Channel Estimation Assisted Bistatic Localization in Terrestrial and Non-Terrestrial Networks

Low-Earth-orbit (LEO) satellites are regarded as a key enabler for 6G communications and localization, due to their large coverage beyond conventional terrestrial networks. In this paper, we propose a downlink LEO base station (BS) bistatic localization framework relying on hybrid beamforming that alleviates the reliance on ultra-fine angle estimation by jointly exploiting time-frequency-spatial observations. A multiple-measurement-vector (MMV) based sparse model is constructed for attaining accurate channel gains and angles from limited pilots and moderate array sizes, where a modified block orthogonal matching pursuit (BOMP) algorithm is proposed to enhance robustness under highly correlated sensing matrices for localization purposes. After geometry-based timing-advance and Doppler pre-compensation at the BS, a two-dimensional (2D) upsampling matched filter having fine delay-Doppler grids is applied to estimate the residual time of arrival (ToA) and Doppler frequency. Then, the final user equipment (UE) position is obtained by intersecting the BS-centered angle-of-arrival (AoA) ray with a bistatic-range ellipse derived from the residual delay. The numerical results under realistic LEO-BS bistatic scenarios demonstrate that the proposed scheme achieves meter-level localization accuracy and highlight the performance gains attained by increasing the number of pilot symbols, subcarriers, and angular resolutions.