Wireless Signal Propagation Prediction with Adaptive Sampling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless signal propagation prediction methods struggle to balance precision and operational efficiency, particularly in the context of 5G deployment, with drive test methods being costly and imprecise, and deterministic models requiring complex calculations and precise input data.
Innovation Solution
A method and apparatus that utilize spatial autocorrelation characteristics to select appropriate algorithms for wireless signal propagation modeling, including Kriging methods and machine learning, and adjust sampling points to improve model precision and accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If drive test mode is used for wireless signal propagation prediction, then coverage areas can be tested, but costs of manpower and material resources are high and estimation deviation exists
Solution Approach 1:
The patent uses machine learning models to create virtual copies of drive test data and propagate signal characteristics across the entire coverage area without physically traversing every location. The model learns from limited sampled points and generates predictions for all locations, replacing the need for exhaustive physical drive tests while maintaining prediction accuracy.
Solution Approach 2:
The patent performs preliminary machine learning model training using historical drive test data and environmental information before actual prediction tasks. This preliminary action creates a ready-to-use prediction system that can quickly generate results without requiring repeated expensive drive tests for each prediction scenario.
2Measurement precision
If deterministic model is used for wireless signal propagation prediction, then precision is slightly improved, but complex calculation process and strict requirement on calculation input restrict wide use
Solution Approach 1:
The patent transforms the deterministic modeling approach by changing parameters from requiring precise construction and restoration of three-dimensional building information to using machine learning models that process environmental data in a more flexible manner. The ML models accept various input formats and automatically learn the complex relationships, reducing the burden of precise input data preparation while maintaining or improving prediction precision.
3Productivity
If statistical model is used for wireless signal propagation prediction, then operation efficiency is improved, but large difference from actual measured signal exists
Solution Approach 1:
The patent creates a composite prediction system that combines machine learning models with environmental data processing capabilities. This composite approach integrates the efficiency of automated data processing with the precision of learned patterns from training data, producing predictions that are both operationally efficient and highly accurate compared to pure statistical models.
Data Source
AI summary
This application discloses a wireless signal propagation prediction method, and the method includes: obtaining S first sampling points in prediction space, and obtaining a first parameter of the prediction space through calculation based on location information of the S first sampling points and corresponding wireless signal received strength; obtaining a target algorithm based on the first parameter, and generating a wireless signal propagation model of the prediction space based on the target algorithm, the location information of the S first sampling points, and the corresponding wireless signal received strength; and obtaining wireless signal received strength of a terminal at any location in the prediction space based on the wireless signal propagation model of the prediction space. This application further discloses a wireless signal propagation prediction apparatus.