On Estimating Nonlinear Frequency Modulated Radar Signals in Low SNR Environments

Abstract

Estimating the signal transmitted by a noncooperating radar is of great use in radar countermeasures. In this article, we propose to improve the noise resilience of the radar signal estimation by constraining the estimate to be in a low-dimensional signal space. In particular, we focus on estimating nonlinear frequency modulated radar signals that have good target resolution characteristics. A flexible 3-D odd polynomial frequency signal model is introduced as the low-dimensional signal space to compute accurate estimates at low signal-to-noise ratios (SNRs) with small number of intercepted signals. The quality of the low-dimensional signal estimates is compared with that of the estimate from the unconstrained high-dimensional space computed using principal component analysis estimator, which is the method in predominant use in the multisensor scenario. It is shown that signal estimation through low-dimensional signal spaces yields more accurate estimates at low SNRs with a smaller number of sensors compared to searching in the unconstrained high-dimensional signal space.