an intelligent broker architecture for context-aware systems a phd. dissertation proposal in...
TRANSCRIPT
An Intelligent Broker An Intelligent Broker Architecture for Context-Aware Architecture for Context-Aware SystemsSystems
A PhD. Dissertation Proposal in Computer Scienceat the University of Maryland Baltimore County
January 2003
Presented by Sangkeun Lee IDS Lab.
Copyright 2008 by CEBT
About this PaperAbout this Paper
CoBrA
Is one of well-known context-aware systems by Harry Chen
Is an agent based architecture for supporting context-aware systems in smart spaces
There are more than 14 papers related to CoBrA
– http://cobra.umbc.edu/
Is the topic of Chen’s PhD thesis, and he’s achieved his PhD. Degree in Dec, 2004
This Paper
Is a PhD. Dissertation Proposal
Is a very early research related to CoBrA
Describes a research plan to develop a broker-centric agent architecture
May help us preparing our degree papers
IDS Lab. Seminar - 2Center for E-Business Technology
Copyright 2008 by CEBT
ContentsContents 1 Introduction
1.1 What’s Context-Aware Computing?
1.2 Problems in Context-Aware Computing
– 1.2.1 Lack of Reusable Context-Aware Mechanisms
– 1.2.2 Limited Resources in Mobile Devices
– 1.2.3 Privacy Issues in Accessing User Information
1.3 Proposed Solution: Context Broker Architecture
– 1.3.1 Use Case
– 1.3.2 Research Contribution
1.4 Proposal Outline
2 Background
2.1 Reasons to Study Context
2.2 Definitions of Context
2.3 Aspects of Context-Aware Computing
– 2.3.1 Enhancing User Interfaces
– 2.3.2 Guiding the Adaptation of System Behavior
– 2.3.3 Enabling Smart Space Applications
2.4 How Do Applications Acquire Contexts?
– 2.4.1 Direct Access to Hardware Sensors
– 2.4.2 Facilitated by a Middle-ware Infrastructure
– 2.4.3 Acquiring Contexts from a Context Server
2.5 Context-Aware Computing Meets Semantic Web
3 Context Broker Architecture
3.1 Objectives
3.2 Modeling the Physical World in Domains
3.3 An Example System of the Context Broker Architecture
3.4 The Design of Domain Context Broker
– 3.4.1 Knowledge Base
– 3.4.2 Inference Engine
– 3.4.3 Context Acquisition Component
– 3.4.4 Broker Behavior
4 Summary and Research Plan
4.1 Initial Approach to Implementation
– 4.1.1 Modeling Ontology and Policy using OWL
– 4.1.2 Building the Inference Engine using Jess
– 4.1.3 Implementing the Broker Behavior using JADE
4.2 Feasibility Study
4.3 Milestones
IDS Lab. Seminar - 3Center for E-Business Technology
I will skip some of the contents that we’ve already discussed many times (e.g. What’s Context-aware computing?, Definitions of Context)
Copyright 2008 by CEBT
Problems in Context-Aware Problems in Context-Aware ComputingComputing
Limited Resources in Mobile Devices
Battery Power Constraints
Information Storage Constraint
Computing Power Constraint
Communication Constraint
– Agents may be lack of sufficient knowledge to communicate with context sources in the environment
e.g. knowing which sensor can provide what information and how to communicate
– Context sources may dynamically join and leave the environment without notifying the agents
IDS Lab. Seminar - 4Center for E-Business Technology
Copyright 2008 by CEBT
Problems in Context-Aware Computing Problems in Context-Aware Computing (cont’d)(cont’d)
Privacy Issues in Accessing User Information
People are worried about how computer systems use and share their personal information
Agents often collect and share user information
Agents also share user information
Lack of Reusable Context-Aware Mechanisms
Many of existing context-aware systems are difficult and costly to build
Because existing systems are tightly coupled with agent implementations, modifying the system often requires modifying the implementation of the agents.
– Growing the number of agents could cause difficulties
IDS Lab. Seminar - 5Center for E-Business Technology
Copyright 2008 by CEBT
BackgroundBackground
Aspects of Context-Aware Computing
Enhancing User Interfaces
Guiding the Adaptation of System Behavior
Enabling Smart Space Applications
How Do Applications Acquire Contexts?
Contexts are acquired by directly accessing low-level context sensors
Contexts are acquired from some kind of middle-ware infrastructures that in turn interact with low-level context sensors
Contexts are acquired from servers that maintain situational knowledge about the environment
IDS Lab. Seminar - 6Center for E-Business Technology
Copyright 2008 by CEBT
Proposed Solution: Context Broker Proposed Solution: Context Broker ArchitectureArchitecture
The author proposes to develop a broker-centric agent architecture (i.e. COBRA) to provide runtime supports for context-aware systems in an Intelligent Meeting Rom environment
The core of the COBRA architecture: Domain Context Broker is responsible for
Maintaining the context model of the domain, which includes domain contexts from the past and at the present
Resolving inconsistencies and ambiguities of the domain contexts through information fusion
Establishing privacy policies with users before sharing their personal information
Providing knowledge sharing service for context-aware agents through agent communications
IDS Lab. Seminar - 7Center for E-Business Technology
Copyright 2008 by CEBT
Context-Aware Computing Meets Semantic Context-Aware Computing Meets Semantic WebWeb
Semantic Web & Context-Aware System
Semantic Web standards will enhance the communication and information sharing
Semantic Web will become a major source of information for determining the context of users and thie associated daily activities
– E.g. When a person enters a room, his location information (his presence in the room) is automatically updated in his personal web site. Immediately, context-aware agents in both remote and local environment can acquire the presence information about the person through the Web without needing any sensing or reasoning.
