Biomedical Engineering Reference
In-Depth Information
Data processing
Even if the main request, in most cases, is to establish a stable performance, will
the issue be focussed on each vulnerable single sensor element and its principles?
There is also an influenced quality demand in the data itself and the processing of
multiple collected quantities. The procedure of measuring a single sensor quantity,
that is merged together with other sensor quantitative data into an overall response
picture, imply that the total amount of data can be processed by sophisticated al-
gorithms. The response picture comprising all the single data in a measurement
cycle can further be processed in order to establish an overall qualitative corre-
lated value. The output parameter represents the participating number of verified
sensor, whose data is fused to valuable information.
Noise
Additive noise from different sources is mainly associated with randomly appear-
ing disturbances and errors. When the sensing elements are collecting the data
in their specific environment, we may presume that the collected data quantity is
already contaminated with noise. The probability may be considered high, that
additional noise is caused by the processing part of a multi-sensor system. The
primary interest, however, is to explore, understand and find countermeasures for
these peculiarities, in order to learn the pattern of the incurring noise. The aim is
typically to find proper techniques to ensure that the noise will cause a minimum
effect and error contribution of a measurement sequence.
Drift
The drift is defined as a gradual change of the defined sensor 's output of its quan-
titative value under obviously constant conditions. Due to drift the outcome value
will be unpredictable and systematically changed. If a number of sensors are in-
volved in a measurement system, there maybe a need to make a change in the
overall direction of an expected drift. There are some techniques to compensate
for the systematic drift in a multi-sensor system. The cause of drift characteristics
are complex and may be an effect from many internal and external sources, e.g.,
changes in the electronic surface or external effects from sunlight, temperature or
change in the air pressure. Techniques are available to suppress the drift charac-
teristics, mainly by software adjustment.
Reproducibility
In multi-sensor systems, there is a process to define the overall qualitative com-
mon value based on the contributing sensors quantitative value, when expressing
the status of the momentarily features in the measurement surrounding. The sens-
ing principles of a number of sensing elements are often selectively sensitive to
each other, which become challenging when estimating the reproducibility of the
Search WWH ::
Custom Search