| |
68 Chapter
5. General Concepts, Toolkits, and Frameworks Object
Use. On Human is suited
for direct measurements of human-centric sensing aspects,
such as Bio Signs/Emotions and Activity. Mutual
Collaboration sensors, such
as the location systems perform best Location sensing,
but also can give hints for
other dimension. Quite interestingly, each sensor placement is meaningful for
it least one sensing
dimension.
Table 5.2: Evaluation
When looking at possible sensor placements
more globally, Table 5.2 depicts that In
Environment and On
Human o er the best sensing results. Analyzing the dominant
factors for each placement, it points out that video and audio are most prominent
for In Environment sensing. However, the perception quality depends on compu-
tational expensive recognition, as video and audio per se can only provide indirect
measurements which have to be analyzed and interpreted. In contrast to that, e.g.
inertial sensors placed On
Object can measure directly Activity which requires only
little computational resources since the sensor result is already more apparent than
a video or audio output. Nevertheless, once an environment has been augmented
with sensors and the necessary computational resources, e.g. smart rooms, the clear
advantage is that applications work without additional instrumentation of users
or objects. Then, such systems can give hints about human-human interaction if
provided with a global view on all activity going on in the environment.
As physical interaction with everyday object mostly involves movements, such as
grasping, moving or turning, the dominant sensor technology for Object
Use are inertial sensors. On
Human placement is well suited for various sensors such as in-
ertial sensors, audio, bio sensors and also video to a certain extend. With regards to
sensing human activity On
Human represents the closest to phenomena placement.
In the real world location is a very determining factor: Whether we are at work,
home, in the car, or even more specific, in a meeting room, lecture hall, kitchen, or
elevator can give hints on our activity, objects, possible interactions with others, and
perhaps even subtle hints on our emotions. Due to this high relevance of location in
our real world Mutual
Collaboration sensors, which mostly exist as location systems,
can provide coarse information about Object
Use, Activity and In
Environment. This also explains why in the first years of context-aware computing mostly location was
|  |
|
| |
|
|
