Transcript bipin-mobile-presentation

Location Information
Management and Moving Object
Databases
“Moving Object Databases: Issues
and Solutions”
Ouri, Bo, Sam and Liqin
1
Contents
Introduction
 Challenges of Location Information
Management

–
–
–
–
Location Modeling
Uncertainty Management
Distributed/Mobile Environment
Location Prediction
2
Contents contd…

Moving Object Databases
– Location Modeling
» Dynamic Attributes
» Data Model
– Linguistics Issues
» Spatial and Temporal Query Language
– Indexing
– Uncertainty/Imprecision Management


Advantages and Disadvantages
Conclusions and Future work
3
Introduction
Location management is managing location
information of mobile devices and using the
data in various applications.
 Location management is a fundamental
component of applications like location
based services, tourist services, disaster
recovery, context awareness, and dynamic
resource discovery.

4
Introduction contd..


Location-based services can be divided into two
categories - Mobile Resource Management, and
Location-Aware Content Delivery
Mobile Resource Management applications
– Mobile workforce management
– Automatic location Management
– Transportation Management
These Systems use location data combined with route
schedules to track and manage service personnel or
transportation systems.
5
Location-aware Content Delivery
Services

These services use location data to tailor the
information delivered to the mobile user to
increase relevancy.
– Delivering driving directions
– Instant coupons to customers nearing a store
– Nearest resource information like local
restaurants, hospitals, ATM machines, or gas
stations
6
Example queries to be answered by a
Location Management System
Retrieve the helicopters that enter region R
within next 10 minutes( Future Query )
 How many times was bus#5 late by more
than 5 minutes at some station( Past Query )
 Send a message when a helicopter is in a
given Region( Trigger )

7
Required Capabilities of a
Location Management System
Modeling of Location Information
 Uncertainty Management
 Spatio-temporal data access languages
 Indexing and scalability issues
 Data mining(traffic and location prediction)
 Privacy and Security
 Synchronization of data from multiple
sensors

8
Location Modeling
Existing DBMS are not well equipped to
handle continuously changing data. Here
data is assumed to be constant unless
explicitly notified.
 To represent moving objects to the database
has two disadvantages

– DBMS cannot handle frequent updates from
many mobile devices
– Frequent updates impose a serious wireless
bandwidth overhead.
9
Using Dynamic attributes to
model data
This approach tries to use DBMS will new
attribute called the “dynamic attribute” to
model data
 Dynamic attribute is used to store location
as a function of time
 The answer to a query depends on database
contents and also the time when the query is
asked.
 A higher level of abstraction is used
10

Dynamic attribute

A Dynamic Attribute A is represented by three
sub-attributes
– A.update_value
– A.update_time
– A.function


The value of dynamic attribute at time
A.update_time + t0 is A.update_value +
A.function( t0 )
The location attribute has sub-attributes L.x, L.y,
L.speed, and L.angle or L.route, where x and y are
of type A
11
Example on Dynamic attribute
Speed = 10miles/unit_time, Direction = North
update_value = 100
update_time = 5
function = 10*t for y direction

Value of y at time 15 = 100 + 10*(15-5)
= 100+100 = 200
12
Linguistics Issues
Traditional query languages such as SQL
are inadequate for expressing queries
involving spatio-temporal data
 Temporal predicates used

– begin_time( … )
– end_time( … )
– now
13
Spatial Query language predicates
DIST( o, n ), where o and n are two mobile
devices
 INSIDE( o, R ), where o is a mobile device
and R is a region

14
Queries using defined predicates

Retrieve the pairs of objects o and n such
that the distance between o and n stays
within 5 miles until they both enter the
polynomial P
RETRIEVE o, n
FROM Moving-Objects
WHERE begin_time( DIST(o,n) <=5 )<=now
and end_time( DIST(o,n) <= 5 ) >=
begin_time( INSIDE(o,P) ^ INSIDE(n,P))
15
Other Temporal Query Predicates

Future queries
–
–
–
–
–
UNTIL
NextTime
Eventually_within( c, g )
Eventually_after( c, g )
Always_for( c, g )
16
Example Future Query

“ Retrieve all objects o that enter the
polygon P within three units of time and
have the attribute PRICE <= 100”
RETRIEVE o
WHERE o.PRICE <=100 ^
Eventually_within( 3, INSIDE(o,P) )
17
Indexing Dynamic attributes
To index queries like “Retrieve the objects
that are currently inside the polygon P” or
“Retrieve the objects whose dynamic attribute
value is in the range[Ac …. Ae] at time t”
 R+ trees cannot be used to index because they
cannot be used to model temporal data

18
Indexing Contd..

Value-time space representation
19
Uncertainty Management
The location of a moving object is
inherently imprecise regardless the update
policy.
 To accommodate uncertainty, a new subattribute to the dynamic attribute named
L.uncertainty should be added.
 “May” and “must” semantics should be
incorporated in the query language

20
Implementation Details

Architecture
21
Implementation Details contd..
DBMS: Oracle database is selected as
DBMS because it handles multi-user access
which is an inherent property.
 A Wrapper application shall be used on
DBMS, because normal DBMS do not have
spatio-temporal query handling capabilities.
 Because the mobile device have movement
involved, updations to the location data is
web enabled, a servlet is used

22
Advantages
Simple approach to handle location
management using DBMS.
 Very few updates to database are needed.
 The architecture shall not cost much to
construct.
 Answers queries that involve both temporal
and spatial data.
 Handles uncertainty of mobile unit location.

23
Disadvantages
The mobile unit should have capability to
give its location uncertainty.
 Dynamic attribute may also need to be
updated frequently if the mobility pattern is
not smooth.
 A function for the dynamic attribute may
not be able to expressed in few cases.

24
Conclusion



Advances in wireless communication and sensor
technology are few forces that are propagating
computing from desktops to outdoor mobile units.
Location management is the key to applications
used by mobile units such as mobile resource
management applications and location-aware
content delivery applications.
Location management is key to answer queries
that involve spatio-temporal data and answer
issues like location modeling, uncertainity
management, and location prediction
25
Future work
Extending the architecture of centralized
database to a distributed database.
 Handle if uncertain data is submitted from
different sensors
 Study the implications of network QoS on
update policies and query processing.
 Integrate with GIS

26