Resilient Quadratic Time-Frequency Distribution for FM Signals with Gapped Missing Data

摘要

Frequency modulated (FM) signals are frequently encountered in radar signal processing. In practice, such signals may experience distortions due to target radar cross section fluctuation, propagation fading, and line-of-sight obstruction. Such distortions make time-frequency presentations and analyses challenging. The problem becomes even more difficult for data with gapped missing samples, i.e., data samples are missing consecutively. In this paper, we address this problem in the context of resilient quadratic time-frequency distributions, which are achieved by applying robust spectrum analysis techniques to the instantaneous autocorrelation functions. In addition, utilization of compressive sensing and sparse reconstruction techniques, in lieu of the conventional Fourier transform, further improves the quality of the estimated time-varying instantaneous frequency signatures.