Machine Learning Ionosphere Delay Estimation for Satellite Positioning
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing satellite positioning systems, such as those described in Patent Document 1, do not effectively improve the accuracy of ionosphere delay and troposphere delay amounts, which are critical for precise positioning.
Innovation Solution
A positioning assistance apparatus and method that utilize machine learning-based ionosphere and troposphere delay models to estimate these delays and calculate their precision, thereby enhancing the accuracy of positioning computations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional satellite positioning methods are used, then positioning can be performed, but positioning accuracy is insufficient due to uncorrected ionosphere and troposphere delays
Solution Approach 1:
The patent introduces local correction information as an intermediary element that mediates between satellite positioning signals and the positioning receiver. This correction information, generated by a generation-side apparatus and transmitted to a use-side apparatus, compensates for ionosphere and troposphere delays without requiring the receiver to perform complex real-time calculations, thus improving accuracy while managing system complexity through centralized processing
Solution Approach 2:
The patent replaces traditional mechanical/geometric correction methods with data-based correction using locally generated correction information. Instead of relying solely on geometric calculations and standard models, the system uses actual measured delay values processed through machine learning models to generate correction data, substituting complex real-time computational mechanisms with pre-processed correction datasets
2Measurement precision
If machine learning-based ionosphere delay models are used, then ionosphere delay estimation accuracy is improved, but model training and computation requirements increase
Solution Approach 1:
The patent applies preliminary action by training machine learning models and generating local correction information in advance during system setup or in low-computation periods. The generation-side apparatus pre-processes positioning data, trains models on historical observations, and generates correction information that can be directly applied by the use-side apparatus without requiring intensive real-time computation, thus reducing energy consumption during actual positioning operations
Solution Approach 2:
The patent uses copying by creating a simplified representation of complex ionosphere and troposphere delay characteristics through machine learning models. Instead of directly computing complex atmospheric delays in real-time, the system copies the essential delay patterns into trained models that can quickly generate correction factors, reducing computational energy requirements while maintaining estimation accuracy
Data Source
AI summary
A positioning assistance apparatus 1 that improves the positioning accuracy includes an estimation unit 2 that estimates a delay amount (ionosphere delay amount or a troposphere delay amount) using a model generated through machine learning (an ionosphere delay model or a troposphere delay model) and a degree-of-precision calculation unit 3 that calculates a degree of precision with respect to a delay amount (an ionosphere delay amount or a troposphere delay amount) calculated through positioning computation, using the estimated delay amount.