119878(1)

0= (13)(V

1+ V2+ V3)

Let 119878(1)119905 119878(2)

119905 119878(3)

119905be the first second and third exponential

smoothing value of velocity V at 119886119905 119887119905 and 119888

119905which stand for

the smoothing coefficients of time series whose values rangeis [0 1] then

119878(1)

119905= 120572V119905+ (1 minus 120572) 119878

(1)

119905minus1

119878(2)

119905= 120572119878(1)

119905+ (1 minus 120572) 119878

(2)

119905minus1

119878(3)

119905= 120572119878(2)

119905+ (1 minus 120572) 119878

(3)

119905minus1

119886119905= 3119878(1)

119905minus 3119878(2)

119905+ 119878(3)

119905

119887119905=

120572

2 (1 minus 119886)2

sdot [(6 minus 5120572) 119878(1)

119905minus 2 (5 minus 4120572) 119878

(2)

119905+ (4 minus 3120572) 119878

(3)

119905]

119888119905=

1205722

2 (1 minus 119886)2[119878(1)

119905minus 2119878(2)

119905+ 119878(3)

119905]

(7)

The above formula can help work out 119886119905 119887119905 and 119888

119905and

then the cubic exponential smoothing speed predictionmodel is described as follows

V119905+Δ119905

= 119886119905+ 119887119905Δ119905 + 119888

119905Δ119905

120579119905+Δ119905

= 120579119905

(8)

The above three methods can be used to estimate thefuture position but in the actual object movement thepositions of moving objects are often changeable whichleads to the errors of predicted values To reduce the errorof prediction the position predicted time cannot exceed acertain range

5 Model Implementation

51 Moving Object Abstract Data Types The moving objectdatabasewas implemented based onOracle Spatial extensionThus the moving object abstract data type includes somebasic data type spatial data types The basic data type is thebasis for all databases including Int Number Varchar andDate Besides based on these types the PLSQL was used todefine the temporal data type supporting temporal data and


Instant

Period

Periods

MPT MO

Temporal Spatial ADT

Spatial-temporal ADT

NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle

middot middot middot

I Point I Line I Polygon I Rectangle I Circle

TAB MO

Figure 6 Abstract data type structure

the spatiotemporal data type supporting spatiotemporal dataThe abstract data type structure for moving object is shownin Figure 6

Temporal data type includes the instant (time) period(time interval) and periods (time interval set) Instant iscomposed of real data and time interval is made of thestarting and ending points of period (sins eins) of instant

Spatial data types are established based on theOracle Spa-tial SDO Geometry data type which includes five attributesSDO GTYPE SDO SRID SDO POINT SDO ELEM INFOand SDO ORDINATES However all forms of spatial datatypes are defined based on SDO Geometry data type andthe interface parameters are relatively complex In order toprovide simple spatial data type interface that is easy to assesswe packaged the SDO Geometry data type into S PointS Line S Polygon S Rectangle and S Circle respectivelySpatial data must have a projection or coordinate systemthe model uses the default projection 8307 namely WorldGeodetic System 1984 (WGS 84) to establish spatiotemporaldata types based on temporal data types and spatial datatypes

Spatiotemporal data types are the inheritance of temporaldata types and spatial data types and spatiotemporal datatypes include I Point I Line I Polygon I Rectangle andI Circle

52 Moving Object Data Model Implementation We usedPLSQL to define the abstract data types of moving objectsvarious types were constructed according to the object-oriented method The base class is spatial data types andtemporal data type classes The class diagram data typedefinitions of all moving objects are shown in Figure 7

521 Implementation of Temporal Data Type Classes Tem-poral data type mainly defines the time point data type

(INSTANT) and time interval data type (Period) INSTANT(YY MM DD HH MI SS) was defined as a basic typeof temporal data and parameters are expressed by yearmonth day hour minute and second The implementationof INSTANT and Period was as follows

CREATE OR REPLACE TYPE ldquoINSTANTrdquo as object(

sdot sdot sdot

constructor function instant(dateStr varchar2)return self as resultconstructor function instant(d date) return selfas resultmember function todate return datemember function tochar return varchar)

CREATE OR REPLACE TYPE ldquoPERIODrdquo ASOBJECT

(m start Datem end DateMEMBER FUNCTION PeriodLengthRETURN NUMBERMEMBER FUNCTION IsValid RETURNINTEGERsdot sdot sdot

MEMBER FUNCTION IsEqual(p period)RETURN INTEGERMEMBER FUNCTION IsCover(p period)RETURN INTEGERMEMBER FUNCTION IsIntersect(p period)RETURN INTEGERsdot sdot sdot )


MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1

I Points I Lines I PolygonsI Rectangles I Circles

I Point I Line I PolygonI Rectangle I Circle

MTrajectory MPT MLine MRectangle MPolygon MCircle MCollection

S Point S Line S PolygonS Rectangle S Circle INSTANT

Period

1lowast

1lowast

1lowast 1lowast 1lowast 1lowastlowast lowast lowast lowast lowastlowast

11 11 11

Figure 7 Class diagram of moving object model

522 Implementation of Spatial Data Types According to themoving object abstract data type spatial data types includeS Point S Line S Polygon S Rectangle and S Circle Heretaking S Rectangle as an example the implementation is asfollows

CREATE OR REPLACE TYPE ldquoS RECTANGLErdquo asobject (

ld pt s point The lower left corner coordi-natesru pt s point The upper right corner coordi-natessrid numberconstructor function s rectangle(ldpt s pointrupt s point) return self as resultmember function ToGeom return mdsyssdogeometry Conversion to SDO GEOMETRYfunction)

523 Implementation of Spatial-Temporal Data Type Basedon temporal data types and the spatial data types and usinginheritance of object-oriented method the fields and meth-ods of spatiotemporal data types (I Point I Line I RectangleI Polygon and I Circle) are inherited from the fields andmethods defined by the temporal data types and spatial datatypes Then the set type of these basic spatial-temporal data

types was also given Here taking I Point as an example theimplementation is as follows

CREATEOR REPLACE TYPE ldquoI POINTrdquo as object (

pt s pointins instantmember function distance(ipt i point) returnnumbermember function timeLength(ipt i point)return integer)

524 Implementation of Moving Object Data Type Movingobject data type is set up based on the above data types Inthis paper we focus on themoving point objectsThemovingpoint objects include the basic information of the movingobject the spatial position and positioning time and theaccuracy velocity and azimuth information the definition isas follows

CREATE OR REPLACE TYPE ldquoMPTrdquo as object (

pt s point spatial coordinate informationspeed numberbearing number moving method and toolsaccuracy number)


Figure 8 Real-time monitoring interface

The moving object trajectory segment is defined as fol-lows

CREATE OR REPLACE TYPE ldquoMOrdquo as object (

mid varchar2(100)mpt1 mpt starting pointmpt2 mpt end pointp period time intervalstate integermember function get mpt(t instant) returni point dynamic function to calculate thelocation at t)

6 Results and Analysis

We implemented the object-oriented all-time-domain mov-ing object database model and applied it to a moving objectmonitoring system This system includes database serversmoving object monitoring center and data collection ter-minals The database server was implemented by extendingOracle 11 g spatial database that each trajectory segment wasstored with the user defined types MO Moving object mon-itoring center uses BS structure The server is mainly usedfor system business logic processing including the businessprocess logic and interface services and themonitoring clientis mainly responsible for themobile terminal monitoring themain function of a mobile terminal including real-timemon-itoring historical trajectory playback and historical locationqueryThemonitoring systemwas developedwith the FineUIframework (httpfineuicodeplexcom) online geographicmap and AMap API for JavaScript (httplbsamapcom) asshown in Figure 8

Themain function of data collection terminal is to collectlocation information in real time and to send the locationinformation to data server through the 3G network andupdate position information by the method proposed in thispapermdashposition-based dynamic threshold position updatingstrategy The terminal system was developed on androidoperation system shown as in Figure 9

A case study was carried out on different moving objectFirstly we installed the data collection software on a smartphone with the global position system Then we harvested

Figure 9 Mobile terminal interface and location interface

the position data on different forms of movements includingbus taxi train and pedestrian Finally we tested the modelfrom the update frequency and the matching degree betweenthe stored moving trajectory and the actual moving one Thedynamic update frequency was compared with the equal-distance updating and the equal-time update one

The parameters settings are as follows In the dynamicthreshold position updating strategy Δ120579 = 20

∘ Δ119889 isdetermined by Table 1 and 119889max = 1000m In equal-distanceupdating strategy Δ119889 = 100m and in equal-time updatingstrategy Δ119905 = 15 sThe updating frequency result is shown asin Table 2

The experimental results are as follows

(1) For moving object such as bus taxi and train thelocation update frequency determined by the locationupdate strategy based on positioning accuracy isabout 30ndash40 of equal-time strategy and equal-distance strategy while it is 88 of that in the train

(2) For moving object such as pedestrian with slowmoving speed and irregular moving trajectory thedynamic threshold update strategy will increase thelocation update frequency in order to record trajecto-ries more accurately because of the rapid changes ofthe azimuth and accuracy information

