Semantic Space : An Infrastructure For Smart Spaces

1
Semantic Space An Infrastructure For Smart
Spaces

• 2006.1.25
• ???

2
Introduction

• Focus on three key issue
• Explicit representation
• Context meanings (or semantics)
• Context querying
• Applications might need to selectively access a
  subset of contexts
• Context reasoning
• Higher-level context
• Semantic Web
• Helps computers and people work better together
  by giving content well-defined meaning
• Using standards to represent machine-interpretable
  information
• RDF, OWL
• Semantic Space
• Offers a united approach to knowledge management
  and information processing

3
The context model

• The context model
• An ontology approach to context modeling
• Common understanding of context structure
• OWL ontologies enable applications to interpret
  contexts
• Hierarchical structure
• Reuse domain ontologies
• Semantic Web tool can support context
  interpretation
• Query, reasoning, knowledge bases
• Designing the context model
• Upper-Level Context Ontology (ULCO)
• Provide a set of basic concepts common across
  different environments
• Three classes of real-world objects (user,
  location, and computing entity)
• One class of conceptual objects (activity)
• Knowledge reuse is one important advantage of
  ontologies

4
The context model
5
The context model

• Marking up real-world contexts
• Represents context as ontology instances and
  associated properties (context-markup)

ltUser rdfaboutRossGellergt ltnamegtRoss
Gellerlt/namegt ltmboxgtross_at_i2r.a-star.edu.sglt/mboxgt
lthomepage rdfresourcewww.i2r.a-star.edu.sg/ros
s/gt ltoffice rdfresourceRoom209/gt ltofficePhon
egt1234lt/officePhonegt ltmobilePhonegt6789lt/mobilePhon
egt ltsupervisorOfrdfresourceJoeyTribbiani/gt lt!
More properties not shown in this
examplegt lt/Usergt
6
The Semantic Space infrastructure
7
The Semantic Space infrastructure

• Context wrappers
• Obtain raw context information and transform them
  into context markups
• Location, environment, light, noise, software
  activity
• Self-Configuring components
• Implemented as UPnP (Universal Plug and Play)
• Context aggregator
• Discover context wrappers and gathers context
  markups
• UPnP control point
• Context knowledge base (CKB)
• Provide persistent context knowledge storage
• Stores the extended context ontology for that
  particular space and context markups
• by user or gathered from context wrappers
• When a context wrapper joins the smart space
• Context aggregator adds the provided context
• When the wrapper leaves
• The aggregators deletes the context

8
The Semantic Space infrastructure

• Context query engine
• Provides an abstract interface for application to
  extract desired context from the CKB
• RDF Data Query language
• ltsubject, predicate, objectgt patterns
• Context reasoner
• Infers abstract, higher-level contexts from basic
  sensed contexts
• Making it easy for developers to realize
  application-specific inferences simply by
  defining heuristic rules
• Jena2 generic rule engine

9
The Semantic Space infrastructure

• Context reasoner
• Sample rules that infer a user's likely situation
  based on context, activity, location, and
  computing entity

10
Conclusion

• Future Work
• Multiple smart spaces
• LAN discovery protocol UPnP
• Privacy concerns
• Support for uncertain contexts
• Source-based context arent always precise

11
ANTS UPnP Sensor Network Architecture

• 2006.11.27
• ???

12
Introduction

• Evolvable Sensor Network
• Several billions of heterogeneous sensor networks
  connected world wide
• How this ubiquitous, heterogeneous, and
  interoperable sensor networks should be handled?
  
• Evolvability
• ANTS An evolvable Network of Tiny Sensors
• Adaptable to new environments and applications
• Intelligent or learning capability
• Capable to accommodate possible new services in
  the future
• Evolvable components without interrupting current
  services

13
Introduction

• Evolvable Sensor Networks Features
• Scalability
• Ease of adaptation to changing necessities
• Adaptability
• Ability to change with environment circumstances
• Upgradeability
• Dynamic updating processes
• Interoperability
• Interaction with external systems
• Modularity
• Module based architecture
• Dependability
• Tolerant to errors during the system operations