IDS Lab. Seminar - 8Center for E-Business Technology
Copyright 2008 by CEBT
Context Broker ArchitectureContext Broker Architecture
SemanticWeb
PervasiveComputing
Software Agents
CoBrACoBrA
CoBrA not CORBA!
Copyright 2008 by CEBT
COBRA ArchitectureCOBRA Architecture
Objectives
Acquiring context from heterogeneous sources
– From sensors, Web, user profiles, behavior pattern
Maintaining consistent contextual knowledge
Enabling knowledge sharing among agents
Protecting the privacy of users
Domain Context Broker
Context sharing
Protecting privacy
IDS Lab. Seminar - 10Center for E-Business Technology
Copyright 2008 by CEBT
The Design of Domain Context BrokerThe Design of Domain Context Broker
IDS Lab. Seminar - 11Center for E-Business Technology
Copyright 2008 by CEBT
Knowledge BaseKnowledge Base
Domain Ontology
Context Model : fundamental concepts for representing contexts
Information Source Model : sensors, user devices, web services, …
Agent Profile Model : types or services that an agent offers, the reason for acquiring certain contexts etc.
Information Privacy Model
Domain Heuristics
Help the broker to resolve knowledge inconsistency caused by imperfect sensing
Help the broker to build and maintain a shared model of context
User Privacy Policies
What information can or can’t be shared, etc
IDS Lab. Seminar - 12Center for E-Business Technology
Copyright 2008 by CEBT
Inference EngineInference Engine
Ontology Reasoning Module
deduce facts that can be concluded from the knowledge in the Knowledge Base in conjunction with the models that are defined in the Domain Ontology
Context Reasoning Module
Reasoning about situational conditions of an entities
Knowledge Maintenance Module
Check logical implication
– “If A is true, then B must be true”
– “B is not true”
– “A is true”
– “B is true”
– Inconsistent!
Apply domain heuristics
– “No one person can be in two different rooms at the same time”
– “Bob is currently attending a meeting in RM 201”
– “Bob is currently typing a paper on his computer in RM 102”
– Inconsistent!
IDS Lab. Seminar - 13Center for E-Business Technology
Copyright 2008 by CEBT
Broker BehaviorBroker Behavior
Privacy Policy Negotiation
IDS Lab. Seminar - 14Center for E-Business Technology
Copyright 2008 by CEBT
Broker BehaviorBroker Behavior
Knowledge Sharing
IDS Lab. Seminar - 15Center for E-Business Technology
Copyright 2008 by CEBT
Use Case: An EasyMeeting ScenarioUse Case: An EasyMeeting Scenario
Alice enters a conference room
The broker detects Alice’s presence
B
Policy says, “can share with any agents in the room”
A
B
The broker buildsthe context model
Web
Alice “beams” her policy to the broker
B
Policy says, “inform my personal agent of my location”
AB .. isLocatedIn ..
Copyright 2008 by CEBT
Use Case: An EasyMeeting Scenario Use Case: An EasyMeeting Scenario (cont’d)(cont’d)
Her agent informs the broker of her
role and intentions
+
The broker tells herlocation to her agent
A
The projector agent wants to help Alice
The projector agentasks slide show info.
B
The projector agent sets up the slides
The broker informsthe subscribed agents
B
Copyright 2008 by CEBT
Summary and Research PlanSummary and Research Plan
Initial Approach to Implementation
Modeling Ontology and Policy using OWL
Building the Inference Engine using Jess
Implementing the Broker Behavior using JADE
Feasibility Study: Validating following hypothesizes
By exploiting the Semantic Web infrastructure, context-aware agents can access a wider range of contexts
By developing a hybrid context reasoning mechanism, we can help capability limited agents to maintain a coherent and consistent model of contexts in a dynamic environment
By enabling agents to share a common model of contexts, we will allow agents to access contexts
By defining a policy-driven mechanism to control how users’ personal contexts are used and shared, we can effective protect the privacy of users
IDS Lab. Seminar - 18Center for E-Business Technology
Copyright 2008 by CEBT
DiscussionDiscussion
Evaluation
Clearly defines the problems to solve
Provides a good conceptual design of the system
shows us how we should plan for a research
Well-written proposal paper
However, more details are required
– Will be presented
Ideas & Discussion
Writing proposal seems very meaningful
How can we show that a new context-aware system is better than existing ones?
IDS Lab. Seminar - 20Center for E-Business Technology