Abstract: In this paper, the automaton that is to base on the combination of fuzzy processing and neural network is constructed successfully. At the same time, fuzzy geometrical features are processed by using fuzzy technology from images of unprocessed electromagnetic wave (simplifying wave of signal), and the wave of signal is classified and recognized. The fuzzy geometrical features are found to be the best for recognizing distorted overlapping waves of signal when Fuzzy Automata (FA) are used. The overall performance is best of FA.