A fuzzy finite state machine implementation based on a neural fuzzy system

The outputs of a feedforward neural network depend on the present inputs only. Difficulties arise when a solution requires memory in such applications as speech processing, seismic signal processing, language processing, or spatiotemporal signal processing. For such applications, the outputs are not only functions of the present inputs but of the present states (or the past inputs and outputs) as well. Fuzzy finite state machines (FFSMs) can be effectively used in these applications. The aim of this study is to show that a FFSM can be realized using a neural fuzzy system (called NeuFuz). In a FFSM, the output and the next state depend on the input and the present state, which in turn is a function of the previous inputs. To accommodate the memory requirement, the feedforward structure of the neural fuzzy system is changed to a recurrent architecture by adding a feedback loop from the output layer to the input layer during the recall mode. The validity of the approach is verified with a temporal pattern-matching experiment