Sensor Technologies - Artificial Human Sensors: Science and Applications - page 93

overall qualitative value in the complete range of operation. In the data analysis
process, a model, or calibration curve can be established, describing the relation-
ship between the environment conditions and the qualitative response measured
for the acquired quality. The calibration has to be secured in a complete multi-
dimensional space of the system's operational range.
Reliability
The sensors are foreseen to behave according to their specifications and be predic-
tive particularly over a long period of time. The single sensor relationship to each
other is also of importance, when the selectivity and internal relations are consid-
ered in calculating the qualitative parameter of a multi-sensor system. Therefore,
specific considerations have to be taken when changing one or more sensors in a
multi-sensor system or replacing a complete sensor array.
Accuracy
Sensors have to exhibit a high sensitivity in a specific domain and preferably some
response sensitivity for related domains. The system accuracy is created as an
additive effect that depends on the different accuracies of each part of the system,
e.g., the sensor elements or electronic processing unit. The drift of each sensor
unit can strengthen or reduce the overall accuracy, if for example, the sensor's
drift results in a positive or negative offset effect. The system accuracy has to be
continuously estimated in coherence with other affecting parameters such as noise,
drift, etc.
The measurement procedure of a multi-sensor system is straightforward in
a sequence. When each sensor element has measured a quantitative value, the
data is collected and calculation takes place. In most multi-sensor applications,
the data is multi-dimensional, with the objectives of processing the data into a
more presentable nature. The multi-dimensional data is often a calculated dy-
namic qualitative value representing the measured picture of the environment of
interest, e.g., a correlation value to the human estimation. These qualitative values
often perform an illustrative connection, comprised in a “picture” that may appeal
as a situational awareness, such as a glimpse of the measured environment, as il-
lustrated in Fig. 4.8. This spot of information may provide a human understanding
and perception related to complement the information. This subject will be further
illustrated in subsequent chapters of this topic.
4.4.3 Complex Sensor Data Fusion Principle
The traditional view of a multi-sensor system may include separate sensors that
are located at a distance from each other. The measured picture of the environ-
ment of interest then may be of importance when fusing data from a wide, partly
measured and not aggregated environment. The distributed organisation of a sen-
sor system network is then an excellent tool, to explain the non-homogeneous