(3) Experiments show that the dynamic threshold posi-tion updating strategy is capable of effectively reduc-ing the position update saving data transmission flowand reducing the data storage without affecting thetrajectory accuracy of the moving objects

(4) We tested thematching degree between the trajectoryderived from the stored position and the actual oneAccording to the results the derived trajectory wassatisfied which was supported by high matchingdegree


Table 2 Comparison between Dynamic position updating strategy and other position strategy

No Total Distance(km)

Total Time(s)

Updating Frequency (numbers) Updating Frequency Percentage ()Different moving object

EDS ETS DTS DTSEDS DTSETS

1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train

4 064 663 6 44 24 4000 545 pedestrianlowastIn this table EDS means Equal-Distance Strategy ETS means Equal-Time Strategy and DTS means Dynamic Threshold Strategy

7 Conclusion

Moving object databases always manage spatial-temporalinformation with massive volume It is necessary to workon spatiotemporal data model adapted to moving objects inorder to store and query the position trajectory informationeffectively We design an all-time-domain moving objectdatabase model combined based on the object-oriented ideaand dynamic attributes in MOST By analyzing the inter-structure and rendering of MOST in detail a position-baseddynamic threshold position updating strategy was proposedto fit this new data model The cubic Hermite interpolationfunction was used to estimate the historical location ofmoving objects Linear extended positioning method veloc-ity mean value positioning method and cubic exponentialsmoothing positioning method were designed and used toestimate the future location of moving object The resultverification showed that this strategy can effectively reducethe transmission and storage of data

Conflict of Interests

The authors declare that they have no conflict of interests

Acknowledgments

This paper is supported in part by the Scholarship Councilof Fujian Province (Fujian Education Cooperation (2013)no 168) by the National Technology Support Project (no2013BAH28F00) and by the National Science Foundationof China (no 41471333) One of the authors would like tothank the members of his committee for their support andthe reviewers for giving the authors constructive suggestionswhich would help them in English and in depth to improvethe quality of the paper

References

[1] P Revesz Introduction to Databases Springer London UK2010

[2] R H Guting M H Bohlen M Erwig et al ldquoA foundation forrepresenting and querying moving objectsrdquo ACM Transactionson Database Systems vol 25 no 1 pp 1ndash42 2000

[3] L Zhao Location-Based Concurrent Query Processing Technol-ogy of Moving Objects Defense Science University of Technol-ogy Changsha China 2010

[4] T Hagerstrand ldquoWhat about people in regional sciencerdquoPapers of the Regional Science Association vol 24 no 1 pp 6ndash211970

[5] G Langran and N R Chrisman ldquoA framework for temporalgeographic informationrdquo Cartographica vol 25 no 3 pp 1ndash141988

[6] M F Worboys H M Hearnshaw and D J Maguire ldquoObject-oriented data modelling for spatial databasesrdquo InternationalJournal of Geographical Information Systems vol 4 no 4 pp369ndash383 1990

[7] K S Hornsby and S Cole ldquoModelingmoving geospatial objectsfrom an event-based perspectiverdquo Transactions in GIS vol 11no 4 pp 555ndash573 2007

[8] Z Y Cao ldquoAn object-oriented spatio-temporal data modelrdquoJournal of Surveying andMapping vol 31 no 1 pp 87ndash92 2002

[9] F Bonfatti and P D Monari ldquoSpatio-temporal modeling ofcomplex geographical structuresrdquo in Proceedings of the IFIPTC5WG511Working Conference on Computer Support for Envi-ronmental Impact Assessment North-Holland 1994

[10] S-Z Yi Y Zhang and L-Z Zhou ldquoSpatial-temporal datamodel for moving area objectsrdquo Journal of Software vol 13 no8 pp 1658ndash1665 2002

[11] N Tryfona ldquoModeling phenomena in spatiotemporal data-bases desiderata and solutionsrdquo inDatabase and Expert SystemsApplications 9th International Conference DEXArsquo98 ViennaAustria August 24ndash28 1998 Proceedings vol 1460 of LectureNotes in Computer Science pp 155ndash165 Springer Berlin Ger-many 1998

[12] N Tryfona and C S Jensen ldquoConceptual data modeling forspatiotemporal applicationsrdquo GeoInformatica vol 3 no 3 pp245ndash268 1999

[13] G Trajcevski O Wolfson K Hinrichs and S ChamberlainldquoManaging uncertainty in moving objects databasesrdquo ACMTransactions on Database Systems vol 29 no 3 pp 463ndash5072004

[14] OWolfson B Xu H Yin and H Cao ldquoSearch-and-discover inmobile P2P network databasesrdquo in Proceedings of the 26th IEEEInternational Conference on Distributed Computing Systems(ICDCS rsquo06) IEEE July 2006

