Station precipitation inversion method and system based on communication and dual-frequency GNSS signals
By using a dual-branch MLP inversion model based on multi-source data acquisition and heterogeneous data decoupling architecture using communication and dual-frequency GNSS signals, the real-time and accuracy problems of existing precipitation monitoring technologies in high-density, minute-level, and end-side localized applications are solved, achieving low-cost, reliable minute-level precipitation monitoring with high-density coverage capability.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- HUBEI PROVINCIAL METEOROLOGICAL SERVICE CENT (HUBEI PROVINCIAL PROFESSIONAL METEOROLOGICAL SERVICE DESK)
- Filing Date
- 2026-04-13
- Publication Date
- 2026-07-17
AI Technical Summary
Existing precipitation monitoring technologies suffer from problems such as poor real-time performance, poor model universality and adaptability, difficulty in maintaining accuracy, insufficient fusion of heterogeneous data, and inadequate hierarchical collaborative processing when applied in high-density, minute-level, and localized end-point applications.
By adopting multi-source data acquisition and preprocessing based on communication and dual-frequency GNSS signals, a dual-branch MLP inversion model with heterogeneous data decoupling architecture is constructed. Through feature filtering and normalization processing, a quantization model adapted to a general-purpose NPU is generated, realizing real-time inversion at the end side and collaborative quality control at the central station, forming a fully closed-loop inversion system at the end side.
It enables minute-level stress-free continuous inversion on low-cost, low-power edge devices, improves monitoring accuracy to the millimeter level (RMSE≈3.42 mm/h), ensures reliable operation in areas without or with weak networks, reduces single-point deployment costs, and has high-density, large-scale coverage capabilities.
Smart Images
Figure CN122412906A_ABST