See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/235300307

Intelligent sensors: Structure and system

Article in Sensor Review · September 1995

DOI: 10.1108/02602289510795978

CITATIONS READS

11 3,558

2 authors, including:

Fikret Yalcinkaya
Kırıkkale University
33 PUBLICATIONS 150 CITATIONS

SEE PROFILE

All content following this page was uploaded by Fikret Yalcinkaya on 08 March 2017.

The user has requested enhancement of the downloaded file.

 Sensor Review
Intelligent sensors: structure and system
E.T. PownerF. Yalcinkaya
Article information:
To cite this document:
E.T. PownerF. Yalcinkaya, (1995),"Intelligent sensors: structure and system", Sensor Review, Vol. 15 Iss 3 pp. 31 - 35
Permanent link to this document:
http://dx.doi.org/10.1108/02602289510795978
Downloaded on: 08 March 2017, At: 12:40 (PT)
References: this document contains references to 0 other documents.
To copy this document: [email protected]
The fulltext of this document has been downloaded 257 times since 2006*
Users who downloaded this article also downloaded:
(1997),"Intelligent sensors", Sensor Review, Vol. 17 Iss 2 pp. 97-98 http://dx.doi.org/10.1108/02602289710170267
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

(1995),"From basic sensors to intelligent sensors: Definitions and examples", Sensor Review, Vol. 15 Iss 4 pp. 19-22 http://
dx.doi.org/10.1108/02602289510102327

Access to this document was granted through an Emerald subscription provided by emerald-srm:549116 []
For Authors
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service
information about how to choose which publication to write for and submission guidelines are available for all. Please
visit www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
Emerald is a global publisher linking research and practice to the benefit of society. The company manages a portfolio of
more than 290 journals and over 2,350 books and book series volumes, as well as providing an extensive range of online
products and additional customer resources and services.
Emerald is both COUNTER 4 and TRANSFER compliant. The organization is a partner of the Committee on Publication
Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive preservation.

*Related content and download information correct at time of download.

 Intelligent sensors: structure
and system
E.T. Powner and F. Yalcinkaya

A modern measurement and other into a single A/D converter and Intelligent sensors and
control system from there the signals are sent to the intelligent transducers
A modern measurement and control digital bus. This second solution is In the existing systems, intelligence is
system[1] could be set up as in Figure preferable, because several sensors are still centralized, but the more the
1; in such a system, a great number of positioned quite close together to form systems become decentralized the
sensors collect information about the one unit. more the intelligence will be
processes available in the system being In the last example, the sensor is distributed, and as a result the more
monitored. Although sensors are connected by an A/D converter to a efficient and effective the systems will
employed in many different applica- microprocessor. The microprocessor become. This will need what is called
tions, the more complex the system has stored the sensor’s specific data in an intelligent sensor. A modern micro-
being monitored becomes, the more its memory, ROM, and also the sensor processor-based intelligent sensor[1]
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

powerful, comprehensive or intelligent signal, and the microprocessor could be composed of a sensing
are the sensors required. transfers a corrected signal to the bus. element, a signal-processing part, and a
In the digital sensor bus of Figure 1, An interesting property of this microprocessor as shown in Figure 2.
a central computer is connected to a subsystem is that the microprocessor If we combine a sensing element
number of input and output devices. It also permits the central computer to with a signal processor on a single chip
picks up information from different send data back to the sensor, which can then what we get is an intelligent
sensor systems via a digital bus and be used to change the measurement sensor. In the near future it will be
either makes some signal processing
range, to effect a recalibration or to possible to put an intelligent sensor on
before sending them out or directly
adjust offset. In such a subsystem, each a microprocessor, but much local signal
sends them to actuators again after
sensor or group of sensors also contains conditioning can be taken over via a
some basic conversion operations.
For instance, the output signal of the a circuit that can recognize addresses; microprocessor used in a system.
sensor is first amplified and corrected this means the circuit can detect when Intelligent transducers are a result of
for offset, non-linearity, and other communication between sensor and recent developments in large scale
types of sensor errors before trans- central computer is wanted. “Today, integration (LSI) and very large scale
ferring it to the bus system. In the most of the described components of a integration (VLSI) technology.
second group of sensors , the signals bus system are still separate and have Transducers can be integrated with
are treated in separate signal-con- their own housing. However, in the signal-processing circuits on the same
ditioning circuits for offset, non- future the integration of these com- chip to perform active functions. These
linearity, temperature sensitivity, and ponents on a single chip will be the are defined as intelligent trans-
so on. Then a single multiplexing norm. This is indicated by the broken ducers[2], in a broad sense, and are
circuit feeds these signals one after the lines in Figure 1. shown in Figure 3.

