Ground moving target classification based on micro-Doppler signature using novel spectral information features

A novel spectral-based feature set is proposed for ground moving target classification using a battlefield surveillance radar. Precise frequencies and coefficients are extracted from the backscattered radar signal which is used as an information in Automatic Target Recognition (ATR) system to solve the problem of ground moving target classification. The proposed ATR system is realized by combining Gaussian Mixture Model (GMM) with Maximum Likelihood (ML) rule. Experimental signals, realized with ground surveillance radar operating in centimetre range wavelengths, are used to test our ATR system. Numerical evaluation results are presented to assess the performance gain provided by the proposed method.