14
Introduction

• ANTS-Hardware
• ANTS-H1, H2, H3, H4
• ANTS-OS
• ANTS-EOS, ANTS-PM
• ANTS-Network Protocols
• ANTS-Network-Architecture
• ANTS-PAD, PAC, Coverage, PPVR
• ANTS-ZigBee-RT, Summa
• ANTS-Middleware
• ANTS-ContextAware
• ANTS-ContextFusion
• ANTS-UPnP, ANTS-EVM
• ANTS-Security
• ANTS-TSF, ANTS-TSR, ANTS-CKM
• ANTS-Localization
• ANTS-PLDL
• ANTS-Synchronization
• ANTS-Csync

15
Middleware for Sensor Networks

• Design Principles
• Providing abstractions and mechanisms for dealing
  with the heterogeneity of sensor nodes
• Supporting development, maintenance, deployment,
  execution of sensing-based applications
• Middleware Functions
• Re-programming Sensor Networks / Application
  Updater
• Application Adapter
• Sensor Database
• Data Placement Replication
• Power Management
• Context Aware
• Etc.

16
Sensor Network Management

• Problem Statement
• How to configure and manage sensor network system
  easily?
• Applying UPnP to the sensor network
• The advantages of UPnP
• support zero-configuration
• allow users automatic discovery of sensor network
  services
• Another Problem
• UPnP is not suitable for the tiny sensor devices.
• BOSS (Bridge Of the SensorS) Architecture
• bridge function between UPnP Control point and
  tiny sensor nodes
• provide the sensor network management services

17
Sensor Network Management

• UPnP-based Sensor Network management Architecture

18
Sensor Network Management

• BOSS Bridge Of SensorS Architecture
• Service Manager
• Discovery service
• Manage service table
• Advertises service to the control point
• Control Manger
• Any of actions included in the device services
• Using SOAP based UPnP
• Event Manager
• UPnP event handling
• GENA (General Event Notification Architecture)
  protocol is formatted using XML

19
ANTS-ContextFusion

• Motivation
• Centralized decision making for contexts may
  cause
• More wasted transfer energy
• Relatively slow response (reaction)
• Reliability problem (e.g., homing attack, fire?)
• Our approach is Two-tiered network architecture
• Low-tier
• data dissemination to BS for satisfying
  fundamental requirements of WSN
• High-tier
• context overlay for optimized context detection
  and in-network actuation

20
ANTS-ContextFusion

• Key features
• User defined context actuation descriptor
  injection
• Node level context detection
• Adaptive to environmental changes
• Fully distributed way
• No interaction with BS
• Early detection of failed context

21
Conclusion

• ANTS is building complete architecture
• We have chances in Sensor Network IT applications
• Starting with Sensor Networks RFID Mobile

22
Paper Work

• 2006.11.27
• ???

23
Paper work 1

• ????? ??
• ??? ??
• sensorDataList
• sensorData
• Device name, networkID, sensorID, deviceType
• Data type, format, value, timestamp
• Options
• Options
• Type
• sensorData location, power, storage, correlated
• deviceData deviceList
• requestData request(device, data, user)

24
Paper work 1

• ?? ??? XML ?? (RFID)

lt?xml version"1.0" encoding"UTF-8" ?gt
ltsensorDataListgt ltsensorDatagt ltdevice
nameReader1gt ltnetworkIDgtnetworkIDlt/networkIDgt
ltsensorIDgtsensorIDlt/sensorIDgt ltdeviceTypegtRF
IDlt/deviceTypegt lt/devicegt ltdatagt lttypegtEPC
lt/typegt ltformatgtStringlt/formatgt ltvaluegt00000
0000000610329037666lt/valuegt lttimestampgt00000000
lt/timestampgt lt/datagt ltoptionsgt ltstoragegt
ltformatgtbooleanlt/formatgt ltvaluegttrueltvaluegt
lt/storagegt lt/optionsgt lt/sensorDatagt ltsens
orDatagt ltdevice nameReader1gt ltnetworkIDgtnetw
orkIDlt/networkIDgt ltsensorIDgtsensorIDlt/sensorIDgt
ltdeviceTypegtRFIDlt/deviceTypegt lt/devicegt lt
datagt lttypegtEPClt/typegt ltformatgtStringlt/forma
tgt ltvaluegt000000000000610329037667lt/valuegt lt
timestampgt00000010lt/timestampgt lt/datagt ltoptio
nsgt ltstoragegt ltformatgtbooleanlt/formatgt
ltvaluegttrueltvaluegt lt/storagegt lt/optionsgt lt/s
ensorDatagt ltsensorDataListgt
25
Paper work 2

• Correlated Data
• Query? ??? ?? ???? ??
• Query??? CEP???? ??
• Stream Data? ???
• E1? Reader 1? ???? ??
• E2? Sensor 1? ???? ??
• 2?? ???(XML)? ??
• E1? E2? correlated data? ??
• 1?? ???(XML)? ??
• ????? ???? data? ??