SA Amplifier A/D
A/D Amplifier Actuator

Correction Bus
controller

SB 1 Signal
conditioning Display or printer

SB 2 Signal Multiplexer A/D Central

conditioning computer Telecommunicaion

SB 3 Signal
conditioning

Bus

Storage
SC A/D Microprocessor

Figure 1. Schematic of a digital sensor bus interface structure

Sensor Review, Vol. 15 No. 3, 1995, pp. 31-35, MCB University Press, 0260-2288 SR Vol. 15 No. 3, 1995 31
 lations which need to measure n
quantities through m transducers such
Sensing Signal that each of them will be a function of
Microprocessor Data lines
element processor variables, yet the output will be a set of
accurate values, or the desired final
output quantity….
Smart sensor The final logic and decision-making
function is signal recognition and
Figure 2. Schematic diagram of an intelligent sensor with microprocessor rejection function which compares the
signals with the features stored in the
memory of the system to perform
recognition and rejection functions.
Sensing Signal
Different means of power reduction
Microprocessor Data lines may be used for low signal bandwidth
element processor
transducers to reduce for instance,
power consumption and the time-
sharing between different sensors, or
Smart sensor operating sensors, at low duty cycles.
Figure 3. Block diagram of an intelligent transducer Intelligent transducers can integrate
different system parameters and decide
what to do.
An intelligent transducer is simply an Signal processing includes the coding The status of any system could be
intelligent sensor, as can be seen in and modulation of transducer output controlled and determined via
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

Figures 2 and 3. The functions per- signals to obtain a better signal-to- instrumentation or control systems by
formed via intelligent transducers[2] noise ratio, signal averaging and measuring several parameters of the
are shown in Table I. redundancy, and the self-checking and system being monitored. The para-
Signal improvement, for example, fault detection system to detect any meters measured may be directly or
includes impedance transformation and malfunctioning of the transducers. indirectly related to the status of the
amplification. Examples of signal Another is the ten pH sensors with an system, or the desired output, in a
improvement with impedance trans- on-chip multiplexing circuit to achieve deterministic or statistical manner. An
the redundancy functions[2]. Addi- intelligent transducer could be de-
formation and signal amplification
tional circuitry can be added that will signed to translate some quantities into
have been reported. One example is the manageable information that could be
provide a single output with an average
capacitive pressure transducer in which displayed or used as an input to control
pH value from the ten sensors. The ten
the oscillator and a low output im- sensors’ output pH value will be more other systems with complex processing
pedance amplifier are incorporated accurate and reliable than that from a circuitry and complicated intelli-
with the capacitive pressure trans- single sensor. gence[2]. For example, pressure,
ducer[2]. The re-processing of signals A very important function of intelli- temperature, engine speeds and exhaust
includes compensation for temperature gent transducers is logic and decision- gases are measured to determine the
and power supply variations, offset making functions. These functions combustion status in order to control a
balance, the filtering of unwanted include the trend of the signal, the car engine’s performance. The current
signals and kinds of non-linearity correlation between several transducer trend is that future transducers will not
correction. variables, and multiple variable calcu- only report the values of pressure,
temperature, gas, etc., but, with built-in
intelligence, also produce an output
indicating how complete the combus-
Signal Signal Signal Logic and tion is or what needs to be done to
sensing improvement Pre-processing processing decision
improve the efficiency of the engine.
Briefly, the most significant features
Multivariable z-transformation Temperature and Coding Correlation of intelligent transducers are:
supply ● improved signal quality from
compensation sensors, with wider margin for
noise and interference;
Amplification Offset balance Redundancy Trend ● improved reliability and accuracy;
filtering
● compatibility with computer and
other system blocks, and
Self-checking
● reduction in size and cost of large
quantity lots[2].
Calculation
Elements of an intelligent
Non-linearity Fault Signal sensor
correction detection recognition An intelligent sensor is a very complex
alarm or rejection system which needs a range of disci-
plines to be applied all together in the
Table I. Functions of intelligent transducers design, implementation and realization

