SUPERVISED CLASSIFICATION OF RADAR TARGETS USING THE MOMENTS SPACE

Abstract

The computation of statistical moments of the radar echo-signals with the objective of translating the decision to the moments space has shown potential in recent work. However, the moments space has not been considered for multiple targets classification, nor its combination with supervised classifiers has been studied. This paper proposes the use of moments as input features for several supervised classifiers and evaluates their performance. Among the considered methods are the Bayesian classifier, k-nearest neighbors, Support Vector Machines and artificial neural networks. The results show the usefulness of the moments space for classifying radar targets with high accuracy, precision and low complexity.