Carrier Switching Model for Mobile Network Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile virtual network operators (MVNOs) face challenges in minimizing the frequency of carrier network switches due to increased battery consumption and downtime, as existing methods do not accurately predict network quality changes and may result in worse network experiences when switching from non-LTE to LTE connections.
Innovation Solution
Implementing a carrier switching model using machine learning algorithms that predict signal strength and radio access technology over a threshold period, based on crowd-sourced user experience ratings and device-specific features, to determine optimal network switches and minimize adverse user experiences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If the UE switches carrier networks frequently to maintain best connection quality, then the network quality is improved, but battery consumption increases and downtime accumulates
Solution Approach 1:
The system performs preliminary actions by predicting future network quality metrics (signal strength, throughput, latency) before actually switching carriers. The machine learning model forecasts network conditions over a threshold period, allowing the UE to anticipate quality degradation and switch proactively rather than reactively, reducing unnecessary switches and associated battery consumption.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring actual network performance metrics and comparing them against predicted values. This feedback loop allows the UE to learn from actual user experiences and refine its switching decisions, ensuring that carrier switches are only executed when predicted improvements are likely to materialize, thereby reducing unnecessary switching and battery drain.
2Reliability
If the UE switches carrier networks to improve connection quality, then network performance is improved, but downtime increases and active connections are broken
Solution Approach 1:
The system performs preliminary actions by predicting future network quality metrics (signal strength, throughput, latency) before actually switching carriers. The machine learning model forecasts network conditions over a threshold period, allowing the UE to anticipate quality degradation and switch proactively rather than reactively, reducing unnecessary switches and associated battery consumption.
Solution Approach 2:
The system dynamically adjusts switching behavior based on real-time network conditions and predicted performance. The UE monitors current network metrics alongside predictions and only executes switches when the predicted improvement exceeds a threshold, making the switching decision dynamic and context-aware rather than static or overly frequent.
3Adaptability or versatility
If the UE switches from non-LTE to LTE carrier networks, then radio access technology is improved, but network quality prediction accuracy decreases
Solution Approach 1:
The system handles parameter changes by training the machine learning model on diverse network conditions including transitions between different radio access technologies. The model learns to accurately predict network quality metrics across various RATs (LTE, non-LTE) by incorporating technology-specific parameters and transition patterns, enabling accurate predictions even when switching between different network technologies.
Data Source
AI summary
A method includes receiving a carrier switching model, determining that user equipment (UE) is transitioning between carrier cells based on at least one connection performance metric of the UE, and obtaining a current UE location and any available carrier networks of carriers relative to the current UE location. The method also includes determining a carrier switch score for each available carrier network using the carrier switching model and determining to switch carrier networks based on the carrier switch score of a currently-connected available carrier network relative to the carrier switch score of another available carrier network. Each carrier switch score is based on at least one of a predicted signal strength or a predicted radio access technology over a threshold period of time for the respective available carrier network. The method also includes causing the UE to switch connection to the other available carrier network.