Subspace-based Joint DOA, Delay and DFO Estimation for GNSS Multipath Signals
Xi Hong,Ning Chang,Wenjie Wang,Qinye Yin
DOI: https://doi.org/10.33012/2018.15992
2018-01-01
Abstract:Positioning performance of Global Navigation Satellite System (GNSS) is usually greatly influenced by multipath of satellite signals. With multi-antenna equipped, GNSS receiver can improve received signal strength (RSS) and mitigate the effects of interference and multipath, but some signal parameters, such as direction-of-arrival (DOA), time delay and doppler frequency offset (DFO), should be known in advance. However, they are hardly acquired fast and accurately due to the low RSS of the GNSS signals. In this paper, we propose a novel subspace-based method to estimate DOA, delay and DFO jointly. By removing sub-carrier from the received GNSS BOC signals, different sub-carrier phases are provided to give fine delay estimates and make some multipath rays incoherent in subspace processing. Forward and backward smooth (FBS) is also applied to build a spatial and spectral cross-correlation matrix with rotation invariance. The joint DOA and delay estimation for all rays is then gotten from its largest eigenvalues and corresponding eigenvectors respectively. A repetition property is then established based on the acquisition and the eigenvectors. With the help of the modulation of GNSS signals, a fine DFO estimate is given. The method in this paper expands the steering vectors in space and frequency, so it could have a higher spatial and temporal resolution in multipath environment. Simulation results show that, compared with the SAGE, the proposed method has a similar performance in light-of-sight (LOS) environment, but has much more accurate estimates for multipath rays with relatively short delay.