Docstoc

Support for context-driven applications in Ambient Intelligence

Document Sample
Support for context-driven applications in Ambient Intelligence Powered By Docstoc
					Support for context-driven applications in
Ambient Intelligence environments
Davy Preuveneers               PhD Defense




                                                 Promotor:
                                   Prof. Dr. ir. Y. Berbers


                                               July 1, 2009
                Context-driven applications
                • Applications
                  – That are aware of my current situation and context
Introduction      – That autonomously adapt their behavior accordingly
Middleware      • Using context
Contributions     –   Who is calling me?
Evaluation        –   What time is it?
Conclusions       –   Where am I?
                  –   What am I doing?

                • Adapt behavior
                  – Ring or vibrate
                  – Redirect to other phone
                  – Leave message



                                                                         2
                Ambient Intelligence
                • Personalized digital environments
                • Sensitive and responsive to presence of people
Introduction
Middleware
Contributions
Evaluation
Conclusions




                                                                   3
                Ambient Intelligence

                Goals of Ambient Intelligence (AmI):
Introduction    •    Devices work in concert
Middleware      •    Support people with everyday activities
Contributions   •    In a natural way
Evaluation      •    Using information and intelligence
Conclusions     •    Hidden in the network connecting these devices


                                    Ambient Intelligence


                    Computing and       Intelligent         Context
                    Communication       Interfaces         Awareness

                                                                       4
Overview
•   Introduction
•   Middleware for context-driven applications
•   Contributions
      1.   Context modeling and reasoning
      2.   Middleware for context management
      3.   Context-driven application adaptation
      4.   Context-aware application collaboration
•   Evaluation
•   Conclusions and future work
                                                     5
                Context-driven applications


Introduction         Examples           Context Types   Human Concern

Middleware
                  Auto Cell Phone
                                                         Convenience
Contributions      Off In Meetings
                                           Identity
Evaluation
                     Tag Photos             Time          Finding Info
Conclusions
                                          Location
                Calendar Reminders        Proximity        Memory
                                           Activity
                     Health Alert          History          Safety

                                             …
                 Auto Lights On / Off                      Efficiency



                                                                         6
                Problem statement

                • Developing context-driven applications
                  is hard and complex
Introduction
                  – Availability of context sources
Middleware
Contributions
Evaluation
Conclusions




                                                       7
                Problem statement

                • Developing context-driven applications
                  is hard and complex
Introduction
                  – Availability of context sources
Middleware
                  – Deduction of high-level from low-level context
Contributions
Evaluation
Conclusions



                              Look up cell tower ID
                              or IP address




                                                                     8
                Problem statement

                • Developing context-driven applications
                  is hard and complex
Introduction
                  – Availability of context sources
Middleware
                  – Deduction of high-level from low-level context
Contributions
Evaluation
                  – Adaptation of applications to changing context
Conclusions



                             Look up cell tower ID
                             or IP address




                                                                 9
                Motivating scenario: E-health

                • Diabetes self-care:
                  – Eat the right foods
Introduction
                  – Get and log daily physical activity
Middleware
Contributions     – Take medications as prescribed
Evaluation        – Test blood glucose regularly
Conclusions




                                                          10
                Motivating scenario: E-health

                • Diabetes self-care with mobile phones:
                   – People forget or don’t keep a detailed log
Introduction
                   – Recalling similar previous situations becomes difficult
Middleware
Contributions      – Create a context-driven recommender application
Evaluation
Conclusions     • Benefits of the application
                   – Time and location monitoring
                   – User input on food consumption and insulin dosage
                   – Find correlations between time/location and activities
                   – Augment blood glucose level logs with contextual data
                   – Use context to find similar situations in the past

                                                                               11
Overview
•   Introduction
•   Middleware for context-driven applications
•   Contributions
      1.   Context modeling and reasoning
      2.   Middleware for context management
      3.   Context-driven application adaptation
      4.   Context-aware application collaboration
•   Evaluation
•   Conclusions and future work
                                                     12
                A middleware approach

                               Applications
Introduction
Middleware                      Middleware

Contributions
                             Operating System
Evaluation
Conclusions
                Benefits:
                •   Hide context sensor heterogeneity
                •   Manage context efficiently
                •   Simplify adaptation of mobile applications
                •   Context-aware discovery and collaboration