32 Sensor Review
 communication processing. Often it
requires a processor of its own,
although it could be realized as part of
Sensor Hardware
Amplification the main processor chip. A multi-drop
element process
bus system is accepted as the natural
form of communication for the intelli-
gent sensor processor. Addressability
Input
variables and integrity of information are another
two important attributes of the
Monitoring
intelligent systems.
As a last point, such an intelligent
Data sensor system is expected to be capable
conversion of[4]:
● converting analogue sensor data to
a digital value;
Data Software
communications process ● correct the value for non-linearity
and temperature drifts;
● send the corrected values over an
IIC serial bus when so requested;
Bus ● self-calibration;
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

● for economic reasons, it is highly

Source : [3] desirable for the chip size of the
Figure 4. Elements of an intelligent sensor complete unit to be as small as
possible.
phases. In addition, an intelligent analogue signals into discrete signals The thrust of VLSI design has been
sensor combines a number of both where a digital processor becomes a predominantly in the digital area. It is
analogue and digital subsystems which must. At this stage it is important to only in recent times that the VLSI
make the job more difficult, because bear in mind that the process of A/D techniques are being applied to
sensors largely operate on signals that conversion is non-linear and represents analogue circuits and, in particular, to
need very complicated digital and a potentially gross distortion of the the intelligent sensor area. It is well
analogue signal processing. The main incoming information, because the known that most instrumentation and
subsystems within the structure of an signals convey the information. This control systems employ a digital micro-
intelligent sensor are[3]: stage may need very important processor, despite the sensors being
● A principal sensing element. software which will control the analogue. Thus, analogue-interfacing
conversion process as a part of a control systems are required, consisting of gain
● Excitation control. loop or accomplish some sort of auto- stages, multiplexers, analogue to digital
● Amplification ( possibly variable ranging for the intelligent system under converters and a communication
gain). investigation. system connecting sensors to the
Compensation is central to the processor. It would be more convenient
● Analogue filtering.
philosophy and main reason for the if the sensor outputs were digital and
● Data conversion. existence of the intelligent sensor: also communication via a serial bus, such as
● Compensation. its needs may affect the basic design of the Intel/Philips IIC system[4]. The
the system. An intelligent sensor may form of such a sensor is shown in
● Digital information processing. be able to compensate for any para- Figure 5 and consists of a single chip
● Digital communication process- meter of the system being monitored containing the sensor proper, a
ing. which is not steady in the acceptable conversion block, corrections/precon-
The principal sensing element is the range. Information processing is, ditioned stage and finally a
without doubt, unique to the intelligent
heart of any intelligent sensor design communications block. Most of the
form of sensors. Although there might
and implementation. Unless it works communication and signal processing
be some overlapping between com-
properly, nothing else works. Excita- pensation and information processing, is readily undertaken by the simple
tion control may take different forms there are some areas where they microprocessor. Since the speed of
depending on application. Amplifica- become fairly independent of each operation is not critical and a serial
tion of the output of the principal other. An intelligent sensor system communication link is employed, a
sensing element is nearly always a needs compressed information and simple serial processor is adequate. It
fundamental requirement. Analogue data to check the integrity of incoming is possible, depending on our appli-
filtering is a fundamental requirement, and outgoing information to make sure cation, to put more than one sensor on
at least to block aliasing effects in the that they are physically reasonable. the same chip and even sometimes it
conversion stage. It is also becoming The information-processing stage may could be possible to use one sensor as a
an attractive requirement especially as destroy information instead of creating measurement element as well as a
digital filtering is most likely going to it or even introduce false information correcting element.
take up too much of the real-time to the system. One basic and one more general
processing power available. Data The final, but extremely important, VLSI-based intelligent sensor are
conversion is a conversion stage of element of an intelligent sensor is shown in Figures 5 and 6.

SR Vol. 15 No. 3, 1995 33

 converter or put the sensor element
Real Sensor Conversion Corrections/ Communication
Bus itself on the microprocessor. Without
world pre-conditioner doubt, this final stage will have a
revolutionary effect on the intelligent
sensors as well as on the measurement
Address and control systems.