[15] P Jin L Yue and Y Gong ldquoDesign and implementation of aunified spatiotemporal data modelrdquo in Advances in Spatio-Tem-poral Analysis ISPRS Book Series in Photogrammetry RemoteSensing and Spatial Information Sciences pp 67ndash76 Taylor ampFrancis 2008

[16] C Z Xue C H Zhou F Z Su Q Dong and J Xie ldquoResearchon process-oriented spatio-temporal data modelrdquo Journal ofSurveying and Mapping vol 39 no 1 pp 95ndash101 2010


[17] Z-M Ding ldquoData model query language and real-time trafficflow analysis in dynamic transportation network based movingobjects databasesrdquo Journal of Software vol 20 no 7 pp 1866ndash1884 2009

[18] Z-M Ding ldquoAn index structure for frequently updated net-work-constrained moving object trajectoriesrdquo Chinese Journalof Computers vol 35 no 7 pp 1448ndash1461 2012

[19] J Yang and Z Fei ldquoBroadcasting with prediction and selectiveforwarding in vehicular networksrdquo International Journal of Dis-tributed Sensor Networks vol 2013 Article ID 309041 9 pages2013

[20] J Yang and Z Fei ldquoStatistical filtering based broadcast protocolfor vehicular networksrdquo in Proceedings of 20th InternationalConference on Statistical Filtering Based Broadcast Protocol forVehicular Networks Computer Communications and Networks(ICCCN rsquo11) pp 1ndash6 Maui Hawaii USA 2011

[21] F Zhang and Q-J Cao ldquoStudy on existing spatio-temporaldata modelsrdquo Journal of University of Shanghai for Science andTechnology vol 27 no 6 pp 530ndash534 2005

[22] L-B Ma and X-C Zhang ldquoResearch on full-period queryoriented moving objects spatio-temporal data modelrdquo ActaGeodaetica et Cartographica Sinica vol 37 no 2 pp 207ndash2222008

Submit your manuscripts athttpwwwhindawicom

VLSI Design

Hindawi Publishing Corporationhttpwwwhindawicom Volume 2014

International Journal of

RotatingMachinery


Hindawi Publishing Corporation httpwwwhindawicom

Journal ofEngineeringVolume 2014


Shock and Vibration


Mechanical Engineering

Advances in


Civil EngineeringAdvances in

Acoustics and VibrationAdvances in



Electrical and Computer Engineering

Journal of


Distributed Sensor Networks


The Scientific World JournalHindawi Publishing Corporation httpwwwhindawicom Volume 2014

SensorsJournal of


Modelling amp Simulation in EngineeringHindawi Publishing Corporation httpwwwhindawicom Volume 2014


Active and Passive Electronic Components


Chemical EngineeringInternational Journal of

Control Scienceand Engineering

Journal of


Antennas andPropagation




Navigation and Observation


Advances inOptoElectronics


Volume 2014

RoboticsJournal of



entities were given Ding carried out the research on theDynamic Transportation Network Based Moving ObjectsDatabase (DTNMOD) and its real-time dynamic transportnetwork flow analysis method based on the spatiotemporaltrajectory of moving objects [17 18] Yang et al developeda protocol based on prediction that significantly boosts thespeed of message transmission over vehicles [19 20] Fromthe above analysis and comparisons it is indicated thatobject-oriented model and moving object model workedbetter in spatiotemporal object modeling and have theadvantages in spatial query time query and spatiotemporalquery But in the continuous moving modeling the datatransmission and data storage also have some disadvantages

The traditional location update method was time pointselection method including equal-time algorithm equal-distance algorithm and dynamic time point algorithm [2122] Equal-time algorithm was the most simple and wasmore effective to achieve and the communication flow isstable but the adaptability was poorWhen the moving objectis static it also records the position which causes invalidstorage Equal-distance algorithm has dramatically increasedcommunication flow in high-speed motion dynamic timepoint algorithm is capable of reducing the position updatebut this method is only suitable for high positioning accuracyand stably changing positioning parameters

According to the deficiencies of continuousmovingmod-eling multidimensional expression and complex spatiotem-poral queries an object-oriented moving object databasemodel supporting all-time-domain query was proposed bycombining static object model and moving object model Inorder to reduce the amount of moving object data trans-mission and storage a dynamic position updating strategywas studied It is suitable for frequent changes of positionof moving objects This model solved the problem of con-tinuous storage multidimensional expression and complexspatiotemporal queries This paper realized the model by aninstance at last