                                                                 13
                A middleware approach

                Context
                •   Current time, location, activity, preferences
Introduction
Middleware
                •   People, services and devices around me
Contributions   •   …
Evaluation
Conclusions         DEFINITION
                    “Any information that can be used to characterize the
                    situation of an entitiy, where an entity can be a person,
                    place or physical or computational object.”,
                    Dey at al. (2001)




                                                                                14
                A middleware approach

                Middleware requirements:
Introduction    1. Implement a semantically expressive and
Middleware         non-ambiguous context representation
Contributions
Evaluation
Conclusions




                                                             15
                A middleware approach

                Middleware requirements:
Introduction    1. Implement a semantically expressive and
Middleware         non-ambiguous context representation
Contributions   2. Deployment on resource constrained hardware
Evaluation         in mobile ad hoc networks
Conclusions




                • PXA271 XScale 13-416 MHz
                • 256kB SRAM, 32 MB SDRAM
                • 32MB Flash
                                                           16
                A middleware approach

                Middleware requirements:
Introduction    3. Support application adaptation at runtime
Middleware
Contributions
Evaluation
Conclusions




                • PXA271 XScale 13-416 MHz
                • 256kB SRAM, 32 MB SDRAM
                • 32MB Flash
                                                               17
                A middleware approach

                Middleware requirements:
Introduction    3. Support application adaptation at runtime
Middleware      4. Context-aware interoperability in dynamic and
Contributions      mobile device and service networks
Evaluation
Conclusions




                ISTAG Scenarios of Ambient Intelligence in 2010; Final Report, Feb. 2001.
                                                                                            18
Overview
•   Introduction
•   Middleware for context-driven applications
•   Contributions
      1.   Context modeling and reasoning
      2.   Middleware for context management
      3.   Context-driven application adaptation
      4.   Context-aware application collaboration
•   Evaluation
•   Conclusions and future work
                                                     19
                   Contributions

                      Context modeling and reasoning
Introduction
Middleware
                           Context management
Contributions
1. Modeling
2. Management
3. Adaptation            Context-driven adaptation
4. Collaboration

Evaluation
Conclusions            Context-driven collaboration



                                                       20
                    1. Context modeling and reasoning

                   CoDAMoS context ontology in OWL
Introduction
                   Major concepts
Middleware            User
Contributions         Environment
                      Platform
1. Modeling
                      Services
2. Management
3. Adaptation      Relevance and
4. Collaboration   Ambiguity
Evaluation            Accuracy
                      Precision
Conclusions           Freshness
                      Spatial coverage
                      Sampling rate
                      …

                                                        21
                    1. Context modeling and reasoning
                    Semantic                            Location
                                                                         is a


Introduction                 Building          Passage                  Room                    Floor
                                                                                 is a
Middleware
Contributions                                       Auditorium         Kitchen          MeetingRoom
1. Modeling
                                         instance
2. Management
                                          00.255               interpreted
3. Adaptation                                                                           Graph

4. Collaboration
                                Shape                                                                    has
                                                                                        has
Evaluation            is a

Conclusions                                                                     Node             Edge
                     Circle             Rectangle        has
                                                                       instance                    has
                                                        shape
                                         instance
                   Quantitative                                                                 Qualitative
                                                Rect_1                       Node_1
                                                                                                           22
                       1. Context modeling and reasoning

                   Ontologies in diabetes use case:
                   • `Upper’ CoDAMoS context ontology
Introduction
                        –   Common general purpose concepts
Middleware
                        –   E.g. time, location, user preferences (base insulin plan)
Contributions
1. Modeling        •     Domain specific ontologies
2. Management           –   Various user activities and tasks
3. Adaptation               • Classified according to exercise level (none, light,
4. Collaboration
                                moderate and high)
                            • E.g. Sleeping, hobbies, profession, sports, …
Evaluation
                        –   Food and nutritional information
Conclusions
                            • Convert amount of food into numeric values
                            • E.g. carbohydrates, proteins, calories, fat, …



                                                                                  23
                       1. Context modeling and reasoning

                   Efficient semantic matching algorithms
                   • Semantic matching with subsumption or
Introduction
                       inheritance relationships
                        –   “Is one concept a subclass of another concept?”
Middleware
                   •     Slow matching for large ontologies
Contributions
1. Modeling
2. Management
3. Adaptation                       Shape

4. Collaboration                  is a

Evaluation
Conclusions              Circle          Rectangle



                             3 classes
                                                           160 classes        24
                    1. Context modeling and reasoning

