Two-Stage Spatial Whitening and Normalized MUSIC for Robust DOA Estimation of GNSS Signals under Jamming
Chuanrui Wang,Xiaowei Cui,Gang Liu,Mingquan Lu
DOI: https://doi.org/10.1109/taes.2024.3472629
IF: 3.491
2024-01-01
IEEE Transactions on Aerospace and Electronic Systems
Abstract:The estimation of the direction of arrival (DOA) of navigation signals is a critical function of a global navigation satellite system (GNSS) array receiver when used for applications such as attitude determination, spoofing detection, and blind adaptive beamforming. However, the presence of high-power GNSS jamming poses significant challenges to accurate DOA estimation. In this paper, we present spatial whitening and normalized multiple signal classification (SWN-MUSIC), an enhanced two-stage MUSIC algorithm for DOA estimation of GNSS signals under jamming. In the first stage, inspired by minimum variance distortionless response (MVDR) beamforming, we employ a spatial whitening matrix for matrix filtering to achieve jamming suppression. In the second stage of the two-stage algorithm, an approach is proposed to account for the effect of matrix filtering when computing the MUSIC pseudospectrum. Compared to previous works on GNSS DOA estimation under jamming, our SWN-MUSIC algorithm exhibits improved accuracy, particularly in challenging scenarios where the DOA of jamming and signals are closely aligned. Additionally, we develop the reduced-dimension SWN-MUSIC (RD-SWN-MUSIC) algorithm to reduce the complexity of applying SWN-MUSIC for two-dimensional DOA estimation using uniform rectangular arrays (URA) under jamming. Compared to standard reduced-dimension MUSIC techniques, in the new RD-SWN-MUSIC, we not only successfully apply dimension-reduced parameter estimation to the more complex pseudospectrum form of SWN-MUSIC, but also avoid unnecessary precision loss by implementing a strategy that conducts separate one-dimensional parameter searches for each DOA parameter. Presented results of numerical simulations in jamming scenarios indicate that the performance of the proposed RD-SWN-MUSIC algorithm closely approaches the corresponding Cramer-Rao lower bound (CRLB). The complexity of the algorithm is significantly lower when compared to the classical MUSIC and its previously proposed versions.