Mobile Bandwidth Prediction Using UE Usage and Geolocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Mobile Virtual Network Operators (MVNOs) lack sufficient insight into UE usage patterns due to limited data access from Mobile Network Operators (MNOs, hindering accurate prediction and management of bandwidth usage.
Innovation Solution
A system utilizing artificial intelligence (AI) to predict bandwidth usage by analyzing data such as previous usage, geolocation, and media events, creating embeddings to encode interactions, and clustering activities to determine optimal plans for bandwidth allocation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If MVNO requests bandwidth based on limited available data, then bandwidth allocation can be simplified, but prediction accuracy deteriorates
Solution Approach 1:
The patent introduces an intermediary system that acts as a bridge between MNO and MVNO. This intermediary processes and analyzes detailed UE usage data (including geolocation, media events, and usage patterns) that would otherwise be inaccessible to MVNO, transforming raw data into actionable bandwidth predictions while maintaining data privacy through controlled access
Solution Approach 2:
The patent segments the bandwidth prediction process into multiple components: data collection from MNO, AI/ML-based analysis of usage patterns, geolocation-based predictions, media event correlations, and aggregated bandwidth forecasting. This segmentation allows MVNO to achieve accurate predictions using processed insights rather than raw data
2Measurement precision
If MNO provides detailed UE data to MVNO, then prediction accuracy improves, but data security and privacy control worsen
Solution Approach 1:
The patent employs an intermediary architecture where detailed UE data remains under MNO control while enabling MVNO to access processed insights. The intermediary system performs AI/ML analysis and aggregates data to produce bandwidth predictions without exposing raw personal information to MVNO, thus maintaining both accuracy and privacy
Solution Approach 2:
The patent creates processed copies of UE usage data through AI/ML models that capture usage patterns and behaviors without revealing actual personal information. These synthetic representations preserve predictive value while eliminating privacy risks associated with sharing raw data
3Device complexity
If MVNO uses traditional bandwidth estimation methods, then system complexity remains low, but resource optimization worsens
Solution Approach 1:
The patent transforms the bandwidth prediction approach by changing key parameters: incorporating geolocation data, media event information, and detailed usage patterns into the prediction model. These parameter changes enable more accurate predictions and better resource optimization, with complexity managed through automated AI/ML processing
Data Source
AI summary
The system obtains data associated with UE, representing interaction between the UE and a network. The data includes: previous bandwidth usage associated with the UE, previous CDR associated with the UE, anticipated geolocation of the UE, a plan associated with the UE, one or more media events, a number of lines associated with the UE, a length of time the UE has been associated with the network, and a unique identifier associated with the UE. The system obtains multiple plans associated with the network, where a plan indicates the bandwidth usage associated with the UE within a predetermined period. Based on the data, the system predicts the bandwidth usage associated with the UE within the predetermined period to obtain a predicted bandwidth usage and determines the plan among the multiple plans accommodating the predicted bandwidth usage. The system requests the bandwidth usage associated with the plan from the network.