Joint time-frequency transform processing for linear and sinusoidal FM coherent ladars

摘要

We examine the signal processing of both linear and sinusoidal frequency modulation (FM) coherent ladar returns from resolved and unresolved targets, which are spread in Doppler. The Doppler spread may be due to target spin, tumbling, or vibration as well as to the applied linear or sinusoidal-FM on the transmitted E-field. Monte Carlo realizations of the target surface random phasor reflector elements interact with the incident Efield producing laser speckle, and the speckled returns are analyzed in this study. The speckle signals are processed (1) using several spectrum (periodogram based) estimators, (2) the conventional "spectrogram" approach, and (3) ten joint time-frequency transforms (JTFT). We show that the Born-Jordan JTFT is superior to the other spectral estimators tested here in suppressing local oscillator laser noise and accurately estimating the target's spectrum for signal processing under speckle target return conditions pertaining to coherent laser radar. A new algorithm which sums particular pixels of the JTFT image is introduced and is shown to be much more robust in low CNR conditions than the JTFT maxima or JTFT centroid processing when utilizing the applied linear or sinusoidal-FM modulation waveform.