Antenna Azimuth Correction Using Weighted Prediction Algorithms
Solution Overview
Problem
Current methods for optimizing antenna feeder systems in mobile communication networks are inefficient due to inaccuracies in antenna operation parameters, which are often not updated correctly and measured inadequately, leading to poor prediction models and high operational costs.
Innovation Solution
A method utilizing big data processing and machine learning algorithms to collect, analyze, and clean operation parameters of antennas, combining multiple prediction algorithms with a weight algorithm to accurately predict and rectify antenna azimuths, thereby improving network quality and maintenance efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional drive test methods are used to acquire mobile communication data, then the data acquisition process is simple and quick, but the accuracy of the acquired data is low and analysis result accuracy cannot be ensured
Solution Approach 1:
The patent replaces the mechanical drive test method with a data processing system that utilizes existing mobile communication data from user equipment. The system processes signaling messages and measurement reports from multiple base stations to acquire and analyze mobile communication data, eliminating the need for physical drive tests while improving data accuracy through comprehensive multi-source data collection and statistical analysis.
2Measurement precision
If a single prediction model is used for antenna azimuth prediction, then the model selection process is simple, but the prediction accuracy is insufficient when data characteristics vary across different areas
Solution Approach 1:
The patent implements a dynamic model selection mechanism that automatically chooses the most appropriate prediction model based on the characteristics of the input data. The system evaluates data patterns and selects from multiple available models (including linear regression, decision trees, and neural networks) to adapt to different scenarios, ensuring high prediction accuracy without requiring manual intervention for model selection.
Solution Approach 2:
The system changes the parameters and structure of prediction models based on data characteristics. Different models with varying complexity and parameter configurations are applied depending on the specific features of the input data, such as data density, spatial distribution, and signal characteristics, thereby optimizing prediction accuracy for different areas and conditions.
3Measurement precision
If comprehensive data collection and multiple prediction models are implemented, then prediction accuracy is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent segments the data processing task into distinct stages: data collection from multiple sources, preliminary filtering and validation, feature extraction, model selection, and prediction. This segmentation allows each stage to be optimized independently and enables parallel processing where applicable, reducing overall computational complexity while maintaining high prediction accuracy through systematic multi-stage analysis.
Data Source
AI summary
Disclosed is a method for implementing antenna azimuth correction based on user data. Five prediction algorithms are combined to avoid some defects of an individual algorithm. By means of a method for applying a plurality of prediction model weights, a plurality of algorithms is organically combined, a high weight is applied to an algorithm having a good prediction effect, and a low weight is applied to an relatively bad algorithm, such that not only the accuracy of data prediction but also the stability of data prediction for different regions are ensured. By drawing a grid-level user data RSRP heat map for analyzed data to obtain sampling point information around a cell and a predicted azimuth, the present invention can find out, by means of deeper analysis, problems in engineering parameters, such as a longitude/latitude error, an azimuth error, and a reversely connected cell.