Comparison of Ultra-wideband Radar Target Classification Methods Based on Complex Natural Resonances

The principal objective of automatic target recognition (ATR) is to identify the target from UWB radar returns. In intelligent vehicles, ATR can be a solution to identify pedestrians, vehicles, traffic signs, etc. Typically radar ATR relies on classifying the characteristics of the target obtained from the backscattered signal. In this paper, we apply Matrix Pencil Method (MPM) to extract complex natural resonances (CNRs) from the late time part of the backscattered field, which is characteristic of the studied target and can be used for its recognition and identification. Next, we compare three classification methods of CNRs: support vector machine (SVM), K-nearest neighbor (KNN) and Naive Bayes classification.