| |
66 Chapter
5. General Concepts, Toolkits, and Frameworks audio.
Inertial sensors placed On Object and On
Human can be used to sense typical movements,
e.g. perceiving the signature at a pen, for identification.
[ Scheirer et
al. 1999
] reports
about using vision, [
Kern et
al. 2002 ]
about using audio worn
On Human for
people identification. Location systems [
Hightower
and Borriello 2001
] can also be used for identifying people at di
erent locations. As these systems
require both sensors worn by human and installed base stations those system appear
in the Mutual
Collaboration column.Table
5.1: Placement vs. Dimension For detecting object
use load-cells [
Schmidt et
al. 2002 ]
have been proven useful
installed both In
Environment and On Object. Object classification with vision
is well established in static settings, whereas occlusion during dynamic use can
be challenging. Audio is another option, if the object use generates characteristic
sounds. The use of inertial force sensors placed On
Object has been successfully used
[ Hickley et
al. 2000 ] [
Schmidt et
al. 1999 ] [
Rekimoto 1996
] [ Harrison et
al. 1998 ]
for object
use. Obviously, motion during Object
Use can be also sensed On Human
but with less quality. Audio On
Human is also possible [
Lukowicz et
al. 2002 ]
but is an indirect measurement compared to On
Object placement. Location systems
can give hints as well for Object
Use, e.g. teleporting X Windows to users current
location [
Richardson et
al. 1994 ]
. Location is the most explored sensing dimension in ubiquitous computing. Load-
cells [
Schmidt et
al. 2002 ]
, vision
[ Brumitt et
al. 2000 ]
and audio
[ Darrell et
al.
|  |
|
| |
|
|