Ground Surveillance Radar Target Classification Based On Fuzzy Logic Approach

The fuzzy logic approach to the automatic classification of moving target detected by ground surveillance radar is presented in this paper. The real audio Doppler signatures of various targets are analyzed by spectrogram. As a result of analysis, input and output variables with corresponding membership function are defined. The set of fuzzy rules is established. The defuzzification of the output fuzzy set is performed by computing the "fuzzy centroid". The three target classes (walking man, running man and wheeled vehicle) are successfully classified.