Radar Target Identification Using Multiclass Sparse Centroids
Ismail I. Jouny · 2023
A Multiclass classifier based on tree of binary classifiers is used to identify radar targets. The binary classifier is a sparse centroid based system. The radar system is a stepped frequency radar using high range resolution profile data. The sparse centroid classifier (developed by Calafiore et al [1]) accomplishes feature selection and recognition simultaneously and is amenable to real time implementation. The system is tested on real radar data in different scenarios and the classification performance is compared to that of an optimal classifier.