GNSS Elevation Mask Learning for Satellite Visibility Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing positioning methods in 5G wireless communication systems face challenges in accurately determining the line-of-sight and non-line-of-sight visibility of satellite vehicles due to obstructions, which affects the precision of Global Navigation Satellite System (GNSS) positioning, and this is not efficiently addressed by conventional 2-D and 3-D map-based approaches.
Innovation Solution
The development of a geography-dependent elevation mask is learned through crowdsourced satellite vehicle signal measurements to predict GNSS positioning quality by identifying and correcting erroneous position estimates, using ephemeris data to determine satellite visibility angles and improve positioning solutions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 2-D and 3-D map-based approaches are used to determine satellite visibility, then positioning methods can be implemented, but positioning accuracy deteriorates due to obstructions and inability to accurately predict line-of-sight and non-line-of-sight conditions
Solution Approach 1:
The patent replaces traditional mechanical map-based visibility determination with a machine learning model that uses crowdsourced signal strength measurements. The ML model processes ephemeris data and signal strength observations to predict satellite visibility and GNSS positioning quality, overcoming the limitations of conventional 2-D/3-D map approaches in accurately determining line-of-sight and non-line-of-sight conditions.
Solution Approach 2:
The system implements feedback by using crowdsourced signal strength measurements from multiple user equipments to continuously train and improve the machine learning model. The model leverages feedback from actual signal observations to refine its predictions of satellite visibility and positioning quality, enabling adaptive improvement over time.
2Device complexity
If conventional map-based approaches are used, then implementation is simpler, but positioning precision deteriorates due to obstructions affecting GNSS signals
Solution Approach 1:
The patent substitutes complex mechanical map-based visibility analysis with a data-driven machine learning system that processes ephemeris data and signal strength measurements. This replacement achieves superior positioning precision by learning patterns from actual signal observations rather than relying on simplified geometric models that fail to account for real-world obstructions.
Solution Approach 2:
The system transforms the problem from geometric visibility calculation to signal strength-based prediction. By changing the parameter from binary line-of-sight/determination to continuous signal strength measurements, the system can more precisely characterize GNSS signal conditions and improve positioning precision while maintaining manageable implementation complexity through ML algorithms.
3Measurement precision
If ephemeris data and signal strength measurements are processed through machine learning, then positioning accuracy improves, but processing complexity increases
Solution Approach 1:
The machine learning model performs self-service by automatically learning from crowdsourced signal strength measurements and improving its own predictions. The system uses the observations from multiple user equipments to train and refine the model, eliminating the need for complex manual configuration or external intervention while achieving high positioning accuracy.
Solution Approach 2:
The ML model serves multiple functions simultaneously: it predicts satellite visibility, determines GNSS positioning quality, identifies line-of-sight and non-line-of-sight conditions, and provides corrections to position estimates. This multi-functionality consolidates what would otherwise require multiple separate processing systems into a single unified model, managing complexity while improving accuracy.
Data Source
AI summary
A method for use in positioning of a user equipment, the method including: acquiring, at the user equipment, a first satellite vehicle signal from a first satellite vehicle of a plurality of satellite vehicles; obtaining, at the user equipment, ephemeris data from the first satellite vehicle signal; and providing to an entity, based at least on the ephemeris data, a plurality of elevation mask indications corresponding to a location of the user equipment and each indicating an azimuth direction and an elevation angle range, if any, of potential satellite vehicle visibility in the indicated azimuth direction.