                   Efficient semantic matching algorithms
                   • Compile ontologies ahead-of-time
Introduction
                   • Incremental encoding, partial loading
Middleware
                   • Orders of magnitude better (cpu/memory) than
Contributions
                       standard ontology reasoners
1. Modeling
                                                    Room                      hasProjector
2. Management
                                                    γ(A): 2                   γ(D): 7
3. Adaptation                                                     A       D
4. Collaboration

Evaluation           Meeting Room                   Lab                   Auditorium
                                      F                       B       E
                       γ(F): 2×13             γ(B): 2×3                   γ(E): 2×7×11
Conclusions
                   Video Conference
                                      G   Computer Lab        C
                     γ(G): 2×13×17          γ(C): 2×3×5



                                                                                         25
                   1. Context modeling and reasoning

                               Binary   HashSet               Prime     Prime
                    Ontology                        Krall
                               Matrix    (avg)                (max)     (avg)
Introduction
                     SUMO       630        244       30        83        42
Middleware
                    Wine &
Contributions                   133        188       33        53        23
                     Food
1. Modeling
                     Pizza       99        194       37        40        23
2. Management
3. Adaptation        Gene
                               20945       301       151       361       82
4. Collaboration    Ontology

Evaluation          Java 1.3    5438       140       68        112       31
Conclusions
                    OpenCyc    25565      1166       350       681       272



                   Compared with other encodings in terms of bits per class
                                                                                26
                   Contributions

                      Context modeling and reasoning
Introduction
Middleware
                           Context management
Contributions
1. Modeling
2. Management
3. Adaptation            Context-driven adaptation
4. Collaboration

Evaluation
Conclusions            Context-driven collaboration



                                                       27
                   2. Middleware for context management

                   Main concerns (vs. other systems):
Introduction
                   • Manage context w/o wasting resources
Middleware         • Keep overhead for applications low
Contributions                                    END USER
                                                 APPLICATIONS       APPLICATION
1. Modeling                                                         ADAPTATION
                                USER PROFILE
2. Management
3. Adaptation
4. Collaboration                                                    TRANSFORM
                        ACQUISITION            STORAGE                  &
Evaluation                                                          INFERENCE

Conclusions                       CONTEXT MANAGING COMPONENTS


                     DATABASE                                             POLICY
                                                          BILLING
                                                                       ENFORCEMENT

                                SENSORS                  OTHER ENABLING SERVICES

                                                                                   28
                   2. Middleware for context management

                   Basic design principles:
Introduction
                   • Input: facade pattern
Middleware         • Transformation: pipe and filter
Contributions        – From: CID = 283F and LAC = 3070
1. Modeling          – To: Lat = 50.8717 and Lon = 4.6875
2. Management
                     – To: Current weather
3. Adaptation
4. Collaboration              Pipe                      Pipe                         Pipe
                                                                                T
Evaluation           GSM                  Filter                      Filter
Conclusions         Network               [infer]                     [infer]



                              CID / LAC             Geo-coordinates                 Weather



                                                                                         29
                   2. Middleware for context management

                   Basic design principles:
Introduction
                   • Input: facade pattern
Middleware         • Transformation: pipe and filter
Contributions      • Storage and pub/sub: blackboard
1. Modeling
                   • Reason engines: chain of responsibility
2. Management
3. Adaptation
                              Reasoning problem
4. Collaboration
                                              Abstract    Spatial
                       App.
Evaluation                                    Handler    Reasoner
Conclusions
                                                         Temporal
                                                         Reasoner

                                            Blackboard   Semantic
                                                         Reasoner
                                                                    30
                       2. Middleware for context management

                   Managing context in diabetes use case:
                   •     Food consumption
Introduction
                   •     Location and activity awareness
Middleware
Contributions      Food consumption
1. Modeling
                   •     Convert user input on meal into numeric values
2. Management
                         with nutrition database component
3. Adaptation
4. Collaboration   •     E.g.: 100 gr French fries
Evaluation              •   38 gram carbohydrates
                                                         Nutrition
                        •   40 gram proteins
Conclusions                                               Table
                        •   456 kilocalories
                        •   16 gram fat          Food                Carbs, Fat
                                              Quantity               Calories, …


                                                                              31
                       2. Middleware for context management

                   Location awareness using cell towers
                   •     Graph of time-stamped cell tower transitions
Introduction
Middleware
Contributions
1. Modeling
2. Management
3. Adaptation
4. Collaboration

