Improved micro-Doppler features extraction using Smoothed-Pseudo Wigner-Ville Distribution

Abstract

Time-frequency representations are commonly used to analyse time-varying spectral density of a time-varying signal. In particular, they can be used to extract features (e.g. blade length, number of blades and rotation rate of rotor) from micro-Doppler signals to provide means to differentiate between the different types of mini-UAV. This paper highlights the exploration of Smoothed-Pseudo Wigner-Viller Distribution (SPWVD) as compared to Short-Time Fourier Transform (STFT) to obtain better resolution in micro-Doppler features extraction via Singular Value Decomposition and Cepstral Analysis. Results have shown that with the proposed method of SPWVD Pre-Window whereby a window is applied to the raw signal before performing SPWVD, has resulted in better resolution in the estimation of blade length and the number of blade flash frequency components. The evaluation of the STFT and SPWVD Pre-Window have been conducted with real UAV data.