Machine Learning Data Transmission for Network Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mobile network planning and resource scheduling methods, relying on manual experience or simple algorithms, face challenges in adapting to diverse service requirements such as ultra-high rates, low latency, high reliability, and massive connections, leading to inefficiencies and high costs due to their limited adaptability and theoretical performance constraints.
Innovation Solution
The implementation of machine learning in mobile networks, where terminal devices collect and transmit data for machine learning processes, enabling network devices to perform optimization and improve communication services through RAN intelligence by configuring terminal devices to collect specific data and reporting it in segments, allowing for better network service provision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If manual experience or simple algorithms are used for network planning and resource scheduling, then device complexity is reduced, but productivity and adaptability deteriorate due to high time consumption and poor adaptability to diverse service requirements
Solution Approach 1:
The patent introduces an intermediary system comprising a data collection module, data processing module, and machine learning model training module that acts as a mediator between raw network data and optimization decisions. This intermediary processing layer automatically collects, processes, and analyzes network data to generate optimization strategies, thereby reducing manual intervention complexity while significantly improving network optimization productivity and adaptability to diverse service requirements
Solution Approach 2:
The patent replaces manual mechanical network planning and scheduling methods with automated machine learning-based systems. By substituting human expert manual configuration with algorithm-driven automated optimization, the system reduces the complexity of network planning operations while dramatically improving productivity through faster, data-driven decision-making that adapts to changing network conditions and diverse service requirements
2Measurement precision
If machine learning models are trained with large amounts of raw data, then measurement precision is improved, but loss of time and energy increase due to data transmission and processing requirements
Solution Approach 1:
The patent segments the data processing workflow into distinct modular components: data collection, data processing, model training, and optimization implementation. Each segment handles specific tasks independently, allowing parallel processing and optimization of individual stages. This segmentation reduces overall processing time while maintaining comprehensive data analysis accuracy by ensuring each segment focuses on its specialized function
Solution Approach 2:
The patent performs preliminary data processing and feature extraction before model training, preparing data in advance to reduce processing time during critical inference phases. By pre-processing data to extract relevant features and remove redundancies beforehand, the system maintains high measurement precision while significantly reducing the time required for actual model training and deployment
3Adaptability or versatility
If comprehensive data collection is performed for machine learning training, then adaptability is improved, but loss of information and data management complexity increase
Solution Approach 1:
The patent extracts only the most relevant and useful features from comprehensive raw network data through automated feature selection and extraction processes. By identifying and extracting key parameters that directly impact network performance and service quality, the system maintains high adaptability to diverse services while reducing data management overhead by eliminating redundant and irrelevant information from the training dataset
Data Source
Figure 1~2
Figure 3~4
Figure 5~6
AI summary
Embodiments of this application provide a data transmission method and apparatus. The method includes: A terminal device sends a first data segment and auxiliary information to a first network device, where the first data segment is one of one or more data segments corresponding to data used for machine learning, and the auxiliary information indicates a target device of the first data segment. According to the method, the first network device can determine the target device based on the auxiliary information, so that massive data that is collected by the terminal device and that is used for machine learning is accurately transferred to a device that is in a network and that performs machine learning. This helps the network provide a better communication service based on a machine learning result, to improve network work efficiency.