Evaluation
Conclusions




                                                                        32
                       2. Middleware for context management

                   Location awareness using cell towers
                   •     Cell tower handover transition probability
Introduction
                   •     Markov chain of first order
Middleware
Contributions
1. Modeling
2. Management
3. Adaptation
4. Collaboration

Evaluation
Conclusions




                                                                      33
                       2. Middleware for context management

                   Location awareness using cell towers
                   •     Handovers depend on current Sj and predecessor state Si
Introduction
Middleware
Contributions
1. Modeling        •     Total duration of connectivity to tower is used for
2. Management            probability of being in that state
3. Adaptation
4. Collaboration

Evaluation         •     Compute probability of connecting to a sequence of three
                         cell towers: O = S2, S4, S7
Conclusions




                                                                                34
                       2. Middleware for context management

                   Activity awareness based on location
                   •     Hard to recognize without additional sensors
Introduction
                   •     Only interested in the exercise level of activity
Middleware
                        •   None: e.g. sleeping
Contributions
                        •   Light: e.g. socializing, hobbies, watching TV
1. Modeling
                        •   Moderate: e.g. profession
2. Management
                        •   High: e.g. profession, sport
3. Adaptation
4. Collaboration   •     Use Hidden Markov Model
Evaluation              •   4 exercise levels as hidden states
Conclusions             •   Discretization of blood glucose as 15 observable states
                        •   Correlate activities with locations




                                                                                35
                   Contributions

                      Context modeling and reasoning
Introduction
Middleware
                           Context management
Contributions
1. Modeling
2. Management
3. Adaptation            Context-driven adaptation
4. Collaboration

Evaluation
Conclusions            Context-driven collaboration



                                                       36
                   Motivating scenario: E-health
                   • Component-based e-diary application
                   • (Non-) functional properties of each component
Introduction
                     are syntactically and semantically specified
Middleware
Contributions
1. Modeling
2. Management
3. Adaptation
4. Collaboration

Evaluation
Conclusions




                                                                      37
                       3. Context-driven application adaptation

                   Adapt GUI component to resources
                   •     Screen size, memory, CPU, …
Introduction
                   •     Graphics support (e.g. virtual machine)
Middleware
Contributions
1. Modeling
2. Management
3. Adaptation
4. Collaboration

Evaluation
Conclusions



                                  Graphical report on bigger screen
                                  Simple text report on small display
                                                                        38
                   3. Context-driven application adaptation

                   1. Collect context information
                     – Hardware, resources, application
Introduction
                       requirements, dependencies
Middleware
Contributions      2. Find deployment
1. Modeling          – Component selection
2. Management
                     – Constraint solving and optimization
3. Adaptation
4. Collaboration   3. Adaptation at runtime
Evaluation           –   Add, remove, replace or relocate components
Conclusions
                     –   Change component configuration
                     –   Behavior defined by policies
                     –   Detect and resolve adaptation conflicts
                                                                  39
                       3. Context-driven application adaptation

                   Application adaptation policies:
                   •     Declarative event-condition-action rules that
Introduction
                         govern application behavior
Middleware
                        •   If (battery.load < 30 %)
Contributions
                              Then disableComponent(‘BluetoothDiscovery’)
1. Modeling
                        •   If (battery.load < 15 %)
2. Management
                              Then invoke(‘UserInterface.warning()’)
3. Adaptation
4. Collaboration   •     Application specific behavior:
Evaluation              •   Normal blood glucose: 80 – 150 mg/dL
Conclusions             •   < 50 mg/dL: hypoglycemic diabetic coma !!!
                        •   > 500 mg/dL: hyperglycemic diabetic coma !!!



                                                                            40
                   Contributions

                      Context modeling and reasoning
Introduction
Middleware
                           Context management
Contributions
1. Modeling
2. Management
3. Adaptation            Context-driven adaptation
4. Collaboration

Evaluation
Conclusions            Context-driven collaboration



                                                       41
Overview
•   Introduction
•   Middleware for context-driven applications
•   Contributions
      1.   Context modeling and reasoning
      2.   Middleware for context management
      3.   Context-driven application adaptation
      4.   Context-aware application collaboration
•   Evaluation
•   Conclusions and future work
                                                     42
                 Evaluation – Case studies
                Middleware for context-aware applications

