Page 50 - Algorithm Collections for Digital Signal Processing Applications using MATLAB
P. 50

38                                                          Chapter 1

              Truncate the portions of the Zsmall, Zmedium and Zlarge sections of the
           output relationship above the corresponding thresholds obtained from output
           fuzzy membership as shown in the figure. Remaining portions are combined
           with overlapping of the individual sections. Combined sections can also be
           obtained by adding the individual sections instantaneously as illustrated in
           the example  given in the section 5.4 Obtain the crisp value Zout which
           divides the combined sections into two halves as shown in the figure. Thus
           crisp value Zout is obtained for the corresponding input crisp value Xinput
           and Yinput using Fuzzy logic system.

           5.3      Why Fuzzy Logic System?


           The Automatic system uses set of  rules framed with  the input  data  to
           decide the output data. For example the automatic system is designed to
           maintain the temperature of the water heater by  adjusting the heat knob
           based on the water level  at that particular instant and temperature at that
           particular instant. In this example water level measured using the sensor at
           that particular instant and temperature measured using the sensor at that
           particular instant are the input to the control system. Control system uses the
           set of rules framed as given below for deciding the position of the heat knob
           which is the output of the control system

           •  If T1<=Temperature <=T2 and W1<=Water level<=W2, Adjust the heat
              knob position to K1. (Say).
           •   If T0<=Temperature <=T1 and W0<=Water level<=W1, Adjust the heat
              knob position to K2. (Say) etc.

             Suppose if the temperature is just less than T1, this rule is not selected.
           The small variation ‘ΔT’ makes the automatic control system not to select
           this particular rule. To overcome this problem instead of using crisp value
           for framing the rule, fuzzy values can be used to frame the rules (i.e) without
           using the actual values obtained through input sensors This is called fuzzy
           base rule as given below

           •  If the temperature of the water is small and water level is high, place the
              heat knob in the high position. (say)
           •  If the temperature of the water is medium and water level is less, place
              the heat knob in the less position. (say) etc.
   45   46   47   48   49   50   51   52   53   54   55