| |
5.3. The Artifact-Based Approach
69 regarded as context and why location was su
cient to provide context-awareness at least to a certain extend. Nevertheless, location systems, can do a lot for context-
awareness, but in direct comparison with On
Object and On Human they are in an
inferior position. 5.2.3
Conclusion This section is a first step towards systematizing the use of sensor technology. There-
fore, six dimension of sensing have been identified representing the sensing goals
for physical interaction. We reviewed existing ubiquitous computing research for
an evaluation of sensing technology with respect to the dimensions of sensing and
physical sensor placement opportunities. These categories should support applica-
tion developers in finding appropriate sensors during system design.
5.3 The Artifact-Based
Approach This approach focuses on distributed perception systems for pervasive and ubiq-
uitous computing scenarios. It is our particular interest that these described ab-
stractions work for di
erent types of applications and are not tailored to a spe-
cific scenario. Accordingly, we set out to discover meaningful layers of abstraction
with regard to embedded and also mobile devices concerning distributed, collec-
tive, and cooperative perception systems. When looking at ubiquitous comput-
ing scenarios it becomes apparent that sensing and context extraction in ubiqui-
tous computing applications is generally related to entities
[ Abowd et
al. 1999, Pascoe 1998
] . To create a common artifact-based model we extracted commonali-
ties of a range of di erent distributed perception systems in ubiquitous computing
applications and application classes. Based on these findings, we consider sensing
and context information to be related to an artifact or more generally to an entity.
Furthermore, artifacts do not only occur in isolation but also in settings where rela-
tionships can be established among artifacts. Key issues reflected in the model are
networking and mobility and their e
ects on perception. Our perception model has three layers of abstraction, namely artifact layer which
is directly related to entities, setting layer, and application layer. We implemented
this model on the Smart-Its platform
[ Beigl et
al. 2003, Smart-Its Project 2004
] and developed various applications building upon it. Also, we assess and discuss the
model with respect to other, independently developed applications in order to show
the applicability of the model to a larger range of ubiquitous computing applications.
The layered perception model is described in the Section 5.3.1. This model has been
implemented on the Smart-Its platform o
ering a perception API, as described in
|  |
|
| |
|
|