Introduction
Middleware
Contributions
Evaluation
Conclusions




                Context-aware adaptation      Diabetes health-care
                of Instant Messaging client   recommender application


                IWT SBO CoDAMoS
                                                                        43
                Evaluation – Case studies
                Middleware support as a service

Introduction
Middleware
Contributions
Evaluation
Conclusions




                 ITEA LOMS

                                                  44
                Evaluation - Tools
                Modeling and simulation support

Introduction
Middleware
Contributions
Evaluation
Conclusions




                                                  45
                Evaluation - Publications
                > 20 international publications
                •   Towards an extensible context ontology for Ambient Intelligence
                •   Encoding semantic awareness in resource-constrained devices
Introduction    •   Automated context-driven composition of pervasive services to
Middleware          alleviate non-functional concerns
                •   Adaptive context management using a component based
Contributions
                    approach
Evaluation
Conclusions




                                                                                 46
Overview
•   Introduction
•   Middleware for context-driven applications
•   Contributions
      1.   Context modeling and reasoning
      2.   Middleware for context management
      3.   Context-driven application adaptation
      4.   Context-aware application collaboration
•   Evaluation
•   Conclusions and future work
                                                     47
                Lessons learned

                •   Model context with ontologies
                    –   Despite performance issue
Challenges          –   Standardization (W3C Working Draft 2009)
Contributions
Publications    •   Component-based design is key
Evaluation
                    –   Syntactically and semantically described black boxes
                    –   Simplify local application adaptation at runtime
Conclusions
                    –   Optimize workflow in context middleware

                •   Our approach is not the holy grail
                    –   Fitness functions, evolutionary and genetic algorithms,
                        distance metrics, learning, mixed-initiative, …
                    –   Trends towards software as a service, mash-ups, …
                    –   Context-oriented programming (COP), languages, …


                                                                             48
                Conclusive remarks
                Evaluation and validation in various projects:
                • IWT SBO CoDAMoS
Challenges
                • IBBT CROSLOCiS
Contributions
                • ITEA LOMS
Publications
Evaluation
Conclusions




                                                                 49
                Future work

                •   Better algorithms and mechanisms for
                    –   Enhanced matching for hybrid context models
Challenges          –   Predicting future contexts to drive proactive behavior
Contributions
                •   Comparing and measuring quality
Publications
                    –   Quality of Context (QoC) / Quality of Information (QoI)
Evaluation          –   Quality of Service (QoS)
Conclusions         –   Quality of Experience (QoE)
                •   Automated testing framework
                    –   Simulate various situations
                    –   Unit testing for adaptive applications
                •   Scalable (horizontal) collaboration
                    –   Interactions in large scale deployments
                    –   Delocalized sensing, decision making, and adaptation

                                                                             50
The end


          51
                Backup slides


Challenges
Contributions
Publications
Evaluation
Conclusions




                                52
                Introduction

                Research challenges of AmI:
Introduction    • Devices disappear from the conscious
Middleware        attention of the user
Contributions
                • Devices communicate
Evaluation
                  – more between one another
Conclusions
                  – more with the environment
                  – less with the user
                 Towards implicit interaction using context
                 information obtained from
                  – other devices
                  – the environment
                  – the user if necessary
                                                          53
                Context-driven applications


                 Existing Examples       Context Types      Human Concern
Introduction
Middleware
                 Auto Lights On / Off    Room Activity       Convenience
Contributions
Evaluation
                Calendar Reminders            Time             Memory
Conclusions
                                        Personal Identity
                    File Systems                              Finding Info
                                            & Time

                    Smoke Alarm          Room Activity          Safety


                 Barcode Scanners        Object Identity       Efficiency



                                                                             54
                Motivating scenario: E-health

                • Replacing diary with mobile phone:
Introduction
Middleware
Contributions
Evaluation
Conclusions




                • Context taken into consideration:
                  – Time, location, activity, history
                  – Blood glucose levels, food intake, …
                                                           55
                       1. Context modeling and reasoning

                   How is our approach different/better?
                   •     Contextual semantics are not enough.
Introduction       •     Need to handle relevance and ambiguity!
Middleware
Contributions
                   Add quality of context information:
1. Modeling
                   •     accuracy
2. Management
3. Adaptation
                   •     precision
4. Collaboration
                   •     freshness
Evaluation         •     sample rate
Conclusions
                   •     min/max values
                   •     spatial coverage
                   •     semantic interpretability
                   •     …
                                                                   56
                   4. Context-aware application collaboration

