Sensor Technologies - Artificial Human Sensors: Science and Applications - page 95

— avoid wasting effort to collect and process irrelevant information,
— and possibly include some intelligent features on a higher level, regarding, e.g.,
the estimates, significance and interface.
A sensor data fusion system may also integrate redundant sensor data as
early as possible that is in the systems lower levels. Complementary informa-
tion is preferably fused at higher levels of the multi-sensor fusion architecture.
The multi-sensor strategies are important parameters, concerning the overall in-
formation performance. The selection of single sensor elements in a multi-sensor
architecture consider the placement in such a way as to get optimum performance
during specific measurement tasks, and for the time selection of measurement ac-
tivities to minimize the dynamically observed system response. The multi-sensor
fusion system is typically based on an ”intelligent” structure that may change strat-
egy under processing and use the best available mode of operation at the right
time. The complex system has an application independent abstraction that models
the sensor function as a software description, performing its input/output charac-
teristics including measuring process and the ability to selectively direct the mea-
surement parameters of task specific interest.
A significant summary of the sensor fusion concept is stated in Biel (2002),
that the strategy of handling sensor responses in a fusion process is comprised as
follows:
“sensor fusion is the process of combining data in such a way that the result pro-
vides more information compared to handling each source separately”.
An extensively used method of data fusion technique particularly in human
related sensing, is the multivariate analysis that comprises one phenomenon by
observing or sensing a number of observations. This method is suitable when mea-
suring more than one variable simultaneously. Further reading is recommended
in Naes (1996) and Esbensen (2000).
4.4.4 The Human Role in the Sensing System
In principle, the basic technologies are comprised of sensing methodology, that of-
ten tries to mimic the human's perception of physical, chemical or biological input
sensations and to convert the measured parameter. This process may be perceived
in a structure, where the human perception can be involved in the measuring pro-
cessing, for example as a sensing unit, processing the measured quantity value in
the system, or be an active part of finding the right decision from several informa-
tion sources.
In order to actively live in complex circumstances, humans have to understand
the environment they are taking an active part in. Also, we have to conceptualise
notions that will enable us to evaluate the impact of different natural phenomena,
as well as of the impact in the interaction process of other human actions. The
human in the system is considered to be an expression, indicating that in the daily