Graph Neural Network Beam Tracking in 5G NR

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

UE beam tracking in 5G NR networks is complex, leading to reduced throughput, and conventional methods require additional prior information such as sensor measurements or AoA estimates, further increasing complexity.

Innovation Solution

A learning-based approach using a Graph Neural Network (GNN) for beam tracking, which predicts future beams based on previous beams and RSRP measurements without requiring additional prior information.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional beam tracking methods using additional prior information (sensor measurements, AoA estimates) are used, then measurement precision may be improved, but device complexity increases

Engineering Contradiction:

Improvebeam tracking precisionVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts and removes the requirement for additional prior information (sensor measurements, AoA estimates) from the beam tracking system. The GNN-based approach achieves accurate beam tracking using only historical beam indices and RSRP measurements that are already available in the system, eliminating the need for extra sensors and complex AoA estimation algorithms.

Inventive Principle:

Principle #2Taking out (Extraction)

Solution Approach 2:

The system serves itself by utilizing its existing data (historical beam indices and RSRP measurements) to train the GNN model for beam tracking. The GNN learns patterns from the system's own operational history and makes predictions without requiring external prior information, making the system self-sufficient and reducing overall complexity.

Inventive Principle:

Principle #25Self-service

2Manufacturing precision

If UE beam tracking via P3 procedure is performed, then beam refinement is achieved, but complexity increases leading to reduced throughput

Engineering Contradiction:

Improvebeam refinement accuracyVSAvoidthroughput

Core Design Contradiction:

Manufacturing precisionVSProductivity

Solution Approach 1:

The GNN model performs preliminary beam prediction by forecasting future optimal beams based on historical data before actual beam tracking is needed. This advance prediction reduces the computational burden during real-time operation, allowing faster beam selection and improving throughput while maintaining refinement accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent applies partial action by using a simplified GNN inference process that leverages pre-trained models and historical patterns. Instead of performing complete complex beam tracking calculations in real-time, the system uses the GNN to predict beams with reduced computational steps, achieving sufficient accuracy while improving throughput.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250055549A1System and method for beam tracking using graph neural networks

Publication Date: 2025.02.13 SAMSUNG ELECTRONICS CO LTD

AI summary

A system and a method are disclosed for performing beam tracking using a GNN. A method of beam prediction by a UE includes receiving an input feature vector; comparing the input feature vector to a lookup table of rated predicted beams; selecting a corresponding predicted beam having a highest rating from the lookup table; and receiving, from a base station, a signal using the selected predicted beam.