2 Moving Object Modeling

21 Moving Object Trajectory Segment Model The motionand change of moving object are continuous In the tradi-tionalmoving object data storagemodel the object-relationaldatabase was used to store the information continuouslyobtained from the moving object at a certain time intervalIt not only increases the amount of the data storage butalso increases the difficulties in the history and future spa-tiotemporal queries of themoving objects and in the complexspatiotemporal behavior queries In order to solve theseproblems dynamic properties methods of the MOST modelare proposed and implemented in this paper The discretemodeling ideas are proposed to support dynamic propertiesof the entire trajectory model and all the trajectories ofmoving objects are separated into many small segments asshown in Figure 1

In Figure 1 119905 axis means the time (119909 119910)means the pointinformation of moving objects and 119901119905

119894is the coordinate

information at 119905119894moment 119898119901119905

119894stands for the position

information of moving objects at 119905119894moment Each discrete

x

ytt1 t2 t3 t4 t5

pt4

pt2

pt3

pt1

pt5

mpt2

mpt1

mpt3

mpt4

mpt5

Figure 1 Moving trajectory decomposition of moving object

segment is abstracted as dynamic changes within a certainperiod of time 119875

In order to support the historical trajectories of movingobjects modeling the quadruple (119860 119875 119865 119878) was put for-ward to support the all-time-domain query All trajectoriessegments expressed by the quadruple constitute continuoustrajectories of moving objects

119860 = (119881119872119875119879119894119872119875119879

119894+1)

= (119881 (119901119905119894 V119894 119887119894 119886119894) (119901119905

119894+1 V119894+1 119887119894+1 119886119894+1))

119875 = [119905119894 119905119894+1)

119865 = 119891 (119905)

119878 = 1 2 3

(1)

Here 119860 stands for basic information of moving objectssuch as moving object ID position direction speed accu-racy and other information 119881 indicates the basic attributeinformation of moving objects such as moving object IDname and type 119872119875119879

119894is the position information of the

moving object in the moment 119905119894119872119875119879

119894= (119901119905

119894 V119894 119887119894 119886119894) 119901119905119894

is the coordinate information at 119905119894moment V

119894is the velocity

at 119905119894 119887119894is the azimuth at 119905

119894 and 119886

119894is the accuracy at 119905

119894 In this

paper the dynamic properties of the model are built on thelocation (119872119875119879

119894) change over time 119875

119875 indicates the time the object moves 119875 is not a fixedtime interval but a right open interval which is uncertainThesize of 119875 is related to the updating strategy of moving objectlocation

119865 represents motion function of moving objects 119891(119905)is the motion function of moving objects and is a time-dependent function by which we can estimate the positionof a moving object in the past at present and in the futurewithin a short time Motion function can be divided intomany types according to the actual situation such as linearfunction and curvilinear function

119878 is the state of the moving objects in the period 119875 Thereare mainly three types of the moving object state 119878 pastcurrent and network failure or interruption respectivelyrepresented by 1 2 and 3 When 119878 = 1 the position is usedto store and query the historical trajectory when 119878 = 2 thatis used to store current position and estimate future positionwhen 119878 = 3 the location information will not be stored untilthe terminal receives the location information again Whenthe update data is firstly or lastly received a new moving



PT

Speed

Accurate

Azimuth

Longitude

Latitude

SRID


Now

None

PT

ST Operation

perationperation

T Operation

S Instant

E Instant

MO Entity

S OS O



119894+1when












p0

p1

p2

p3

p4

f0f1

A

B

D

C

f2

f3

f4

f5

f6















GPS data package

Data analysis

Position data


Azimuth threshold


Start

End

Position updating

Yes

NoYes

No

No

Yes

No

Yes

t minus t0 gt tmax

Velocity = 0




















1198981199011199052(119901119905(1199092 1199102) V2 119886 119887) in the period 119875

119891 (119905) = 1199041 + V sdot Δ119905

= (1199091+V1+ V2

2sdot Δ119905 sin120572 119910

1+V1+ V2

2sdot Δ119905 cos120572)

(2)

t

v

y

x

yx

S(tj)G(tj)

f(tj)

998400x

t998400

x998400

y998400

998400y















119895)


119895) and


119895) = 119910119895 119866(119905119895) = 119909119895

Then use 119878(119905119895) and 119866(119905


time 119905Let 119878(119905

119895) = 119910




then

119878 (119905119895) =

3

sum119895=0

[119910119895120572119895 (119905) + 119898119895120573119895 (119905)] (3)



119895(119905) and 120573





120572119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895minus1minus 119905119895

) 119905119895minus1

le 119905 le 119905119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895+1minus 119905119895

