The BeiDou Navigation Satellite System (BDS) features a heterogeneous constellation, which poses challenges in accurately characterizing the spatiotemporal correlations of multipath effects and effectively mitigating these effects using a single model. The widely used sidereal filtering (SF) and multipath hemispherical map (MHM) models cannot simultaneously reveal the spatiotemporal correlations of multipath. Therefore, we propose a satellite-dependent multipath model based on spatiotemporal correlations to mitigate multipath effects in BDS precise point positioning (PPP). Compared with the high-reliability MHM model with quality control and high-precision SF model with denoising, the static PPP positioning accuracy in the three-dimensional direction using the proposed model can be improved by up to 17.5 and 7.2%, respectively, and 9.3 and 3.5% improvements can be obtained in the kinematic PPP, respectively. For the convergence time of MHM and SF models, up to 0.5 and 49.2 min can be shortened in the static PPP using the proposed model. The proposed model not only exhibits superior accuracy in comparison to the high-reliability MHM and high-precision SF models but also has a relatively short convergence time.

北斗卫星导航系统（BDS）具有异构星座特征，这对准确表征多径效应的时空相关性并使用单一模型有效减轻这些效应提出了挑战。广泛使用的恒星滤波（SF）和多径半球图（MHM）模型不能同时揭示多径的时空相关性。因此，我们提出了一种基于时空相关性的卫星相关多径模型，以减轻北斗精密单点定位（PPP）中的多径效应。与带质量控制的高可靠MHM模型和带去噪的高精度SF模型相比，采用该模型的三维方向静态PPP定位精度可分别提高17.5%、7.2%和9.3%。运动学 PPP 可以分别获得 3.5% 和 3.5% 的改进。对于 MHM 和 SF 模型的收敛时间，使用所提出的模型在静态 PPP 中可以缩短最多 0.5 分钟和 49.2 分钟。与高可靠性 MHM 和高精度 SF 模型相比，所提出的模型不仅具有更高的精度，而且具有相对较短的收敛时间。