Biomedical Engineering Reference
In-Depth Information
single sensor may not perceive all the components of a qualitative impression. The
design of a well-adjusted multi sensor-system, e.g., a multidimensional selectivity-
based sensor array in close symbiosis with a straightforward pattern recognition
system, is a very powerful tool. The complex system is then able not only, to pro-
vide a complete picture that corresponds to the environment of interest, but also
make necessary corrections, due to error or bias in the system operational princi-
ple.
However, when using multi-sensors, two illustrative application principles
can be viewed. The choice of principles is of basic interest when designing the
system, and is described as follows:
— When using a number of sensors in a system with similar operational princi-
ples, we are looking at an elaborate redundant sensor system that will secure
the measured feature we intend to quantify. This may be proceeded by using
the same sensor in parallel, to ensure that even if one or more sensors are er-
roneous, then we have enough sensor power to provide the right performance.
Also, we may complement the power of a sensor system with different sensor
principles by using different sensor types with a variety of measurement tech-
niques in order to get a selective range of multi-sensors measuring the same
parameter. These types of multi-sensor systems are often used in highly se-
cured functions where system failure is not an option.
— In multi-sensor systems, for example, in electronic nose applications, we aim
for a broad selection of classifications of detectable compounds in the measured
air volume. The selectivity is of importance to detect various chemical com-
pounds in the measured volume. Therefore, the aim is to select a wide amount
of sensors that provide a selective performance and complement each other.
The multi-sensor principle will get a wide selectivity and perform a structure
of quantity values that give rise to a virtual qualitative range of features. The
qualitative value will provide measurements that are not detectable by only
one single sensor and are expected to demand a more complex data analysis.
This concept also requires a strategy for accuracy and calibration standards.
The basic terminology and primarily performance for a multi-sensor system is
slightly more demanding when analysing system performance. Some of the more
important characteristics can be described in the following parameters:
Response time
The multi-sensor system would be able to have a relatively high response time than
a single sensor system, in order to be able to secure that each sensor has enough
time to collect, convert and provide data, i.e., data processing to present a quali-
tative or quantitative response parameter within a predetermined time frame. In
the time frame, the recovery time and the initiating time to prepare for a new mea-
surement sample is often included.
Search WWH ::
Custom Search