) 119905119895le 119905 le 119905

119895+1(119895 = 0 1 2)

120573119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (119905 minus 119905119895) 119905

119895minus1le 119905 le 119905

119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (119905 minus 119905119895) 119905

119895le 119905 le 119905

119895+1(119895 = 0 1 2)

(4)





119905| and 120579

119905


1003816100381610038161003816 =1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(5)


1003816100381610038161003816V119905+Δ1199051003816100381610038161003816 =

1

119898

119898

sum119894=1

1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(6)





1 1199052 1199053 119905

119896 Besides




4


0= (13)(V

1+ V2+ V3)

Let 119878(1)119905 119878(2)

119905 119878(3)





119878(1)

119905= 120572V119905+ (1 minus 120572) 119878

(1)

119905minus1

119878(2)

119905= 120572119878(1)

119905+ (1 minus 120572) 119878

(2)

119905minus1

119878(3)

119905= 120572119878(2)

119905+ (1 minus 120572) 119878

(3)

119905minus1

119886119905= 3119878(1)

119905minus 3119878(2)

119905+ 119878(3)

119905

119887119905=

120572

2 (1 minus 119886)2

sdot [(6 minus 5120572) 119878(1)

119905minus 2 (5 minus 4120572) 119878

(2)

119905+ (4 minus 3120572) 119878

(3)

119905]

119888119905=

1205722

2 (1 minus 119886)2[119878(1)

119905minus 2119878(2)

119905+ 119878(3)

119905]

(7)


119905and


V119905+Δ119905

= 119886119905+ 119887119905Δ119905 + 119888

119905Δ119905

120579119905+Δ119905

= 120579119905

(8)





Instant

Period

Periods

MPT MO



NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle



TAB MO










sdot sdot sdot






MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of




PT

Speed

Accurate

Azimuth

Longitude

Latitude

SRID


Now

None

PT

ST Operation

perationperation

T Operation

S Instant

E Instant

MO Entity

S OS O



119894+1when












p0

p1

p2

p3

p4

f0f1

A

B

D

C

f2

f3

f4

f5

f6















GPS data package

Data analysis

Position data


Azimuth threshold


Start

End

Position updating

Yes

NoYes

No

No

Yes

No

Yes

t minus t0 gt tmax

Velocity = 0




















1198981199011199052(119901119905(1199092 1199102) V2 119886 119887) in the period 119875

119891 (119905) = 1199041 + V sdot Δ119905

= (1199091+V1+ V2

2sdot Δ119905 sin120572 119910

1+V1+ V2

2sdot Δ119905 cos120572)

(2)

t

v

y

x

yx

S(tj)G(tj)

f(tj)

998400x

t998400

x998400

y998400

998400y















119895)


119895) and


119895) = 119910119895 119866(119905119895) = 119909119895

Then use 119878(119905119895) and 119866(119905


time 119905Let 119878(119905

119895) = 119910




then

119878 (119905119895) =

3

sum119895=0

[119910119895120572119895 (119905) + 119898119895120573119895 (119905)] (3)



119895(119905) and 120573





120572119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895minus1minus 119905119895

) 119905119895minus1

le 119905 le 119905119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895+1minus 119905119895

) 119905119895le 119905 le 119905

119895+1(119895 = 0 1 2)

120573119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (119905 minus 119905119895) 119905

119895minus1le 119905 le 119905

119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (119905 minus 119905119895) 119905

119895le 119905 le 119905

119895+1(119895 = 0 1 2)

(4)





119905| and 120579

119905


1003816100381610038161003816 =1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(5)


1003816100381610038161003816V119905+Δ1199051003816100381610038161003816 =

1

119898

119898

sum119894=1

1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(6)





1 1199052 1199053 119905

119896 Besides




4


0= (13)(V

1+ V2+ V3)

Let 119878(1)119905 119878(2)

119905 119878(3)





119878(1)

119905= 120572V119905+ (1 minus 120572) 119878

(1)

119905minus1

119878(2)

119905= 120572119878(1)

119905+ (1 minus 120572) 119878

(2)

119905minus1

119878(3)

119905= 120572119878(2)

119905+ (1 minus 120572) 119878

(3)

119905minus1

119886119905= 3119878(1)

119905minus 3119878(2)

119905+ 119878(3)

119905

119887119905=

120572

2 (1 minus 119886)2

sdot [(6 minus 5120572) 119878(1)

119905minus 2 (5 minus 4120572) 119878

(2)

119905+ (4 minus 3120572) 119878

(3)

119905]

119888119905=

1205722

2 (1 minus 119886)2[119878(1)

119905minus 2119878(2)

