Terminal Positioning Label Correction for AI/ML Model Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing AI/ML models for terminal device positioning suffer from low accuracy due to unreliable label information, especially in scenarios with large Non-Line of Sight diameters, leading to positioning errors greater than 10 meters.
Innovation Solution
A method for acquiring label information involves a first network device determining a correction parameter to refine the label information of a data set, using actual information from terminal devices to correct initial label information, thereby improving the reliability of the data set for training an AI/ML model.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional positioning methods are used, then the system is simple to implement, but the positioning accuracy deteriorates with large Non-Line of Sight diameters
Solution Approach 1:
The patent applies preliminary action by pre-collecting actual positioning information from terminal devices and pre-calculating correction parameters before the AI/ML model training process. The network device obtains actual position information through traditional positioning methods or GPS, then calculates correction parameters in advance to improve the reliability of label information for model training, thereby enhancing positioning accuracy without increasing real-time system complexity
Solution Approach 2:
The patent introduces an intermediary element - the correction parameter - that mediates between traditional positioning methods and AI/ML model training. The correction parameter acts as a bridge to refine label information by comparing initial label information with actual positioning information, thereby improving positioning accuracy without requiring fundamental changes to the system architecture
2Measurement precision
If label information is not corrected, then the data processing is simple, but the positioning accuracy deteriorates due to unreliable label information
Solution Approach 1:
The patent implements feedback by using actual positioning information to calculate correction parameters that refine label information. The network device obtains actual position information, compares it with initial label information, and uses the difference as feedback to generate correction parameters. This feedback mechanism improves the reliability of label information for AI/ML model training, thereby enhancing positioning accuracy
Solution Approach 2:
The patent applies parameter changes by modifying label information through correction parameters. The correction parameters adjust the initial label information based on the difference between initial and actual positioning information. This parameter transformation improves the accuracy of label information without requiring complex reprocessing of the entire dataset
Data Source
AI summary
A method for acquiring label information and a terminal device are provided. The method includes following operations. The first network device determines a correction parameter. The first network device determines label information corresponding to each data in the data set based on the correction parameter. The data set and the label information corresponding to each data in the data set are used to train a first model.