Micro‐Doppler feature extraction for wideband imaging radar based on complex image orthogonal matching pursuit decomposition

Abstract

The micro‐Doppler (m‐D) features induced by targets with micro‐motions provide important information for automatic radar target recognition. In this study, the m‐D effect induced by rotational micro‐motion in wideband radar is analysed, and an algorithm for the extraction of these m‐D features is proposed. By making use of the amplitude and phase information of ‘range‐slow‐time image’, a dictionary with m‐D signal atoms is constructed in the complex image space. The orthogonal matching pursuit algorithm in vector space is then extended to the complex image space to decompose the range‐slow‐time image and to extract the m‐D features of the target. The proposed algorithm can extract the m‐D features in the presence of migration through range cells of micro‐motional scatterers, and can also work well when the sampling rate in slow‐time domain is lower than the Nyquist sampling rate. Simulations are given to validate the effectiveness and robustness of the proposed method.