Gonzalez

judgedrunkshipServers

Nov 17, 2013 (3 years and 6 months ago)

85 views

Collaborative Context Recognition for
Mobile Devices

Software for Context
-
Aware

Multi
-
User Systems

Professor Joao Sousa

David Gonzalez


Overview

Summary of Huuskonen CCR

Theory Abstract.

Model Interpretation.

Implementation Close look.

Long
-
term context


Related works.

Recommendations.

Theory Abstract


Once upon a time....

Mobile Devices(MD) were too limited(e.g. Power
computing, Energy dependent, not common).


Well, still is like that but they are “ubiquitous”.


PCs are not “wearable”, but MDs are.



MD User Interface are limited, but they are
Communication Hubs.


Theory Abstract


Human Computer Interaction(HCI) must
integrate Sensors to engage a real Context
experience.


Sense of:


Location


Social Situation


Tasks


Activities


Must be easy to the user, but the
implementation is not trivial.


Theory Abstract


Context Awareness (CA):


Humans are a “Rank
-
A” CA animals, because:


We use CA for primitive functions like
Survival, Reproduction and Subsistence.


Imitate and Learn is a common behavior, so
We are Context
-
driven individuals.


The issue is how transfer this to Machines.

Simple Model for Human Behavior

CA

Lost

Doubt

Do

Ask

Imitate

Mobile Context Awareness


This is the first step to allow CCR.


It merges IA and HCI.


Examples:


Location


Environmental Sensors


Biometrics


Acceleration sensors


Multimedia

Application Area


Geomarketing


Jaiku


Clarity Brickstream


Nintendo 3DS


Latitude by Google

Long
-
term goal


State CCR as part of global Initiative.

This is not isolated research, but a common
effect of Computing Paradigm Shift.


Establish improvements to the current
architecture.

Till now the architectures work, but lack of
new frameworks to ease the inherent
flexibility of this kind of systems.

Model Interpretation



A CCR Looks like:

Context

Awareness

Context


Recognition

Context


Reasoning

Sensors signals

Process

Method

Signal

Processin
g

Weighted

Voting

Protocol


CCR

Server

Model Interpretation



A CCR System Looks like:

Context

Awareness

Context


Recognition

Context


Reasoning

Sensors signals

Process

Actor

Mobile

Device

Mobile

Device

Group

CCR

Server

Implementation Close look

Symbian S60, IOS

Apache Tomcat

Windows, Linux

Actor

Mobile

Devices

CCR

Server

Development up to present


State CCR as part of global Initiative:


2008, Bannach


Context Recognition
Network


2005, Sung & Blum


Wearable computing


2003, Huuskonen


CCR for MD

Recommendations


New SW Platforms are requires, in this
particular case: Android.


Stronger Architecture are required in the
Business layer, specifically Web Services.


Ontologies are proposed, not yet
implemented.

Architecture

ideas

Data Access

Business

Presentation

Data Mining for
new Contexts
rules

More Flexibility

and spreadable

with Web Services

Rich User Interfaces,

Context Aware like DK