                   Collaborate if local adaptation not possible
                   •   Trigger to adapt arrives at bad time
Introduction
                   •   Not enough/right resources
Middleware
Contributions      Managing context in diabetes use case
1. Modeling
                   •   Display report on device with bigger screen
2. Management
3. Adaptation
                   •   Sync with bluetooth glucose meter
4. Collaboration
                   Use protocols to discover alternatives
Evaluation
                   1. Enhance service discovery protocol with
Conclusions
                      context-awareness
                   2. Device announces services and context
                   3. Protocols replicate service state
                                                                     57
                   4. Context-aware application collaboration

                   Replicate services
Introduction
                   • Reduce effects of disconnections
Middleware         • Enhance interaction modalities
Contributions
1. Modeling
2. Management
3. Adaptation
                       SERVICE MIGRATION
4. Collaboration

Evaluation
Conclusions                                         STATE
                                               SYNCHRONIZATION


                       SERVICE DIFFUSION



                                                                 58
                   4. Context-aware application collaboration

                   Replicate services
                   •   Handover is quicker than local adaptation
Introduction
Middleware
Contributions
1. Modeling
2. Management
3. Adaptation
4. Collaboration

Evaluation
Conclusions




                                                                   59
                Evaluation - Publications
                1.       Context modeling and reasoning
                     •      Towards an extensible context ontology for Ambient Intelligence
                     •      Encoding semantic awareness in resource constrained devices
Challenges
                2.       Middleware for context management
Contributions        •      Adaptive context management using a component based approach
Publications    3.       Context-aware application adaptation
Evaluation           •      Automated context-driven composition of pervasive services
                     •      Multi-dimensional dependency and conflict resolution for self-
Conclusions                 adaptable context-aware systems
                4.       Context-aware application collaboration
                     •      Context-driven migration and diffusion of pervasive services on the
                            OSGi framework
                5.       Case studies
                     •      Context-aware adaptation for component-based pervasive
                            computing systems
                     •      Mobile phones assisting with health self-care: a diabetes case study



                                                                                              60
                Conclusive remarks
                1. Model context and reasoning
                    OWL ontologies with support for ambiguity
Challenges
                    Algorithms for efficient semantic matching
Contributions
                2. Manage context
Publications
                    Adaptive middleware not wasting resources
Evaluation
                    Vertical scalability
Conclusions
                3. Use context for adaptation
                    Context-aware adaptation of component-
                     based applications and services
                    Mechanisms to detect adaptation conflicts
                4. Context-driven collaboration
                    Enhanced service discovery protocols
                    Replicating services for better experience
                                                                  61
• Service Discovery Protocols (SDPs)
   –   Jini             (Sun Microsystems)
   –   UPnP             (Microsoft)
   –   Salutation       (Salutation Consortium)
   –   SLP              (IETF standard: RFC2608, RFC2609, RFC3059)
   –   Bluetooth SDP    (Bluetooth Special Interest Group)

• Characteristics:
   –   Configuration
   –   Communication
   –   Service descriptions
   –   Discovery
   –   …

                                                                 62
   Service Provider




Service Deployment
Infrastructure (SDI)




Service Users

                       CoDAMoS Service Platform   63
   Service Provider
                                                                  Comp. Library   Compositions




Service Deployment
                                               Service Provider
Infrastructure (SDI)
                                                     SDI
                       Context Specification




Service Users

                               CoDAMoS Service Platform                                          64
   Service Provider
                                                                     Comp. Library   Compositions
                                                    Rs
                       Rs  Rh



                Rh


                           Rh, User profile, Location, …


Service Deployment
                                                  Service Provider
Infrastructure (SDI)
                                                        SDI
                         Context Specification




Service Users

                                  CoDAMoS Service Platform                                          65
   Service Provider
                                                                          Comp. Library          Compositions
                                                    Rs
                       Rs  Rh



                Rh
                                                                           Service S1

                           Rh, User profile, Location, …


Service Deployment
                                                  Service Provider
Infrastructure (SDI)
                                                        SDI
                         Context Specification                       Generated Implementation




                                                                                        Service S1’’




Service Users

                                  CoDAMoS Service Platform                                                      66
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’
                                                                                                             Service S4




                                                              Users
Service Users

                                             CoDAMoS Service Platform                                                      67
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’
                                                                                                             Service S4




