A First Course in FUZZY and NEURAL CONTROL

210 CHAPTER 6. NEURAL CONTROL

time=time+0.1;

end

We obtain a plot of the generated training pattern vector as shown in Figure
6.8.

Figure 6.8. Training pattern vector
We then compute and plot the gradient of the training pattern vector, shown
in Figure 6.9.

dp = gradient(p);

plot(dp,í*í);

Figure 6.9. Gradient of training pattern vector

The gradient of the pattern vector allowsus to develop a prototype proportional
gain function and the pattern vector itself can be used to develop a prototype
integral gain function. These are obtained inMatlabas follows:

for i=1:50

tprop(i)=10.0*abs(dp(i));

tint(i)=exp(-abs(p(i)));

end