Classification Of Non-Fluctuating Complex Targets Using Modwpt Feature Extraction

Abstract

The classification of complex targets is one of the challenging tasks in the radar research field. The Radar Cross-Section is
the prime attribute for the target classification. The RCS of the complex targets is an amalgamation of various complicated
shapes where computational electromagnetic approaches require tremendous computations. A dedicated approach for
modelling the complex targets with simple scatters like a sphere, circular cylinder, frustum, and circular disc are developed.
The developed model features are extracted using the MODWPT extraction method and are given to different classifiers
specifically SVM, k-NN, and ANN. The classification of the non-fluctuating complex targets by the classifier of the highest
accuracy is determined.