Source :[4] The role of intelligence in

Figure 5. Basic intelligent sensor configuration smart sensors and sensor
systems
The roles of intelligence sensors are to
Sensor-01 Address enhance design flexibility and realize
new sensing functions, and additional
roles are to reduce loads on central
processing units and signal trans-
Bus
Conversion Processing Communication mission lines by distributed informa-
tion processing in the lower layer of the
system.
Sensor-nn
The sensor intelligence performs
Serial microprocessor
distributed signal processing at the
lower layer in the bottom-up structure
of the sensing system. The role of
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

Source :[4]
signal processing in intelligent sensors
Figure 6. General purpose intelligent sensor system system can be summarized as:
An intelligent sensor: a system of sensor, the function is referred to as ● signal improvement for extraction
sensor signal fusion. of useful features of the objects;
approach
I think both of the following figures Another function is the tuning of ● fortification of intrinsic attributes
will help to understand more easily sensor parameters to optimize the total of sensor device.
what is meant by system approach to system performance. Finally, in
The important role of sensor intelli-
intelligent sensors. Actually an intelli- general, the property of information
gence is to improve the signal selec-
gent sensing system generally has a processing done in the lower layer is
tivity of individual sensors. This
hierarchical structure as human sensory more directly related to hardware and
includes simple operations of output
system having multi-layer as shown in less hardware-related in the higher
from multiple sensor devices for
Figure 7. layer. For the same reason, the
feature extraction. However, this does
The top layer acts like the brain of algorithm of information processing is not include the optimization of device
the intelligent system where the most more flexible and more knowledge- parameters or signal integration from
intelligent information processing is oriented in the higher layer and less multiple sensor devices, because this
being conducted. Processing is cen- flexible and less knowledge-related in requires knowledge of the sensor
tralized in the top layer. Processed the lower layer. Processing in each devices and the objects.
information is not dependent on layer can be characterized as follows: Mainly there are three roles of
operating principle and physical ● signal processing in the lower intelligence in the middle layer of
structure of sensors. In contrast, layer; intelligent sensing systems. The first
information processing in the lower ● information processing in the role of intelligence is to extract
layer is strongly dependent on the middle layer; essential features of the object while
sensor’s underlying principles and the second role is to organize multiple
structures. A group of intelligent ● knowledge processing in the upper outputs from the lower layer and
sensors or ordinary sensors on the layer[5]. generate average output. The middle
lower layer collects information from The most popular concept of an layer combines or integrates the
the external objects like our distributed intelligent sensor is an integrated coming-out signals from multiple
sensory organs. Signal processing of monolithic device combining a sensor sensors. Then the extracted features are
these intelligent sensors is done in a with a microcomputer within one chip. utilized for recognition of the situation
distributed and parallel manner. However, such a device has not yet by upper layer intelligence.
The role of sensor intelligence in the been realized. The development stages Signals from sensors of different
lower layer is limited within the signal of such intelligent sensors[6] are measurands are combined and the
processing of signal selectivity and the illustrated in Figure 8. results give us new, useful information.
characteristics compensation of indi- Technology is now in the second Ambiguity or imperfection in the signal
vidual sensor devices. This layer is stage of the developmental stage which of a measurand can be compensated by
much more hardware-dependent. means that technology is able to another measurand. This is sensor signal
In the middle layer, there are integrate some signal-conditioning fusion. This feature of processing
intermediate signal-processing func- circuitry on the sensor element itself creates a new phase of information.
tions. One function of intermediate while at the same time integrating A/D The third role of the middle layer is
signal processing is the integration of on the microprocessor. The final goal is the parameter tuning of the sensors to
signals from multiple sensors. When to integrate the sensor element with optimize the total system performance.
the signals come from different types signal-conditioning circuitry and A/D Optimization is done based on the

