250 CHAPTER 8. APPLICATIONS
are many applications to the wide area of pattern recognition, and applications
to areas that involve processing of fuzzy information, such as medicine and eco-
nomics. When fuzzy systems are developed to solve appropriate problems, their
typical characteristics include rapid and smooth responses.
Neural networks are more suited than fuzzy logic systems to problems that
require a large number of input variables, for example, too many for a rule-based
fuzzy system to be feasible. On the other hand, they require large data sets for
training. Neural networks are widely used. We mention a few general areas:
telecommunication, pattern recognition, quality control,financial and economic
forecasting, speech recognition, and many automotive applications.
Neural networks can be used to design membership functions for fuzzy logic
systems. Designing appropriate membership functions is at the heart of fuzzy
applications. Neural networks can automate the process, saving time and sim-
plifying the tuning process. Many applications are to consumer products such
as washing machines, vacuum cleaners, photocopiers, and air conditioning sys-
tems. They are also applied in thefinancial industry. Below, we illustrate two
studies of fuzzy control for real-world applications.
8.2 Cooling scheme for laser materials
In this section, we present a fuzzy logic-based approach to the cooling of laser
materials developed by N.S. Prasad and N.R. Prasad [58]. The controller design
is based on the performance characteristics of commercially available thermo-
electric coolers. Simulation results are presented and discussed and the feasibil-
ity of implementing such a controller is evaluated.
Solid-state laser materialsare susceptible to heating effects caused by opti-
cal pumping. The medium basically gets heated up due to the absorption of
the pump radiation and non-radiative decay processes. The heating and subse-
quent cooling gives rise to thermal effects. Thermal gradients result in stresses
and strains that produce variations in refractive index, leading to thermo-optic
distortions. The distortions lead to self-focusing, depolarization, and other un-
desired effects that is deleterious to the lasing medium and hence the laser
performance. There is extensive on-going research to characterize thermal be-
havior and to use the results for the design and development of various cooling
schemes to minimize detrimental effects [13, 34, 48].
Conventional cooling techniques include passive cooling by surrounding air,
liquid cooling, micro-channel cooling, and thermoelectric cooling [12, 38]. Ther-
moelectric coolers (TECs) are current operated devices that can provide preci-
sion cooling. They are easy to implement and operate and have the additional
advantages of size and weight. The theory and operation of thermoelectric cool-
ers is widely reported in the literature and they are commercially available [44].
Advances in semiconductor diode-laser pumped solid-state laser systems have
necessitated the use of more efficient and miniaturized cooling schemes. As a
consequence, TECs are becoming attractive in view of space constraints and
operational elegance.