Multipath signals formed by signal reflection coming from objects in the vicinity of Global Navigation Satellite System (GNSS) receivers result in a degradation of the tracking performance and an increase in the positioning error. By estimating the parameters of both line-of-sight signal and the multipath signals, superior multipath mitigation, spoofing suppression, and localization can be attained. We propose using the multiple sparse Bayesian learning method together with the joint angle and delay estimation technique in GNSS multipath environment to fully exploit the sparsity present in both the spatial and the temporal domains. We also extend the techniques to the estimation of fractional Doppler frequency besides the angle and delay. To counteract the intrinsic drawbacks of sparse representations, two different algorithms based on on-grid and off-grid estimators are proposed to either reduce the complexity or enhance the resolution such that the proposed multipath mitigation approach can be adapted to various GNSS practical situations. Subsequently, a third algorithm with improved resolution is obtained by applying the Space Alternating Generalized Expectation–Maximization algorithm to refine the MSBL-based joint angle and delay estimates. Simulation results indicate that the three proposed algorithms can effectively resolve the GNSS multipath signals and have better performance than existing methods even in severe situations, like the cases of signals with low carrier-to-noise-power-density ratio and spatially and temporally correlated multipath.

由来自全球导航卫星系统（GNSS）接收器附近物体的信号反射形成的多径信号会导致跟踪性能下降并增加定位误差。通过估计视线信号和多径信号的参数，可以实现出色的多径缓解，欺骗抑制和定位。我们建议在GNSS多路径环境中使用多重稀疏贝叶斯学习方法以及联合角度和延迟估计技术来充分利用时空域中的稀疏性。除了角度和延迟之外，我们还将技术扩展到分数多普勒频率的估计。为了解决稀疏表示的固有缺点，提出了两种基于并网估计和离网估计的不同算法，以降低复杂度或提高分辨率，从而使所提出的多径缓解方法可以适应各种GNSS实际情况。随后，通过应用空间交替广义期望最大化算法来改进基于MSBL的关节角度和延迟估计，从而获得分辨率提高的第三种算法。仿真结果表明，所提出的三种算法可以有效地解决GNSS多径信号，即使在恶劣的情况下，如载噪功率密度比低，时空相关的多径信号，也能比现有方法具有更好的性能。 。