V THE HYBRID NEURAL NETWORK/FUZZY LOGIC
MODEL
The proposed hybrid vector control model is shown in
fig. 7. For a varying command speed with a predictable
trajectory, the switch S switches to the neural network
model. For a constant speed operation, S switches to the
fuzzy logic model.
A trained neural network model will be able to
anticipate the optimum output that follows if the
command speed varies according to a predetermined
pattern. Meanwhile a fuzzy model can adjust better to the
unpredictable disturbances that occur during the period of
constant speed operation. By switching to the appropriate
models by referring to the command speed requirements,
a control model that combines the benefits of both models
can be realized.