An Artificial Perceptual Model Approach - Artificial Human Sensors: Science and Applications - page 107

5.1 INTRODUCTION

The traditional artificial sensor approach for the identification of qualitative
parameters is illustrated in Fig. 5.2. This basic principle is that each fraction of
what we consider a qualitative sensing, e.g., odour, taste or other experiences
related to the human perception ability, leaves a characteristic pattern, fingerprint
or response map. Based on the sensor system's operational principle, the initial
process begins by establishing a measurable object. This first part in a measure-
ment cycle will organise the sensors, i.e., refresh, clean or other measures taken
to make the sensors able and ready to take the measurement. Then the measured
object is attached to the sensor elements, i.e., by a chamber or finding a proper
time occasion. The measurement is done at a proper time stamp to ensure that
a common comprehensive picture is taken at the same moment. The signal re-
sponses collected for each sensor will establish a number of single spot measure-
ments which convert the chemical, biological or physical reaction in the chosen
environment into a joint and cooperative value of correspondent electrical param-
eter. Since there is an expectation, that the response from each single sensor is able
to exhibit one characteristic response for only a single reaction, some sensors may
naturally exhibit a response profile that is similar for several different responses,
i.e., the concept of sensor selectivity. The degree of selectivity in the sensor array
and the type of environment sensing responses that can be detected, are highly
dependent on the choice and numbers of selected sensors in the sensor array. In
some special applications, the mounting and place of each sensor can be of in-
terest, mainly due to the application and the way of applying compounds on the
array. The process, however, is dependent on the circumstances that each sensor
response will be a fraction of the overall result. The result is then highly dependent
on the single sensor contribution that is working in its normal condition, without
exceptional variations in temperature, irregular application of compound or flow,

Figure 5.2. A principal quality identification system, illustrating that the sensors may be
directly active in its environment... or at a remote distance and detecting different parts of
the object.