UE Beam Switch Neural Network Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for detecting beam failure and determining new candidate beams in 5G communication systems do not adequately consider all factors in complex cellular environments, leading to inefficiencies due to assumptions based on Block Error Rate (BLER) thresholds that are affected by estimation noises and interference.
Innovation Solution
Implementing machine learning-based systems and methods where a terminal device uses a neural network to measure and analyze beam performance metrics such as RSRP, RSRQ, SINR, and BLER, allowing for more accurate beam switch calculations and reporting to the base station for improved beam management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional BLER threshold methods are used for beam failure detection, then the detection process is simple, but the accuracy is poor due to estimation noises and interference
Solution Approach 1:
The patent transforms the beam failure detection problem from using a single BLER parameter to using multiple parameters including RSRP, RSRQ, SINR, and BLER. This multi-parameter approach allows the system to capture more aspects of channel quality and make more accurate beam failure detection decisions, resolving the contradiction between simplicity and accuracy.
Solution Approach 2:
The patent introduces a machine learning model as an intermediary between raw measurement data and beam failure detection decisions. The model processes multiple input parameters (RSRP, RSRQ, SINR, BLER) and produces a comprehensive beam quality assessment, enabling accurate detection while maintaining system manageability through the structured ML approach.
2Adaptability or versatility
If conventional link recovery function is used, then the implementation is straightforward, but it does not consider all factors in complicated cellular communication environments
Solution Approach 1:
The patent creates a universal beam management framework that handles multiple functions including beam failure detection, candidate beam identification, and beam switch decisions within a single machine learning-based system. This multi-functional approach enables the system to adapt to various cellular environments while maintaining a unified implementation structure.
Solution Approach 2:
The patent performs preliminary measurements of multiple parameters (RSRP, RSRQ, SINR, BLER) before making beam failure detection decisions. By gathering comprehensive data in advance and processing it through the machine learning model, the system can make more informed decisions that account for all relevant environmental factors.
3Measurement precision
If machine learning based beam switch calculation is implemented, then the accuracy of beam switch is improved, but the device complexity increases
Solution Approach 1:
The patent implements a self-service machine learning model that runs on the UE device itself, enabling the terminal to autonomously perform beam failure detection and beam switch decisions using locally processed measurements. This self-service approach improves beam switch accuracy by making decisions at the edge while managing complexity through efficient local computation.
Data Source
AI summary
Methods and systems for beam switch measurement and reporting are provided. In some embodiments, the method includes (1) receiving, by the terminal device, configuration information of beams currently used by the terminal device; (2) receiving, by the terminal device, configuration information of candidate beams; (3) performing, by the terminal device, a first measurement on the beams currently used by the terminal device and a second measurement on the candidate beams; and (4) generating, by the terminal device, a beam switch decision for the beams currently used by the terminal device by applying a neural network on results of the first and second the measurements.