Noise robust time of arrival estimation method using hierarchical Bayesian based compressed sensing algorithm

摘要

A microwave radar system is a useful tool for all-weather type remote sensing, such as terrain surface measurement. It is well known fact that a resolution of time-of-arrival (TOA) is strictly determined by the frequency bandwidth of transmitted signal. As super-resolution technique beyond such limitation, a compressed sensing (CS) algorithm has come under spotlight because a sparse assumption is well established in typical radar situations. However, the original CS method suffers from lower TOA resolution in insufficient SNR level. To address with such problem, this paper introduces a hierarchical Bayesian based CS algorithm. This method introduces the stochastic model derived from cross-correlation response as a priori information for CS reconstruction as hyper-prior distribution. The results of numerical simulation show that the proposed method enhances an accuracy for signal reconstruction, even in lower SNR situations.