Sensor Technologies - Artificial Human Sensors: Science and Applications - page 89

4.4.2.3 SINGLE- AND MULTI-SENSOR SYSTEMS

The use of a single sensor system is directly aiming towards the focus on the spe-
cific sensor performance and thus provides a limited understanding of the prede-
termined measurement goal and expected singular result. A single sensor system
is probably more understandable in terms of required performance and parame-
ters required when considering selectivity in range, sensibility, specificity or sens-
ing features for a number of collaborative sensors achieving a common measure-
ment goal.
A central point in a dynamic sensor system can, with advantage, be defined by
the sensor specification or, in the case of multi-sensors, also by the overall sensor
performance. Some measurement applications have a clear defined aim to perform
single and dirigible sensor output, while others request loosely defined parame-
ters. In reality, one single sensor is typically not able to measure all aspects of a
complex entity without the combination of other sensors with different selectivity.
Then the single sensor sensitivity is directed towards one particular range of inter-
est. In case of multi-sensors, the aim is to coordinate the selectivity from different
sensors in building-up a multi-dimensional selectivity space. An interesting ben-
efit of using sensor arrays is the ability to also find vaguely defined parameters in
applications, where the aim may be to detect or identify single qualitative param-
eters that are not clearly defined. In reality however, many sensor characteristics
are indicatively responsive to several different defined parameters, where in these
cases, they are considered to be non-selective. If a non-selective sensor is used
alone, confusion may occur when indicating different measurement values from
different defined units that can generate a similar response pattern. The drawback
may then be that the measured value is difficult to relate to a certain sensor out-
put and its state of a particular defined parameter. In the case of multi-sensors,
e.g., a sensor array, a design that constitutes a multi-sensor system of locally gath-
ered sensor elements that will provide a multi-dimensional output. A strategically
designed sensor array may achieve higher performance, by using non-selective
sensor elements that are able to strengthen the measurement capability, Pettersson
(2008). The non-selective sensors can together measure a “picture” of the environ-
ment, corresponding to a similar picture of a mountain chain where each single
measured sensor value can be expressed as a mountaintop or valley, i.e., the mea-
sured response plot containing each sensor sensitivity to a common defined pa-
rameter. The full response picture is described as the included sensors' collected
values are inserted in a descriptive pattern of sensitivity. The mathematical tools
used for pattern recognition can be put together in a measurement picture through
a joint analysis of the sensor signals. This procedure is called multi-dimensional
analysis.
The strength of complex sensor array applications aims at sensing complex
parameters, e.g., qualitative values as bad odours, fresh taste or beautiful sight.
The qualitative parameters that usually are requested in these measurements are
in a manner connected to the human understanding and prerequisite. Normally, a