Passive low SNR tracking by spatial-temporal fusion sliding-window Radon transforms

Abstract

In the presence of multi-path propagation, attenuation of signal spectral components and target motion, the estimates of time delay between random wideband signals in low signal-to-noise (SNR) scenarios are characterized by a high probability of large errors. This results in increase the number of sensors to obtain desired performance levels (and therefore results in a series of further problems of managing the sensors). In this paper we propose a novel approach to time delay estimation (TDE) based on spatial-temporal fusion of local Radon transforms (LRD) in a sliding window, which significantly decreases probability of large errors of TDE. The properties of the cross-correlation function (CCF) in low SNR scenarios for linear and nonlinear filtering of sequences of CCF outputs are considered. The Efficiency of the proposed method in terms of normal error variance and the probability of large errors of TDE using numerically simulated data is discussed.