34 Sensor Review
 Upper layer Knowledge processing Sensor system intelligence Conclusion
Total control
Intelligent sensors are a rapidly
expanding area of research which will
Concentrated central processing Integration inevitably bring extraordinary changes
Intelligent Intelligent
materials structures
with in the current medical, industrial,
(Digital serial processing) computers
avionics, consumer products, space re-
search, manufacturing, and many other
Middle layer Information processing Figure 9. Different methods of sensor system
real life problems. Therefore, at this
intelligence
Intermediate control, tunning and stage what is needed urgently for intelli-
gent sensor design is a deeper under-
optimization of lower level sensor signal intelligent because they choose
standing of the basics of sensor systems
only the signal we are after; they
and their application to technology.
integration and fusion suppress the noise and undesirable
Unless the basics of traditional sensor
effects; they fit very well the object systems are understood, new intelligent
Lower layer Signal processing material with sensor material to sensor design will be very difficult.
Sensing and signal conditioning realize nearly ideal signal selec- Specifically, the area of medicine will
tivity. These materials will be very get much more benefit from intelligent
Intelligent sensors useful especially for biosensor sensors, and probably much more from
applications. intelligent material-based intelligent
Distributed parallel processing
(2) The use of a functional mechanical sensors, i.e. biosensors, although nowa-
Analogue days microcomputer-based intelligent
structure, the so-called intelligent
structure. They are called intelli- sensors are much more popular.
Source :[5]
gent, because they have the ability
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

Figure 7. Structure of an intelligent sensing to choose only the signal we are References
system looking for, and to suppress the 1. Middlehoek, S. and Audet, S.A.,
noise or undesirable effects. This Silicon Sensors, Academic Press, New
extracted feature and knowledge about method needs silicon technology York, NY, pp. 331-3.
the target signal. The knowledge is and silicon micromachining which 2. Ko, W.H. and Fung, C.D., “VLSI and
given by the upper layer as a form of intelligent transducers”, Sensors and
are extremely high technology. Actuators, Vol. 2, 1982, pp. 239-50.
optimization algorithm. This approach requires VLSI
3. Brignell, J. and White, N., Intelligent
technology. Sensor Systems, Institute of Physics
Approaches to make sensor systems
(3) Integration with microcomputers Publishing, 1994, pp. 142-5, pp. 235-
intelligent 41.
There are three different approaches to or microprocessor. This method is
very much used and preferred 4. George, M., Hascard, M.R., Koh, S.N.
realize sensor intelligence: and Kong, R.Y., “Intelligent sensor
technique. The reason behind this
(1) The use of specific functional tendency is the availability of the with serial bus interface”, 7th
materials, so-called intelligent Australian Microelectronics Con-
systems that will be needed for ference Proceedings, 1988, pp. 127-34.
materials. Intelligent materials are design and implementation.This
vitally important for intelligent 5. Yamasaki, H. and Takahashi, K.,
method is most popular and it is “Advanced intelligent sensing system
sensor design. It needs a very deep usually represented as the integra- using sensor fusion”, Proceedings of
understanding of chemistry, tion of sensor devices and the 1992 International Conference on
physics and mathematical for- microprocessors. The algorithm is Industrial and Electronic Instrumenta-
mulation of the facts at hand. programmable and it can be tion and Automation, Cat. No.
Intelligent materials are called changed even after it is designed. 92CH3137-7, Vol. 1, 1993, pp. 1-8.
6. Hashimoto, Y., et al. (Ed.),
Measurements Techniques in Plant
Science, Academic Press, New York,
I S SC A/D µP NY, 1990, pp. 25-42.

Further reading
II S SC A/D µP Giachino, J.M., “Smart sensors”, Sensors
and Actuators, Vol. 10, 1986, pp. 239-48.
Polla, D.L., Muller, R.S. and White, R.W.,
“Integrated multisensor chip”, IEEE
Electron Device Letters, March 1986,
pp. 255-6.
S SC A/D µP Wise, K.D., “Integrated sensors: interfacing
electronics to a non-electronic world”,
Sensors and Actuators, Vol. 2, 1982,
pp. 229-37.
III S SC A/D µP

S µP E.T. Powner (Professor) and F. Yalcinkaya

are both based in the School of Engineering,
The University of Sussex, Falmer,
Smart sensor or intelligent sensor Brighton, UK.
Figure 8. Development stages and trends in integration with microprocessor

SR Vol. 15 No. 3, 1995 35

 This article has been cited by:

1. Kuncup Iswandy, Andreas KönigA Novel Adaptive Resource-Aware PNN Algorithm Based on Michigan-Nested Pittsburgh
PSO 477-484. [CrossRef]
Downloaded by KIRIKKALE UNIVERSITY At 12:40 08 March 2017 (PT)

View publication stats