ML-Based Secondary Cell Selection for Carrier Aggregation Balancing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing carrier aggregation systems fail to optimally distribute user equipment among available cells, leading to overloading and suboptimal user throughput due to uniform cell selection without considering load and radio quality variations.

Innovation Solution

A network node device employs machine learning models, specifically reinforcement learning and recurrent neural networks, to predict and dynamically set secondary cell weight parameters based on downlink throughput and spectral efficiency, enabling optimal secondary cell selection for carrier aggregation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If all UEs are assigned the same set of primary and secondary cells, then cell configuration is simplified, but cells become fully loaded and user throughput deteriorates

Engineering Contradiction:
Improvecell configuration simplicityVSAvoiduser throughput
Core Design Contradiction:
Ease of manufactureVSProductivity

Solution Approach 1:

The patent applies local quality by assigning different secondary cell sets to different UEs based on their specific characteristics (mobility, service type, location). Instead of uniform cell configuration, the system customizes cell assignments for each UE group, thereby preventing cell overload and optimizing throughput for each user's specific needs.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent implements dynamics by enabling the network to dynamically adjust secondary cell assignments based on real-time conditions. The system can modify which UEs aggregate which secondary cells according to changing load conditions and radio quality, allowing the system to adapt to varying demands and maintain optimal performance.

Inventive Principle:
Principle #15Dynamics

2Quantity of substance

If the network adds more cells than UEs can aggregate, then spectrum resources are increased, but load distribution becomes suboptimal without individual cell consideration

Engineering Contradiction:
Improvespectrum resourcesVSAvoidload distribution efficiency
Core Design Contradiction:
Quantity of substanceVSProductivity

Solution Approach 1:

The patent applies parameter changes by using machine learning models to predict and optimize cell weight parameters. The system adjusts secondary cell weight parameters based on predicted load and radio quality conditions, enabling intelligent load distribution across available spectrum resources rather than uniform allocation.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent implements feedback mechanisms where the network monitors actual cell performance and uses this information to refine machine learning predictions. The system continuously learns from observed data to improve its ability to distribute UEs across cells, optimizing load distribution as more cells are added to the network.

Inventive Principle:
Principle #23Feedback

3Productivity

If machine learning models are used to predict and optimize secondary cell weights, then downlink throughput and spectral efficiency are improved, but system complexity increases

Engineering Contradiction:
Improvedownlink throughputVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies self-service by implementing self-learning mechanisms where the machine learning models automatically adapt to network conditions without manual reconfiguration. The system uses historical data and real-time observations to autonomously optimize cell weight parameters, reducing the need for complex manual intervention while maintaining high throughput performance.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP4589875A1Machine learning -based secondary cell selection for carrier aggregation
Publication Date: 2025.07.23 NOKIA SOLUTIONS & NETWORKS OY
  • EP4589875A1 patent drawingFigure 1
  • EP4589875A1 patent drawingFigure 2
  • EP4589875A1 patent drawingFigure 3

AI summary

Devices, methods and computer programs for machine learning (ML) -based secondary cell selection for carrier aggregation (CA) are disclosed. At least some example embodiments may allow dynamically predicting a secondary cell weight parameter which may influence a network node decision in selecting secondary cells for user devices.