Implementing fuzzy logic control with a biologically plausible neural net

Abeles (1982) and Alkon et al. (1989) stressed the fact that neurons possess an average firing rate, and may fire both above or below this average. The equations developed by the author (1991) (called the RX equations) include an average firing rate. The RX equations have been shown to be useful in solving elementary control problems. In fact, control, on a par with a fuzzy logic controller, may be achieved by a network of as few as three neurons, two input and one output-a three-neuron controller (TNC). In this paper we discuss the analogies between fuzzy logic control and control exercised by a TNC and the restrictions existing in application of the TNC technique.