119905+ 119878(3)

119905]

(7)


119905and


V119905+Δ119905

= 119886119905+ 119887119905Δ119905 + 119888

119905Δ119905

120579119905+Δ119905

= 120579119905

(8)





Instant

Period

Periods

MPT MO



NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle



TAB MO










sdot sdot sdot






MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of



p0

p1

p2

p3

p4

f0f1

A

B

D

C

f2

f3

f4

f5

f6















GPS data package

Data analysis

Position data


Azimuth threshold


Start

End

Position updating

Yes

NoYes

No

No

Yes

No

Yes

t minus t0 gt tmax

Velocity = 0




















1198981199011199052(119901119905(1199092 1199102) V2 119886 119887) in the period 119875

119891 (119905) = 1199041 + V sdot Δ119905

= (1199091+V1+ V2

2sdot Δ119905 sin120572 119910

1+V1+ V2

2sdot Δ119905 cos120572)

(2)

t

v

y

x

yx

S(tj)G(tj)

f(tj)

998400x

t998400

x998400

y998400

998400y















119895)


119895) and


119895) = 119910119895 119866(119905119895) = 119909119895

Then use 119878(119905119895) and 119866(119905


time 119905Let 119878(119905

119895) = 119910




then

119878 (119905119895) =

3

sum119895=0

[119910119895120572119895 (119905) + 119898119895120573119895 (119905)] (3)



119895(119905) and 120573





120572119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895minus1minus 119905119895

) 119905119895minus1

le 119905 le 119905119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895+1minus 119905119895

) 119905119895le 119905 le 119905

119895+1(119895 = 0 1 2)

120573119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (119905 minus 119905119895) 119905

119895minus1le 119905 le 119905

119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (119905 minus 119905119895) 119905

119895le 119905 le 119905

119895+1(119895 = 0 1 2)

(4)





119905| and 120579

119905


1003816100381610038161003816 =1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(5)


1003816100381610038161003816V119905+Δ1199051003816100381610038161003816 =

1

119898

119898

sum119894=1

1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(6)





1 1199052 1199053 119905

119896 Besides




4


0= (13)(V

1+ V2+ V3)

Let 119878(1)119905 119878(2)

119905 119878(3)





119878(1)

119905= 120572V119905+ (1 minus 120572) 119878

(1)

119905minus1

119878(2)

119905= 120572119878(1)

119905+ (1 minus 120572) 119878

(2)

119905minus1

119878(3)

119905= 120572119878(2)

119905+ (1 minus 120572) 119878

(3)

119905minus1

119886119905= 3119878(1)

119905minus 3119878(2)

119905+ 119878(3)

119905

119887119905=

120572

2 (1 minus 119886)2

sdot [(6 minus 5120572) 119878(1)

119905minus 2 (5 minus 4120572) 119878

(2)

119905+ (4 minus 3120572) 119878

(3)

119905]

119888119905=

1205722

2 (1 minus 119886)2[119878(1)

119905minus 2119878(2)

119905+ 119878(3)

119905]

(7)


119905and


V119905+Δ119905

= 119886119905+ 119887119905Δ119905 + 119888

119905Δ119905

120579119905+Δ119905

= 120579119905

(8)





Instant

Period

Periods

MPT MO



NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle



TAB MO










sdot sdot sdot






MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of














1198981199011199052(119901119905(1199092 1199102) V2 119886 119887) in the period 119875

119891 (119905) = 1199041 + V sdot Δ119905

= (1199091+V1+ V2

2sdot Δ119905 sin120572 119910

1+V1+ V2

2sdot Δ119905 cos120572)

(2)

t

v

y

x

yx

S(tj)G(tj)

f(tj)

998400x

t998400

x998400

y998400

998400y















119895)


119895) and


119895) = 119910119895 119866(119905119895) = 119909119895

Then use 119878(119905119895) and 119866(119905


time 119905Let 119878(119905

119895) = 119910




then

119878 (119905119895) =

3

sum119895=0

[119910119895120572119895 (119905) + 119898119895120573119895 (119905)] (3)



119895(119905) and 120573





120572119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895minus1minus 119905119895

) 119905119895minus1

le 119905 le 119905119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895+1minus 119905119895

) 119905119895le 119905 le 119905

119895+1(119895 = 0 1 2)

120573119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (119905 minus 119905119895) 119905

119895minus1le 119905 le 119905

119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (119905 minus 119905119895) 119905

119895le 119905 le 119905

119895+1(119895 = 0 1 2)

(4)





119905| and 120579

119905


1003816100381610038161003816 =1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(5)


