Micro-Doppler analysis of vibrating targets in SAR

Abstract

Vibrating targets cause phase modulation of the azimuth phase history for a SAR system. The phase modulation may be seen as a time-dependent micro-Doppler frequency. It is useful to analyse such signals with Cohen's class time - frequency methods due to their superior resolution potential. The method must be chosen with care, as the effective time integration window of the kernel should not be longer than about a cycle of the modulation signal. Using the adaptive optimal kernel method (AOK), the authors obtained good results from a controlled experiment where two oscillating corner reflectors were placed within a SAR scene collected by the US Navy APY-6 radar. The oscillation frequency and amplitude were calculated from the time -frequency distributions and the results agree well with available ground truth. To resolve the instantaneous frequencies, the kernel time extent had to be sufficiently short. quadratic time - frequency methods. The methods are useful due to the high time -frequency resolution that may be obtained. We show that the method must be chosen with care if the aim is to reveal the time-dependent Doppler frequency caused by the phase modulation. Generally, the method must not use a time window longer than about one cycle of the modulation. With these considerations in mind, the adaptive optimal kernel (AOK) method has been applied to an APY-6 SAR collection containing two oscillating corner reflectors. The oscillations, although being only millimetres in amplitude, are clearly seen in the resulting time - frequency distributions. The results agree well with ground truth. 2 Theory A stationary point target within a SAR scene will trace out a unique phase history as the SAR antenna is moved relative to the scene (2). The phase history is related to the position of the point target, and the SAR processor estimates the position and complex amplitude from the given phase history. If the position of the point target changes during the time of data collection for the SAR image, a target motion- related phase term is introduced into the phase history. In this paper we consider the case of small sinusoidal motions.