Cyclostationary Analysis of Micro-Doppler Signature for ISAR Ship Classification

This paper presents an approach to inverse synthetic aperture radar (ISAR) ship classification by exploiting the cyclostationary properties of micro-Doppler (m-D) signatures from rotating shipborne components. Unlike traditional methods that attempt to eliminate m-D effects as noise, our approach leverages these signals to enhance target classification. By analyzing the cyclic spectral density (CSD) of radar echoes, we estimate key features of rotating parts such as frequency and radius before full ISAR image formation. We show cyclostationary analysis outperforms traditional m-D analysis methods on weak reflectors in noisy environments. We demonstrate both effectiveness and feasibility of this approach using efficient algorithms, showing promising results in lab experiments.