5G Handover Success Prediction via Machine Learning Clustering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The increasing number of small cells in 5G wireless networks leads to a higher number of handovers, making traditional manual settings of handover parameters inefficient and inaccurate, resulting in poor user experience and resource wastage due to handoff and radio link failures.

Innovation Solution

The use of machine learning and artificial intelligence to analyze handover data, cluster cells with similar behaviors, predict future handover success rates, and automatically update network parameters to improve handover success rates without manual operator intervention.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If the number of small cells is increased to meet traffic demands, then network capacity and coverage are improved, but the number of handovers increases leading to more handoff failures and radio link failures

Engineering Contradiction:
Improvenetwork capacityVSAvoidhandover success rate
Core Design Contradiction:
ProductivityVSReliability

Solution Approach 1:

The system performs preliminary actions by predicting future handover success rates using machine learning models before handovers occur. Historical handover data is analyzed to forecast upcoming failures, allowing the system to proactively adjust handover parameters and configure cells to prevent failures before they happen, rather than reacting after failures occur.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements dynamic adjustment of handover parameters based on predicted failure patterns. Instead of static manual configuration, the machine learning model continuously forecasts handover success rates and automatically updates handover parameters in real-time to adapt to changing network conditions, traffic patterns, and cell behaviors.

Inventive Principle:
Principle #15Dynamics

2Measurement precision

If manual setting of handover parameters is performed for each cell, then parameter accuracy may be improved, but significant manpower and time are required making the process inefficient

Engineering Contradiction:
Improvehandover parameter accuracyVSAvoidparameter configuration time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The system enables self-service by allowing the network to automatically configure and optimize its own handover parameters through machine learning. The ML model autonomously analyzes historical data, predicts future performance, and adjusts parameters without human intervention, making the network self-optimizing and eliminating the need for manual operator configuration.

Inventive Principle:
Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical manual process of parameter setting with an automated intelligent system. Instead of operators manually analyzing data and configuring parameters cell-by-cell, a machine learning model automatically performs prediction and optimization, substituting human manual work with computational intelligence that processes data faster and more accurately.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

3Ease of operation

If traditional manual methods are used to set handover parameters, then operational simplicity is maintained, but the process becomes inefficient and inaccurate for large numbers of small cells

Engineering Contradiction:
Improveparameter setting simplicityVSAvoidparameter configuration efficiency
Core Design Contradiction:
Ease of operationVSProductivity

Solution Approach 1:

The machine learning system provides a universal solution that handles parameter optimization for all cells simultaneously through a single automated process. Instead of requiring separate manual configuration for each cell, the ML model processes multiple cells in parallel, applying learned patterns across the entire network to achieve both simplicity and high productivity.

Inventive Principle:
Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system automatically changes handover parameters based on data-driven insights from the machine learning model. Instead of relying on operator judgment and manual adjustment, the system dynamically modifies parameters such as handover thresholds, timing, and cell configuration based on predicted failure patterns, achieving both ease of operation and improved efficiency.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP3972339A1Handover success rate prediction and management using machine learning for 5g networks
Publication Date: 2022.03.23 ACCENTURE GLOBAL SOLUTIONS LTD
  • EP3972339A1 patent drawingFigure 1
  • EP3972339A1 patent drawingFigure 2
  • EP3972339A1 patent drawingFigure 3

AI summary

Aspects of the present disclosure provide systems, methods, and computer-readable storage media that leverage artificial intelligence and machine learning to predict future handover success rates of cells of a wireless network and to determine updated network parameter values that may be used to improve the future handover success rates. Clustering algorithms may be applied to handover success rates associated with the cells to group the cells into different clusters. Machine learning (ML) models may be trained based on historical handover data associated with cells of the clusters to predict future handover success rates based on current handover success rates. The output of the ML models may be used to determine one or more updated network parameter values to be implemented by the wireless network to improve future handover success rates at one or more cells.