AI Confidence-Guided Beam Management for Reliable Selection

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing beam management systems struggle to reliably select a target beam pair from multiple optimal beam pairs, leading to potential incorrect predictions and increased measurement overheads, especially in AI-based beam management scenarios.

Innovation Solution

A method for beam management that selects a target beam resource based on confidences associated with each beam resource, using an AI model to evaluate and quantify the reliability of beam resources, reducing measurement overheads by scanning only when confidences are low.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If multiple optimal beam pairs are obtained by prediction based on AI model, then beam management efficiency is improved, but the reliability of target beam pair selection deteriorates due to inability to evaluate prediction confidence

Engineering Contradiction:

Improvebeam management efficiencyVSAvoidtarget beam pair selection reliability

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent introduces confidence information as an intermediary element between the AI model's beam pair predictions and the final selection decision. This confidence information acts as a mediator that enables reliable evaluation of prediction quality without requiring complete re-measurement of all beam pairs, thus resolving the contradiction between efficiency and reliability.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces the traditional mechanical approach of physically measuring and evaluating all beam pairs through exhaustive scanning with an AI-based predictive system that provides confidence estimates. This substitution allows the system to achieve both high efficiency through prediction and high reliability through confidence-based selection, eliminating the need for complete mechanical re-evaluation.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Reliability

If exhaustive beam scanning is performed to ensure reliable beam selection, then selection reliability is improved, but measurement overhead increases significantly

Engineering Contradiction:

Improvebeam selection reliabilityVSAvoidmeasurement overhead

Core Design Contradiction:

ReliabilityVSLoss of energy

Solution Approach 1:

The patent performs preliminary action by using the AI model to predict optimal beam pairs and their confidence levels before actual measurement occurs. This preliminary prediction filters out unlikely beam pairs, allowing the system to perform measurements only on a reduced set of candidate beams, thereby maintaining reliability while significantly reducing measurement overhead.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

Instead of performing complete exhaustive scanning of all possible beam pairs, the patent applies partial action by measuring only the subset of beam pairs that the AI model predicts to be optimal with high confidence. This partial measurement approach is sufficient to achieve reliable beam selection without the energy cost of full scanning.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250274928A1Method for beam management, and device

Publication Date: 2025.08.28 GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD

AI summary

Provided is a method for beam management. The method for beam management includes: selecting a target beam resource from K beam resources based on confidences respectively corresponding to the K beam resources, wherein K is an integer greater than or equal to 1.