AI Mobility Management for Low-Latency 5G Handover Continuity

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

5G mobile networks face challenges in maintaining session continuity and low latency during handovers due to frequent changes in UPF serving areas, especially with the introduction of edge computing, leading to potential session interruptions and packet loss.

Innovation Solution

The introduction of an Intelligent Mobility Element (IME) that utilizes machine learning and AI to anticipate handovers, making dynamic resource pre-allocation decisions to maintain session continuity by relocating IP addresses and application sessions proactively, ensuring seamless transitions between cell sites.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If machine learning models are trained on large volumes of network data, then model accuracy and prediction capability are improved, but training time and computational resources are increased

Engineering Contradiction:
Improvemodel accuracyVSAvoidtraining time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies preliminary action by pre-training machine learning models on historical network data before deployment. Models are trained in advance on labeled historical data to establish baseline accuracy, then fine-tuned with smaller amounts of additional data when needed, avoiding the need to train from scratch each time predictions are required.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the training process into multiple phases: initial pre-training on historical data, followed by fine-tuning on smaller targeted datasets. This segmentation allows the system to achieve high accuracy through pre-training while minimizing the time and resources needed for subsequent model updates and adaptations.

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If machine learning models are trained on large volumes of network data, then model accuracy and prediction capability are improved, but computational resources and energy consumption are increased

Engineering Contradiction:
Improvemodel accuracyVSAvoidcomputational resources
Core Design Contradiction:
Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The patent applies preliminary action by pre-training machine learning models on historical network data before deployment. Models are trained in advance on labeled historical data to establish baseline accuracy, then fine-tuned with smaller amounts of additional data when needed, avoiding the need to train from scratch each time predictions are required.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent segments the training process into multiple phases: initial pre-training on historical data, followed by fine-tuning on smaller targeted datasets. This segmentation allows the system to achieve high accuracy through pre-training while minimizing the time and resources needed for subsequent model updates and adaptations.

Inventive Principle:
Principle #1Segmentation

3Reliability

If mobility management functions are centralized in the core network, then network control and optimization are improved, but latency and response time for mobile devices are increased

Engineering Contradiction:
Improvenetwork controlVSAvoidlatency
Core Design Contradiction:
ReliabilityVSSpeed

Solution Approach 1:

The patent segments mobility management functions between centralized network entities (AMF, SMF) and distributed network edge components (RAN, UPF). Critical functions like authentication and session management remain centralized for reliability, while data forwarding and local policy enforcement are distributed to reduce latency.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces the User Plane Function (UPF) as an intermediary component that bridges the centralized control plane and the mobile device. The UPF handles data plane operations locally at the network edge, acting as a mediator that maintains centralized control while enabling fast local responses to reduce latency.

Inventive Principle:
Principle #24Intermediary (Mediator)

4Speed

If handover processes are simplified for faster connection switching, then mobility speed and user experience are improved, but connection stability and data loss prevention are worsened

Engineering Contradiction:
Improvehandover speedVSAvoidconnection stability
Core Design Contradiction:
SpeedVSReliability

Solution Approach 1:

The patent applies preliminary action by performing handover preparation steps before the actual handover occurs. The target gNB is pre-configured with necessary context information, and resource allocation is prepared in advance, enabling the handover to execute quickly without compromising stability through rushed or incomplete procedures.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback mechanisms during handover where the network monitors handover progress and device status in real-time. If issues are detected during the simplified handover process, the system can trigger corrective actions or rollback procedures to maintain connection stability while preserving the overall speed advantage of simplified handover.

Inventive Principle:
Principle #23Feedback

Data Source

PatentEP4309417B1Intelligent mobility in communication networks
Publication Date: 2026.05.13 BOOST SUBSCRIBERCO LLC
  • EP4309417B1 patent drawingFigure 1
  • EP4309417B1 patent drawingFigure 2
  • EP4309417B1 patent drawingFigure 3

AI summary

Embodiments are directed towards facilitating intelligent mobility including session and handover management within communications networks. Handovers for user devices in a network are predicted, current and predicted status data of the network are obtained, and a tradeoff between a cost for pre-allocation of resources prior to predicted handovers and a gain in quality of service is analyzed based on the current and predicted status data to cause resource pre-allocation in the network.