Fuzzy logic controller with neural network

摘要

The fuzzy logic controller (1) has measured process values and inputs (x1-xn) that are received by an input neuron layer (E). This layer connects with an intermediate layer (A) where the input values are fuzzified using inference rules. A further neuron layer (MAX) provides values for an output layer (B) to generate control values (U1-Un). This layer provides a de-fuzzification of the values and the system uses a learning process. The output layer employs a priority based evaluation.