ML Channel Quality Prediction for 5G Spectral Efficiency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current wireless communication systems face challenges in efficiently estimating channel quality status (CQS) information, particularly in 5G networks, which leads to increased computational burden on user equipment (UE) and reduced spectral efficiency.
Innovation Solution
The proposed method employs machine learning, specifically reinforcement learning-based neural networks, to intelligently predict channel quality indicators (CQI), pre-coding matrix indices (PMI), and rank indices (RI) for secondary frequency bands without the need for reference signals, leveraging reports from primary frequency bands.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine learning-based prediction is used to estimate CQS information, then computational burden on UE is reduced and spectral efficiency is improved, but prediction accuracy may be compromised compared to traditional reference signal-based estimation
Solution Approach 1:
The patent introduces a machine learning model as an intermediary between reference signals and CQS information. The model processes reference signal data and network parameters to predict CQS values, acting as a mediator that translates raw signal data into quality metrics without requiring direct UE computation of all CQS parameters from reference signals.
Solution Approach 2:
The system performs preliminary processing of reference signals and network parameters through the machine learning model before final CQS determination. By pre-computing predictions based on historical data and current network state, the system prepares CQS estimates in advance, reducing the need for real-time UE computation and enabling faster scheduling decisions.
2Measurement precision
If traditional reference signal-based CQS estimation is used, then measurement accuracy is maintained, but computational burden on UE increases and latency increases
Solution Approach 1:
The patent replaces the mechanical computation process performed by UE with a machine learning-based prediction system. Instead of UE directly calculating CQS from reference signals through complex algorithms, the system uses a trained model to predict CQS values, substituting computational mechanics with intelligent prediction that requires minimal real-time processing.
Solution Approach 2:
The machine learning model creates a copy or approximation of the CQS estimation function. Rather than performing full reference signal processing, the model replicates the essential CQS determination capability through learned patterns from training data, providing accurate estimates without the computational overhead of traditional methods.
3Reliability
If continuous reference signal reporting is implemented, then CQS information accuracy is maintained, but device complexity and energy consumption increase
Solution Approach 1:
The system applies partial action by using the machine learning model to predict only the necessary CQS information based on available reference signals and network parameters. Instead of requiring complete reference signal processing and continuous reporting, the model generates sufficient CQS estimates from partial inputs, reducing UE processing requirements while maintaining adequate accuracy for scheduling decisions.
Data Source
AI summary
The disclosure relates to a communication method and system for converging a fifth-generation (5G) communication system for supporting higher data rates beyond a fourth-generation (4G) system with a technology for Internet of Things (IoT). The disclosure may be applied to intelligent services based on the 5G communication technology and the IoT-related technology, such as a smart home, a smart building, a smart city, a smart car, a connected car, health care, digital education, smart retail, security and safety services. A method for predicting channel quality status (CQS) in a wireless network by a base station (BS) is provided. The method includes receiving channel quality indicator (CQI) reports, pre-coding matrix index (PMI) reports, and rank index (RI) reports of first frequency band of a plurality of frequency bands in the wireless network, storing the CQI reports, the PMI reports, and the RI reports of the first frequency band, and predicting the CQI, the PMI and the RI of a second frequency band of the plurality of frequency bands associated with the UE based on the CQI reports, the PMI reports, and the RI reports of the first frequency band.