                  UI: speech                                                                       UI: graphical, speech




                                                              Users
Service Users

                                             CoDAMoS Service Platform                                                      68
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’       Cooperation
                                                             Service S4                                      Service S4




                  UI: speech                      Context awareness/discovery                      UI: graphical, speech




                                                              Users
Service Users

                                             CoDAMoS Service Platform                                                      69
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’
                                                                                                             Service S4

                                                                                   Mobility




                                                              Users
Service Users

                                             CoDAMoS Service Platform                                                      70
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’
                                                                                                             Service S4

                                                                                   Mobility




                                                              Users
Service Users

                                             CoDAMoS Service Platform                                                      71
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’
                                                             Service S4                                      Service S4

                                                                                   Mobility




                                                              Users
Service Users

                                             CoDAMoS Service Platform                                                      72
   Service Provider
                                                                                      Comp. Library         Compositions
                                                               Rs
                                Rs  Rh



                   Rh
                                                                                       Service S

                                      Rh, User profile, Location, …


Service Deployment
                                                             Service Provider
Infrastructure (SDI)
                                                                   SDI
                                    Context Specification                       Generated Implementations



                Service-User SDI                           Service-User SDI                           Service-User SDI

                 Device resources
                                                                                                   Service S1’’
                                                        Service S3
           Service S1’
                                                                                                        Service S2’’
                  Service S2’       Cooperation
                                                             Service S4                                      Service S4

                                                                                   Mobility


                  UI: speech                      Context awareness/discovery                      UI: graphical, speech




                                                                Users
Service Users

                                             CoDAMoS Service Platform                                                      73
                              user
            usesService*

                                 usesPlatform*



          providesService*              hasEnvironment
service                      platform                    environment




                                                                       74
                                            preference
                                              profile
   • User
                                                 isa

                               hasProfile                hasProfile

activity             role                     profile                 mood

                                                         hasProfile
      hasActivity*
                            hasRole*

 task                                                                 user
                               hasTask*

    usesService*                                          usesIODevice*
                                              i/o
                     service
                                            device
                                                                             75
                                                                                          power

        • Platform                                                                       resource

                                                                                         memory
                        rendering operating     virtual      middle-                     resource
supportsModality*         engine   system      machine        ware
                                                                                           cpu
                                                                                         resource
             modality
                                   isa                                 resource           storage
                                                                isa               isa    resource
             providesService*
service                         software         hardware                                 network
                                                                                         resource

         requiresPlatform*               providesSoftware*
                                                                                   isa      input
                                                                          i/o
                                platform                                                   device
                                                providesHardware*       device
                    hasEnvironment
                                                              usesIODevice*
                                                                                           output
                             environment                                 user              device
                                                                                                    76
• Service
       task                                           software


              usesService*                providesService*
                             service

hasServiceProfile            hasServiceModel        hasServiceGrounding


       service               service                   service
       profile               model                    grounding

                                                                          77
        • Environment
                                               platform
                                                hasEnvironment

                                     environment

                      hasLocation*        hasTime*           hasEnvironmentalCondition*

isRelativeTo*
                      location          time                 environmental
                            isa                                condition
                                                                         isa

         relative         absolute
                                               temperature        lighting              noise
                isa
         address                                       pressure              humidity
                                                                                                78
• Required Resources
  – Specify minimum required memory:

     <profile:serviceParameter>
      <compprofile:RequiredResources
       rdf:ID="requiredMemory">
       <profile:sParameter>
         <context:Memory>
          <context:bytes>1048576</context:bytes>
         </context:Memory>
       </profile:sParameter>
      </compprofile:RequiredResources>
     </profile:serviceParameter>

                                                   79
•   User preferences for the communication service:
<context:UserPreference rdf:ID="VideoEncoding1">
 <context:prefProperty rdf:resource="#VideoQuality"/>
 <context:prefValue rdf:resource="&concepts;#High"/>
 <context:prefCondition>
  <context:Location>
   <context:where rdf:resource="&concepts;#Office"/>
  </context:Location>
 </context:prefCondition>
 <context:prefCondition>
  <context:Time>
   <context:when rdf:resource="&concepts;#NineToFive"/>
  </context:Time>
 </context:prefCondition>
</context:UserPreference>                                 80
81
82
83
84
85
•   PETs:
    – Anonymous credential systems:
       • User receives pseudonymously non-transferable and revocable credential from
         issuer
       • Different shows to verifier are not linkable
       • Selectively disclose attributes to the verifier
       • Based on pseudonyms, blind signatures and zero-knowledge proofs
    – Zero-knowledge proof: Ali Baba’s Cave                                            P
       • Alice wants to prove to bob that she knows how
          to open the secret door between R and S.
       • Bob goes to P
       • Alice goes to R or S
       • Bob goes to Q and tells Alice to come from one                    Q
          side or the other of the cave                                R       S
       • If Alice knows the secret, she can appear from
          the correct side of the cave every time
       • Bob repeats as many times until he believe
          Alice knows to open the secret door
                                                                                           86
Previous
Work