1003816100381610038161003816V119905+Δ1199051003816100381610038161003816 =

1

119898

119898

sum119894=1

1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(6)





1 1199052 1199053 119905

119896 Besides




4


0= (13)(V

1+ V2+ V3)

Let 119878(1)119905 119878(2)

119905 119878(3)





119878(1)

119905= 120572V119905+ (1 minus 120572) 119878

(1)

119905minus1

119878(2)

119905= 120572119878(1)

119905+ (1 minus 120572) 119878

(2)

119905minus1

119878(3)

119905= 120572119878(2)

119905+ (1 minus 120572) 119878

(3)

119905minus1

119886119905= 3119878(1)

119905minus 3119878(2)

119905+ 119878(3)

119905

119887119905=

120572

2 (1 minus 119886)2

sdot [(6 minus 5120572) 119878(1)

119905minus 2 (5 minus 4120572) 119878

(2)

119905+ (4 minus 3120572) 119878

(3)

119905]

119888119905=

1205722

2 (1 minus 119886)2[119878(1)

119905minus 2119878(2)

119905+ 119878(3)

119905]

(7)


119905and


V119905+Δ119905

= 119886119905+ 119887119905Δ119905 + 119888

119905Δ119905

120579119905+Δ119905

= 120579119905

(8)





Instant

Period

Periods

MPT MO



NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle



TAB MO










sdot sdot sdot






MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of




120572119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895minus1minus 119905119895

) 119905119895minus1

le 119905 le 119905119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (1 + 2119905 minus 119905119895

119905119895+1minus 119905119895

) 119905119895le 119905 le 119905

119895+1(119895 = 0 1 2)

120573119895 (119905) =

(119905 minus 119905119895minus1

119905119895minus 119905119895minus1

)

2

sdot (119905 minus 119905119895) 119905

119895minus1le 119905 le 119905

119895(119895 = 1 2 3)

(119905 minus 119905119895+1

119905119895minus 119905119895+1

)

2

sdot (119905 minus 119905119895) 119905

119895le 119905 le 119905

119895+1(119895 = 0 1 2)

(4)





119905| and 120579

119905


1003816100381610038161003816 =1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(5)


1003816100381610038161003816V119905+Δ1199051003816100381610038161003816 =

1

119898

119898

sum119894=1

1003816100381610038161003816V1199051003816100381610038161003816

120579119905+Δ119905

= 120579119905

(6)





1 1199052 1199053 119905

119896 Besides




4


0= (13)(V

1+ V2+ V3)

Let 119878(1)119905 119878(2)

119905 119878(3)





119878(1)

119905= 120572V119905+ (1 minus 120572) 119878

(1)

119905minus1

119878(2)

119905= 120572119878(1)

119905+ (1 minus 120572) 119878

(2)

119905minus1

119878(3)

119905= 120572119878(2)

119905+ (1 minus 120572) 119878

(3)

119905minus1

119886119905= 3119878(1)

119905minus 3119878(2)

119905+ 119878(3)

119905

119887119905=

120572

2 (1 minus 119886)2

sdot [(6 minus 5120572) 119878(1)

119905minus 2 (5 minus 4120572) 119878

(2)

119905+ (4 minus 3120572) 119878

(3)

119905]

119888119905=

1205722

2 (1 minus 119886)2[119878(1)

119905minus 2119878(2)

119905+ 119878(3)

119905]

(7)


119905and


V119905+Δ119905

= 119886119905+ 119887119905Δ119905 + 119888

119905Δ119905

120579119905+Δ119905

= 120579119905

(8)





Instant

Period

Periods

MPT MO



NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle



TAB MO










sdot sdot sdot






MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of



Instant

Period

Periods

MPT MO



NumberInt Varchar

Base DT

MO ADT

Date

SDOGeometry

S Point

S Points

S Line

S Polygon

S Rectangle

S Circle



TAB MO










sdot sdot sdot






MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of



MO

1 11 1 1 1 1 1 1

1

1

1

1 1 1 1

1





Period

1lowast

1lowast


11 11 11

































Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of























Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of





Total Time(s)



1 520 870 52 58 19 365 328 Bus

2 710 956 71 70 22 310 314 Taxi

3 11444 3600 1144 240 101 88 721 Train


7 Conclusion




Acknowledgments


References

























VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of










VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of



VLSI Design



RotatingMachinery





Shock and Vibration



Advances in







Journal of





SensorsJournal of








Journal of










Volume 2014

RoboticsJournal of


Documents

An All-Time-Domain Moving Object Data Model, Location ...faculty.ung.edu/jyang/research_files/papers/wuqy.pdf · To estimate the posi-tion of moving objects at 𝑡time, trajectory