Nonparametric Discriminant Analysis Applied to the Extraction and Selection of Radar Target Signature Features

摘要

A nonparametric criterion of discriminant analysis is applied to a large scale data set, of radar returns from aircrafts, for frequency selection. The performance of the optimum sets characterized is evaluated using Monte-Carlo simulation. The effect of the number of frequencies and the a priori information of the azimuth angle on the performance of the pattern recognition system is examined. Also the effect of multiple measurements is examined.