Enhancing VINS with Smart Feature Grading: Overcoming Cautious and Excessive Removal of Dynamic Features for Robust Urban Localization

Satellite navigation can help measure water vapor, but most stations lack weather sensors. This limits accurate real-time estimates. We developed a method using artificial intelligence weather models to calculate two key atmospheric quantities locally and quickly anywhere. Tests with three models gave water-vapor errors below one millimeter for five-day forecasts, better than current methods. This can improve real-time weather monitoring and forecasts when more satellite data are used.