| |
70
Chapter 5. General Concepts, Toolkits, and Frameworks
Section 5.3.2. In Section 5.3.3 two applications, realized with the artifact based
model, are discussed in more detail. Furthermore, we show how scenarios, which
have been implemented before the model existed, could also benefit from the artifact
model. Finally, Section 5.3.4 concludes with a review of our findings.
5.3.1 The Perception Model
Acquiring information about the real world through sensors is a prerequisite for
context-aware applications in ubiquitous computing
[
Abowd and Mynatt 2000
]
.
Considering the context in which sensing happens provides vital information. As
objects and people in the real world are not isolated, but in context and relation
with others, the perception model evaluating the sensor readings should reflect this
connectedness: a spoon, fork and knife in close co-location may indicate a meal -
similarly the perception model has to evaluate the sensor readings regarding the
artifacts the sensors are attached to. Sensors and sensor readings have little mean-
ing by themselves, but connected - physically and logically - to an entity they can
provide important information. We believe that ubiquitous computing systems may
be well modelled by properties of and relations among artifacts where sensors are
attached to.
An artifact-based model allows to specify issues on a single artifact level, but also
acknowledges the fact that artifacts may be part of a larger networked system (set-
ting). When looking from a more global system’s perspective some artifacts may
be statically coupled whereas others may dynamically and continuously group and
regroup in a more open and ad-hoc fashion.
From our analysis we can state the following observations for distributed sensing
systems for ubiquitous computing scenarios. (1) Sensing, data capture, and percep-
tion are related to entities which hold important information for the processing from
sensor data to context information. (2) Perception is dependent on the applications
that make use of the data acquired. (3) The specific device to which the sensor is
attached as well as the environment in which the device is operated are important
when selecting the perception methods and architecture.
From these observations several questions arise:
How is data captured?
What types of data are captured and how is data provided? In particular what
meta data needs to be attached to readings?
Where does data reside and where is data stored over longer terms?
How is data accessed and what are the communication models and paradigms?
|  |
|
| |
|
|