26
Paper work 2

• Correlated Data
• CEP? ?? ??
• Explicit correlation
• ?? ??? Event? ???? ?? ??
• E1(reader1, t1, EPC)
• E2(reader2, t1, EPC)
• E3(sensor1, t1, 20)
• (reader1, reader2, sensor1)? ??
  eventName_correlated_1 ????
• ECA Rule? define ??? ?? ??
• E1 Observation(sensor1, t1, null)E2
  Observation(sensor2, t1, null)ON ......ACTION
• (sensor1, sensor2)? ?? eventName_correlated_2 ????

27
Paper work 2

• Correlated Data
• Data Provider?? ??
• Implicit correlation
• Time based correlate
• Subscribe(sensor, time) ??? time ???? ??
• Count based correlate
• Reference count? ???? ??
• ??? ??? correlated data? ??? ??
• Client based correlate
• ???? ?? ?? time or count based correlate ???
• Remote data? ?? ?? ??? ???(client)? ??? data?
  ???? ??
• ???(client)? ???? ?? ? time or count based
  correlate ??

28
Paper work 2

• Data Provider ?? Component
• Subscribe manager
• Subscribe? ?? table?? ? pattern ??
• Client?? Subscribe ??
• Correlated data generator
• Time, count, client? ???? ??
• Correlated data manager
• ??? correlated data? ??
• Implicit correlated data? ?? ??
• Garbage correlated data ??

29
Paper work 2

• Time based correlate

Subscribe(sensor1, 5) Subscribe(sensor1, 10)
Subscribe(sensor2, 6) Subscribe(sensor2, 10)
Subscribe(sensor3, 20) Subscribe(sensor4, 10)
Subscribe(sensor5, 60) Subscribe(sensor6, 60)
Subscribe(sensor7, 60) Subscribe(sensor8, 30)
Subscribe(sensor9, 30) Subscribe(sensor10, 60)
Subscribe(sensor5, 60) Subscribe(sensor6, 60)
Subscribe(sensor7, 60) Subscribe(sensor10, 60)
Subscribe(sensor1, 10) Subscribe(sensor2, 10)
Subscribe(sensor4, 10) Subscribe(sensor8, 30)
Subscribe(sensor9, 30) Subscribe(sensor1, 5)
Subscribe(sensor2, 6) Subscribe(sensor3, 20)
Subscribe(sensor1, 5) timeCorrelated10_1 Subscrib
e(sensor2, 6) Subscribe(sensor3,
20) timeCorrelated60_1 timeCorrelated30_1
6 XML
12 XML
30
Paper work 2
Using countCorrelated_1

• Count based correlate

Subscribe(sensor1) Subscribe(sensor2)
Subscribe(sensor1, 5) Subscribe(sensor2, 6)
Subscribe(sensor1) Subscribe(sensor2)
Subscribe(sensor1) Subscribe(sensor2)
Subscribe(sensor1) Sunscribe(sensor4) timeCorrel
ated10_1 Subscribe(sensor2) Subscribe(sensor1)
Subscribe(sensor3, 20) Subscribe(sensor2, 15)
Subscribe(sensor4) Subscribe(sensor5)
Subscribe(sensor1, 5) Subscribe(sensor5)
Subscribe(sensor6) Subscribe(sensor2, 6)
Subscribe(sensor1) Subscribe(sensor2)
Subscribe(sensor1, 5) Subscribe(sensor2, 6)
Subscribe(sensor1) Subscribe(sensor2)
Subscribe(sensor1) Subscribe(sensor2)
Subscribe(sensor1) Sunscribe(sensor4) timeCorrel
ated10_1 Subscribe(sensor2) Subscribe(sensor1)
Subscribe(sensor3, 20) Subscribe(sensor2, 15)
Subscribe(sensor4) Subscribe(sensor5)
Subscribe(sensor1, 5) Subscribe(sensor5)
Subscribe(sensor6) Subscribe(sensor2, 6)
countCorrelated_1 countCorrelated_1
countCorrelated_1 countCorrelated_1
countCorrelated_1 countCorrelated_1
Sunscribe(sensor4) timeCorrelated10_1
countCorrelated_1 Subscribe(sensor3, 20)
countCorrelated_1 Subscribe(sensor4)
Subscribe(sensor5) countCorrelated_1
Subscribe(sensor5) Subscribe(sensor6) countCorre
lated_1
Count sensor1 7 sensor2 7 sensor4 2 sensor5
2 sensor3 1 sensor6 1 timeCorrelated10_1 1
countCorrelated_1
17 XML
21 XML
31
Paper work 2

