Neural Network Traffic Quality Estimation in Wireless Networks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for user plane traffic quality analysis in wireless communication networks face limitations, including inability to perform root cause analysis, lack of end-to-end coverage, and dependency on RTCP reports and user equipment availability.
Innovation Solution
An apparatus and method that utilize a neural network-based system to estimate user plane traffic quality by probing bidirectional traffic flows, allowing for end-to-end quality assessment without relying on RTCP reports or specific user equipment configurations.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If user plane probes are deployed at gateway nodes to monitor RTP/RTCP traffic, then end-to-end coverage is achieved, but the system becomes dependent on RTCP reports and user equipment availability
Solution Approach 1:
The patent introduces a neural network-based estimation system as an intermediary that infers downlink quality metrics from uplink probe data. This mediator eliminates the need for direct RTCP report collection from user equipment, allowing quality assessment without depending on UE availability or RTCP report generation.
Solution Approach 2:
The patent replaces the traditional mechanical approach of directly measuring downlink traffic quality through RTCP reports with a neural network-based estimation system. This substitution uses machine learning models to predict quality metrics based on uplink measurements, eliminating the need for complex bidirectional probing infrastructure.
2Measurement precision
If detailed RTP packet inspection is performed at probes, then measurement precision is improved, but device complexity and processing requirements increase
Solution Approach 1:
The patent extracts only the essential uplink probe measurements (packet loss, jitter, duration) that are sufficient for quality assessment, eliminating the need for comprehensive bidirectional RTP packet inspection. This extraction approach maintains measurement precision while significantly reducing processing complexity.
Solution Approach 2:
The patent uses neural network models to create virtual copies of downlink quality measurements based on uplink probe data. Instead of actually measuring downlink traffic directly, the system generates estimated quality metrics through machine learning, reducing the need for complex inspection infrastructure.
3Ease of operation
If RTCP reports are used for quality monitoring, then ease of operation is improved, but loss of information occurs when RTCP reports are unavailable
Solution Approach 1:
The patent enables the uplink probe system to self-service by inferring downlink quality information from its own uplink measurements. The neural network model processes uplink probe data and automatically generates estimates for downlink quality metrics, eliminating the need for separate RTCP report collection and preventing information loss when RTCP is unavailable.
Data Source
AI summary
An apparatus for user plane traffic quality analysis in a wireless communication network. The apparatus includes an interface configured to be coupled to a user plane probe arranged on a bidirectional user plane traffic flow path of user plane traffic flowing through the wireless communication network between a first terminal and a second terminal of the wireless communication network, and an estimation unit coupled to the interface.