GNSS Localization Using SNR Data and 3D Maps

The patent introduces satellite SNR measurements as an intermediary indicator to infer whether the signal path is blocked or reflected. Instead of directly measuring position, the system uses SNR as a mediator to detect propagation conditions, which then informs the particle filter to weight or reject certain position hypotheses, thereby resolving the contradiction between obtaining position estimates and avoiding reflection errors

The patent replaces reliance on purely geometric time-of-flight measurements with a probabilistic framework that incorporates signal quality metrics (SNR). This substitution transforms the deterministic geometric positioning approach into a stochastic method that accounts for signal propagation uncertainties, enabling the system to handle reflected and blocked signals more robustly

The patent segments the continuous space into discrete particle hypotheses and processes each independently through parallel evaluation of SNR consistency with 3D map shadows. This segmentation allows the computationally intensive particle filter to be distributed across multiple particles, making the complex computation more manageable and enabling implementation on resource-constrained mobile devices

The patent implements a practical particle filter that uses a finite number of particles (excessive but not infinite) rather than exhaustive search. This partial action approach provides sufficient accuracy improvement while keeping computational complexity at acceptable levels for mobile deployment, balancing precision gains with device capabilities