• User based correlate

Subscribe(sensor1) middleware_1
Subscribe(sensor2) middleware_1
Subscribe(sensor5, 5) middleware_1
Subscribe(sensor1) middleware_1
Subscribe(sensor2) middleware_1
Subscribe(sensor1) middleware_1
Subscribe(sensor1) middleware_1
Subscribe(sensor3, 20) middleware_1
Subscribe(sensor4) middleware_1
Subscribe(sensor2, 6) middleware_1 Sunscribe(se
nsor4) application_1 timeCorrelated10_1 applicat
ion_1 Subscribe(sensor5, 5) application_1
timeCorrelated10_1 application_1
Subscribe(sensor5, 5) application_1 Subscribe(sen
sor1) middleware_2 Subscribe(sensor2)
middleware_2 Subscribe(sensor5, 10)
middleware_2 Subscribe(sensor6, 10)
middleware_2 Subscribe(sensor1) middleware_4
Subscribe(sensor1) middleware_1
Subscribe(sensor2) middleware_1
Subscribe(sensor1) middleware_2
Subscribe(sensor5, 5) middleware_1
Subscribe(sensor1) middleware_1
Subscribe(sensor2) middleware_1
Subscribe(sensor1) middleware_4
Subscribe(sensor2) middleware_2
Subscribe(sensor1) middleware_1
Sunscribe(sensor4) application_1
timeCorrelated10_1 application_1
Subscribe(sensor1) middleware_1
Subscribe(sensor5, 5) application_1
Subscribe(sensor3, 20) middleware_1
Subscribe(sensor4) middleware_1
Subscribe(sensor5, 10) middleware_2
Subscribe(sensor6, 10) middleware_2
timeCorrelated10_1 application_1
Subscribe(sensor5, 5) application_1 Subscribe(sen
sor2, 6) middleware_1
Local data ???? ??
Client? ??
32
Paper work 2
??? correlated data Local data sensor1?
sensor2 countCorrelated_1 middleware_1 Remote
data (sensor5,10) ? (sensor6,
10) timeCorrelated10_2 middleware_2
Subscribe(sensor1) middleware_1
Subscribe(sensor2) middleware_1
Subscribe(sensor1) middleware_2
Subscribe(sensor5, 5) middleware_1
Subscribe(sensor1) middleware_1
Subscribe(sensor2) middleware_1
Subscribe(sensor1) middleware_4
Subscribe(sensor2) middleware_2
Subscribe(sensor1) middleware_1
Sunscribe(sensor4) application_1
timeCorrelated10_1 application_1
Subscribe(sensor1) middleware_1
Subscribe(sensor5, 5) application_1
Subscribe(sensor3, 20) middleware_1
Subscribe(sensor4) middleware_1
Subscribe(sensor5, 10) middleware_2
Subscribe(sensor6, 10) middleware_2
timeCorrelated10_1 application_1
Subscribe(sensor5, 5) application_1 Subscribe(sen
sor2, 6) middleware_1
countCorrelated_1 middleware_1 Subscribe(sensor1)
middleware_2 Subscribe(sensor5, 5)
middleware_1 countCorrelated_1 middleware_1 Subs
cribe(sensor1) middleware_4 Subscribe(sensor2)
middleware_2 countCorrelated_1 middleware_1 Suns
cribe(sensor4) application_1 timeCorrelated10_1
application_1 countCorrelated_1 middleware_1 Subs
cribe(sensor5, 5) application_1
Subscribe(sensor3, 20) middleware_1
Subscribe(sensor4) middleware_1
timeCorrelated10_2 middleware_2 timeCorrelated10
_1 application_1 Subscribe(sensor5,
5) application_1 countCorrelated_1 middleware_1
20 XML
17 XML
33
Paper work 3

• ????? ??? ??? ??
• ???? ???
• ?? ??(XML)? ??? peer-to-peer??
• ???? ?? ??? ??
• Application? ???? data? ?? middleware? ?? ??
• Device data ? ???? ?? middleware??
• Remote middleware? data? ??

