Manufacturing Automation is a key technology necessary for improving the quality and quantity
of manufactured products.This technology is specially important in developing countries due to the shortage of skilled workers in those countries.Several automatic techniques are available for detecting faults in different machine
components and locating the immediate faulty areas.Computer Numerical Control(CNC)machines,robots,surface milling machines,lathes and compressors
are among hundreds of widely used machines in manufacturing.Every one of these tools is
designed with the idea of repetition in mind.On an assembly line,each component repeats a task
over and over again,most of the time without a break.The cost associated with manufacturing a
piece is multiplied by hundreds and thousand of pieces manufactured at site.These pieces could be
related to automobiles,refrigerators,toys,pharmaceuticals,clothing,computers etc.Precision has
found more and more demand everyday for the past decade.Due to great amounts of competition in
the line of manufacturing,unlike the past,only companies with the best products survive.
Due to the repetitive nature of manufacturing,most people working in these sites become more
and more valuable for the creation of their product.This is due to the fact that they gain experience
and their productivity increases with their experience.Unfortunately, one cannot say the same thing
about most machines on the assembly line.They start wearing out and they hardly learn anything
from their past experiences.In an optimal manufacturing process,it is desirable for the system to learn and improve its performance
with experience.It is also desirable that any components of the system which are no longer
performing optimally should be found and reported to the human supervisor in an automatic manner
so that they could be replaced with a new piece.
Controls In Manufacturing
A wide class of controllers in this employ predefined gains and do not take into consideration
the nonlinear dynamics in these machines. [1, 2] These gains are based on linear approximations of
these highly nonlinear systems and are tuned to different tasks manually.These tuning jobs usually
take hours and sometimes days and during this time the machines are not operable.The result is that
these machines are not utilized to their full potential in terms of speed and precision.With a more
sophisticated control strategy,it is possible to compensate for the complicated effects of nonlinearities which have in the past been considered as disturbances in most systems.Two classes of newly developed control systems which are mostly geared toward manufacturing
applications are called Repetitive and Learning control systems.A repetitive controller [15, 16] is
designed for processes which operate in cycles.Repetitive controllers assume that there is continuity
between the last point of a repetition and the first point of the next repetition.In Learning Controllers the initial conditions are reset to the
same value at each repetition.The Learning Self-tuning Regulator has also shown both in computer simulations and in experimental
setups to be highly robust and to reduce the total error of systems by considerable amounts.
Another important need of manufacturing systems is to be able to monitor different components
automatically.Lots of research has been done in this area in the past decade.The health of components
such as bearings could be monitored by placing inexpensive accelerometer on the body of
the machine.Similar type of research has been done for monitoring compressors which could seriously malfunction
and blow up in cases.Automatic monitoring systems could shut these systems down before they could
cause any danger.In such cases,lives might be saved using these automatic monitors.
Using these new technologies in control and monitoring of manufacturing processes could be very
practical and valuable.Using these techniques,better quality products could be manufactured in
addition to the increased speeds of production.The down-time of manufacturing processes is reduced
extensively using condition monitoring techniques and the expertise of human workers could be used
in much more useful ways.Perfect integration of the above techniques could amount to an optimal
manufacturing process.