Privacy
Enhancements




               87
•   2 dimensions:
     – CPU – Memory

•   Structural adaptation
     – Add or remove
       component

•   Behavioral adaptation
     – Shift in 2D space

•   Extra dependency
     – Add an extra dimension




                            88
• Resource vector:
  change in resource
  requirements

• Free resource
  space: hatched area




                   89
   S


          network
         connection




                      Node B
Node A

                               90
 Location    Type     Accuracy   Limitation    Sensing    Battery   CPU
technique                                     Frequency   Usage     Load

Bluetooth Physical 5 - 10 m      Outdoors     1 sec       30 % / h 12 %
GPS

WLAN        Physical < 10 m      >= 3 APs     1 sec       41 % / h 17 %
RSSI
WLAN        Physical < 100 m     >= 1 AP      30 sec      38 % / h 13 %
MAC
IP          Symbolic Country/City Online      4 min       34 % / h 5 %
Address                           Database

Personal Symbolic user                        4 min       12 % / h 2 %
Agenda

                                                                           91
– Experiment: Location-awareness



WiFi signal             Bluetooth
   strength                 GPS




                    – Select the best
                      component given the
Personal
    Information       current context
Manager
                    – Save as much energy as
                      possible
                                            92
       Experiment: Location-awareness
          • Resource aware location-awareness
               – Play music on nearest device
               – Auto-set status of IM client
               – Go to power-safe mode in meeting rooms
           • Two identical PDAs: without and with adaptation




WiFi SSID + signal               WiFi + GPS + PIM:
   strength: no adaptation          runtime adaptation
                                                               93
•   Diminution of context ambiguity:
    –   Context quality parameters in a vector space model
    –   Context queries/responses are compared with distance
        function R


•   In a d-dimensional quality parameter space Q:
    –   Context information vector: c
    –   Context query vector: q
    –   Empirically defined weights: w




                                                               94
           x             x    x
                  x     x             x
                x o o o
                              x               x
            x o o  o
Context               o   x
query                   o         x       x
               x    o
          x               x
                   x
             x




                                                  95
•   Why relevance feedback?
    –   Compute better/faster results next time based on this feedback
    –   Update/expand queries or revise results, boost weights
                 Query vector q1 = α . original query vector q0
                         + β . positive feedback context vector ck
                         - γ . negative feedback context vector cl
    –   Backpropagation to delivering node is initiated whenever:
        •   Irrelevant: max hops reached or no more peers, last node did not mark
            the message as relevant
        •   Unused: information is relevant but not used
        •   Duplicate: message already passed through this node through a quicker
            path
    –   Algorithm: see paper

                                                                               96
• Generating a large scale artificial network
  –   k nodes in unit square, 1st in center and other randomly distributed
  –   Weighted location dependent wiring
  –   Nodes nj are connected to at most m existing nodes ni (i<j, j=2..k) that
      each minimize a different function Fm:

                         Fm(ni, nj) = H(ni, nj) + wm.D(ni, nj)

      where H = number of hops between 2 nodes, D = Euclidean distance, and
      wm influences geographical dependency
  –   Simulation in OMNET++ network simulator



                                                                                 97
HOST A     HOST B




 Service




                    98
           HOST B



HOST A




 Service




           HOST C

                    99
HOST A    HOST B




Service   Service




                    100
101
• Similarity analysis
   – Find similar situation for current measurements
      • 50% glucose level, 25% food intake, 25% activities
   – Also compare with 3 previous measurements
      • 4 combined scores 50%, 30%, 10% and 10%




                                                             102
• Location and activity accuracy based on user
  feedback
   – Reliable?


• High accuracy
   – 4 exercise levels
   – Inactive people




                                                 103

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:19
posted:6/27/2012
language:English
pages:103