34
Paper work 3

• ???
• XML? ???? remote data? ??
• ??? ? ?? ??? ???
• user(application, middleware)? ??
• ??? ?? middleware
• user? ????? (XML) ?? ??

35
Paper work 3

• XML ?? (deviceData)
• ???? ???? ?? ??

// ???? ?? ?? ?? lt?xml version"1.0"
encoding"UTF-8" ?gt ltsensorDataListgt ltsensorData
gt ltdevice nameMiddleware2gt ltnetworkIDgtnetwor
kID(remote middlewareID)lt/networkIDgt ltsensorIDgt
lt/sensorIDgt ltdeviceTypegtMiddlewarelt/deviceTypegt
lt/devicegt ltdatagt lttypegtlt/typegt ltformat
gtlt/formatgt ltvaluegtlt/valuegt lttimestampgt000000
00lt/timestampgt lt/datagt ltoptionsgt ltdeviceLi
stgt ltdevicegt ltnetworkIDgtnetworkIDlt/networ
kIDgt ltsensorIDgtsensorIDlt/sensorIDgt ltdevi
ceNamegtReader100lt/deviceNamegt ltdataTypegtEPClt/
dataTypegt lt/devicegt . . .
ltdevicegt ltnetworkIDgtnetworkIDlt/networkIDgt
ltsensorIDgtsensorIDlt/sensorIDgt ltdeviceNam
egtTemperatureSensor5lt/DeviceNamegt ltdataTypegtT
emperaturelt/dataTypegt lt/devicegt lt/deviceLis
tgt lt/optionsgt lt/sensorDatagt ltsensorDataListgt
36
Paper work 3

• XML ?? (request Data)
• ?? ??? ?? data

// remote sensor ??? ?? lt?xml version"1.0"
encoding"UTF-8" ?gt ltsensorDataListgt ltsensorData
gt ltdevice nameMiddleware1gt ltnetworkIDgtnetwor
kID(local middlewareID)lt/networkIDgt ltsensorIDgtlt
/sensorIDgt ltdeviceTypegtMiddlewarelt/deviceTypegt
lt/devicegt ltdatagt lttypegtlt/typegt ltformatgt
lt/formatgt ltvaluegtlt/valuegt lttimestampgt0000000
0lt/timestampgt lt/datagt ltoptionsgt ltrequestgt
ltdevicegt ltnetworkIDgtnetworkIDlt/networkIDgt
ltsensorIDgtsensorIDlt/sensorIDgt ltdeviceNa
megtReader100lt/deviceNamegt ltdataTypegtEPClt/data
Typegt lt/devicegt ltusergt ltformatgtIPv4lt/
formatgt ltvaluegtxxx.xxx.xxx.xxxlt/valuegt lt/
usergt lt/requestgt lt/optionsgt lt/sensorDatagt lt
sensorDataListgt
37
Paper work 3

• XML ?? (request Data)
• ?? data

// remote sensor ??? ?? lt?xml version"1.0"
encoding"UTF-8" ?gt ltsensorDataListgt ltsensorData
gt ltdevice nameMiddleware1gt ltnetworkIDgtnetwor
kID(local middlewareID)lt/networkIDgt ltsensorIDgtlt
/sensorIDgt ltdeviceTypegtMiddlewarelt/deviceTypegt
lt/devicegt ltdatagt lttypegtlt/typegt ltformatgt
lt/formatgt ltvaluegtlt/valuegt lttimestampgt0000000
0lt/timestampgt lt/datagt ltoptionsgt ltrequestgt
ltdatagt ltformatgtEPClt/formatgt ltvaluegt0
00000000000610329037666lt/valuegt lt/datagt ltu
sergt ltformatgtIPv4lt/formatgt ltvaluegtxxx.xx
x.xxx.xxxlt/valuegt lt/usergt lt/requestgt lt/o
ptionsgt lt/sensorDatagt ltsensorDataListgt