ML-Based Wireless Connectivity Management for 5G NR Outage Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current 5G networks face challenges in maintaining high Time on NR (ToNR) due to complex deployment scenarios, network mobility strategies, and varying UE capabilities, leading to convoluted measurement and optimization processes and potential NR Radio Link Failures.
Innovation Solution
A method using a management node and Machine Learning (ML) models to predict signal strengths for viable LTE and NR cell combinations, identifying optimal cell pairs for handover when the NR signal strength falls below a threshold, thereby reducing NR outage areas and improving ToNR.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional measurement and optimization processes are used for ToNR, then network coverage can be improved, but the process becomes highly convoluted and difficult to manage
Solution Approach 1:
The patent replaces traditional manual measurement and optimization processes with Machine Learning models that automatically predict signal strengths and identify optimal cell combinations. The ML models process location information and network data to determine handover decisions, substituting complex human-driven optimization with automated intelligent algorithms.
2Reliability
If NR cell coverage area is reduced to optimize signal quality, then connection reliability improves, but the area where NR connectivity is available decreases
Solution Approach 1:
The patent performs preliminary handover to LTE cells before NR signal is completely lost by predicting future signal strengths using ML models. When the model predicts that NR signal will fall below threshold, the system proactively initiates handover to a suitable LTE-NR cell combination, preventing NR connection failure before it occurs.
Solution Approach 2:
The system uses ML models that continuously analyze location information, signal strength measurements, and network conditions to provide feedback on predicted signal strengths. This feedback loop enables dynamic adjustment of handover decisions based on real-time conditions, optimizing the balance between coverage area and connection reliability.
3Reliability
If handover operations are increased to maintain NR connectivity, then Time on NR improves, but network overhead and complexity increase
Solution Approach 1:
The patent implements self-service through ML models that autonomously predict signal strengths and identify optimal handover targets without requiring extensive human intervention. The system automatically processes location data, evaluates multiple cell combinations, and makes handover decisions based on predicted outcomes, reducing the need for manual network optimization.
4Measurement precision
If ML prediction is used for signal strength, then handover accuracy improves, but computational requirements increase
Solution Approach 1:
The patent applies partial action by using ML models selectively - only when location information is available and when prediction can provide value. The system processes location data and network conditions through ML models to predict signal strengths for candidate cell combinations, applying computational resources only where needed to improve handover accuracy rather than continuously.
Data Source
AI summary
The present application relates to a computer implemented method for managing connectivity of a wireless device in a cellular communication network, wherein the wireless device is operable to connect to a cell of a first radio-access technology, RAT, and to a cell of a second RAT. The method includes receiving location information for the wireless device, wherein the wireless device is connected to a first RAT cell hosted by the serving radio access node, and is also connected to a second RAT cell, and wherein a signal strength of the second RAT cell, received at the wireless device, has fallen below a trigger threshold. The method further includes identifying neighbour first and second RAT cells, and assembling, a candidate set of cell combinations.