Mobile Network QoE Modeling With KPI Decoupling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional network analytics systems lack the ability to provide detailed troubleshooting for user-perceived service quality issues and require extensive training data and tailored models for each service provider, especially in 5G networks, making them costly and difficult to maintain.
Innovation Solution
A unified QoE framework that calculates QoE metrics using a generic loss model, KPI coupling, and machine learning parameter optimization, independent of specific service providers, to estimate service degradation based on accessibility, integrity, and retainability factors, using a scalar value on a standardized scale.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional NA systems continuously monitor basic network KPIs, then network performance can be tracked, but they cannot provide adequate detailed troubleshooting or identify user-perceived service quality issues
Solution Approach 1:
The patent segments the service quality assessment into multiple dimensions by introducing QoE metrics that separately evaluate accessibility, integrity, and retainability factors. This segmentation allows detailed troubleshooting of specific service quality aspects while maintaining comprehensive monitoring coverage.
Solution Approach 2:
The patent introduces QoE metrics as an intermediary layer between basic network KPIs and user-perceived service quality. These QoE metrics translate raw network data into meaningful service quality assessments, bridging the information gap without requiring direct access to user devices.
2Measurement precision
If tailored QoE models are created for each service provider, then accurate service quality assessment can be achieved, but the system becomes costly and difficult to maintain
Solution Approach 1:
The patent creates a universal QoE framework that can assess service quality across multiple service providers and service types using a single standardized methodology. The framework defines common metrics (accessibility, integrity, retainability) that apply universally, eliminating the need for provider-specific models while maintaining assessment accuracy.
Solution Approach 2:
The patent uses parameter-based configuration to adapt the universal QoE framework to different service types and providers. By changing parameters such as metric weights, thresholds, and evaluation criteria rather than creating entirely new models, the system maintains flexibility and accuracy across diverse services without increasing maintenance complexity.
3Measurement precision
If service-specific QoE models are developed for each service type, then accurate QoE calculation can be achieved, but extensive training data is required for each model
Solution Approach 1:
The patent develops a universal QoE model that calculates service quality across all service types using a single set of core metrics and algorithms. This universal approach eliminates the need to collect and process extensive training data for each individual service type, as the model learns general service quality patterns that apply across diverse services.
Solution Approach 2:
The patent adapts the universal QoE model to different service types by adjusting parameters such as metric importance weights and evaluation thresholds rather than retraining entire models. This parameter-based adaptation requires minimal service-specific data while maintaining calculation accuracy for voice, video, data, and other service types.
Data Source
AI summary
Provided herein is a unified Quality of Experience (QoE) framework for determining QoE metrics for a plurality of mobile services based, in part, on the Key Performance Indicators (KPIs) that are collected for the services. The QoE metrics are scalar values, fall on a unified scale, and eliminate any dependencies that may exist between the KPIs used in determining the QoE metrics. The unified QoE framework includes a generic QoE calculation module having a loss model component, a KPI coupling calculation component, and a machine learning (ML) parameter optimization component. The QoE calculation module correlates network event information and calculates various resource and/or network KPIs. Additionally, the QoE calculation module calculates service KPIs for a predetermined number of traffic types, and estimates factors due to losses, drops, and soft drops. Additionally, the internal functional parameters of the underlying KPI are determined and/or optimized without requiring external intervention or the initial setting of parameters.