Methods and systems for evaluating voice over Wi-Fi (VoWiFi) call quality

Patent metadata

Field	Value
Publication number	US-9992123-B2
Application number	US-201615199265-A
Country	US
Kind code	B2
Filing date	Jun 30, 2016
Priority date	Jun 30, 2016
Publication date	Jun 5, 2018
Grant date	Jun 5, 2018

Abstract

A method, a device, and a non-transitory storage medium provide receiving a plurality of voice call quality values and values for a plurality key performance indicators (KPIs) related to voice over Wi-Fi voice call quality; selecting a subset of the plurality of KPIs based on a correlation between each KPI and the voice call quality value; performing a plurality of discrete regression analyses based on the subsets of the plurality of KPIs and the voice call quality values to generate a plurality of regression results; determining an accuracy for each of the plurality of regression results; assigning weights to each of the plurality of regression results based on the determined accuracies; and combining the plurality of regression results using the assigned weights to generate a final combined estimated VoWiFi voice call quality algorithm that accurately predicts the voice call quality value based on values for the selected subset of the plurality of KPIs.

First claim

What is claimed is:

1. A computing-device implemented method comprising:
receiving a plurality of training data points, wherein each data point in the plurality of training data points includes a voice call quality value and values for a plurality of key performance indicators (KPIs) related to voice over wi-fi voice call quality;
determining, for each KPI of the plurality of KPIs and using the plurality of training data points, a correlation between the KPI and the voice call quality value;
selecting a subset of the plurality of KPIs based on the correlation;
performing a plurality of discrete regression analyses based on the subsets of the plurality of KPIs and the voice call quality values to generate a plurality of regression results;
determining an accuracy for each of the plurality of regression results;
assigning weights to each of the plurality of regression results based on the determined accuracies;
combining the plurality of regression results using the assigned weights to generate a final combined estimated voice over Wi-Fi (VoWiFi) voice call quality algorithm that accurately predicts the voice call quality value based on values for the selected subset of the plurality of KPIs;
receiving a new subset of the plurality of KPIs; and
determining the estimated VoWiFi voice call quality value based on the combined estimated VoWiFi voice call quality algorithm and the new subset of the plurality of KPIs.

2. The computing-device implemented method of claim 1 , wherein the voice call quality value comprises a traditionally determined Perceptual Objective Listening Quality Assessment (POLQA) score.

3. The computing-device implemented method of claim 1 , wherein the selected subset of the plurality of KPIs comprises at least two KPIs.

4. The computing-device implemented method of claim 3 , wherein the selected subset of the plurality of KPIs are selected from real time protocol (RTP) Packet loss percentage, RTP Jitter, RTP Latency, RTP Gap Ratio, RTP Session Duration, packet data convergence protocol (PDCP) throughput, radio link control (RLC) Throughput, the number of physical downlink shared channel physical resource blocks (PDSCH.PRB), a PDSCH block error rate (BER), and a handover happening indicator value based on the correlation of each KPI to the voice call quality value.

5. The computing-device implemented method of claim 1 , further comprising:
receiving the new subset of the plurality of KPIs from a mobile device; and
outputting, to the mobile device, an indication of the estimated VoWiFi voice call quality value.

6. The computing-device implemented method of claim 1 , further comprising:
determining the correlation using a Spearman correlation technique.

7. The computing-device implemented method of claim 1 , wherein performing the plurality of discrete regression analyses comprises performing a number of discrete regression analyses, wherein the number ranges from 3 to 8.

8. The computing-device implemented method of claim 7 , wherein the number of discrete regression analyses comprise a linear regression, a second order polynomial regression, a third order polynomial regression, a lasso regression, a ridge regression, an elastic regression, a generalized additive model regression, and an adaptive local weight scatterplot smoothing (A-LOESS) regression.

9. The computing-device implemented method of claim 7 , further comprising:
assigning different weights to the plurality of regression results based on the determined accuracy,
wherein the total of all weights combines to a value of one.

10. The computing-device implemented method of claim 9 , further comprising:
assigning an updated weight to the combined estimated VoWiFi voice call quality algorithm to generate the final estimated VoWiFi voice call quality algorithm.

11. The computing-device implemented method of claim 10 , wherein the updated weight is calculated using a quadratic programming technique based on the different weights assigned to plurality of regression results.

12. A device comprising:
a communication interface;
a memory, wherein the memory stores instructions; and
a processor, wherein the processor executes the instructions to:
receive a plurality of training data points, wherein each data point in the plurality of training data points includes a voice call quality value and values for a plurality key performance indicators (KPIs) related to voice over wi-fi voice call quality;
determine, for each KPI of the plurality of KPIs and using the plurality of training data points, a correlation between the KPI and the voice call quality value;
select a subset of the plurality of KPIs based on the correlation;
perform a plurality of discrete regression analyses based on the subsets of the plurality of KPIs and the voice call quality values to generate a plurality of regression results;
determine an accuracy for each of the plurality of regression results;
assign weights to each of the plurality of regression results based on the determined accuracies;
combine the plurality of regression results using the assigned weights to generate a final combined estimated VoWiFi voice call quality algorithm that accurately predicts the voice call quality value based on values for the selected subset of the plurality of KPIs;
receive, via the communication interface, a new subset of the plurality of KPIs; and
determine the estimated voice over Wi-Fi (VoWiFi) voice call quality value based on the combined estimated VoWiFi voice call quality algorithm and the new subset of the plurality of KPIs.

13. The device of claim 12 , wherein the voice call quality value comprises a traditionally determined Perceptual Objective Listening Quality Assessment (POLQA) score.

14. The device of claim 12 , wherein the selected subset of the plurality of KPIs comprises at least two KPIs selected from real time protocol (RTP) Packet loss percentage, RTP Jitter, RTP Latency, RTP Gap Ratio, RTP Session Duration, packet data convergence protocol (PDCP) throughput, radio link control (RLC) Throughput, the number of physical downlink shared channel physical resource blocks (PDSCH.PRB), a PDSCH block error rate (BER), and a handover happening indicator value based on the correlation of each KPI to the voice call quality value.

15. The device of claim 12 , wherein the processor to perform the plurality of discrete regression analyses further executes the instructions to perform a number of discrete regression analyses, wherein the number ranges from 3 to 8.

16. The device of claim 15 , wherein the number of discrete regression analyses comprise a linear regression, a second order polynomial regression, a third order polynomial regression, a lasso regression, a ridge regression, an elastic regression, a generalized additive model regression, and an adaptive local weight scatterplot smoothing (A-LOESS) regression.

17. The device of claim 15 , wherein the processor further executes the instructions to:
assign different weights to the plurality of regression results based on the determined accuracy,
wherein the total of all weights combines to a value of one.

18. The device of claim 17 , wherein the processor further executes the instructions to:
assign an updated weight to the combined estimated VoWiFi voice call quality algorithm to generate the final estimated VoWiFi voice call quality algorithm.

19. A non-transitory, computer-readable storage medium storing instructions executable by a processor of a computational device, which when executed cause the computational

Assignees

• Verizon Patent & Licensing Inc

Inventors

• Ouyang Ye
• Becht Carol
• Iyer Krishna Pichumani
• Su Le

Classifications

• H04L65/608

  Electricity · mapped topic

• H04W84/12

  WLAN [Wireless Local Area Networks] · CPC title

• H04W24/08

  Testing, {supervising or monitoring} using real traffic · CPC title

• H04L43/08

  Monitoring or testing based on specific metrics, e.g. QoS, energy consumption or environmental parameters · CPC title

• H04L47/2416 Primary

  Real-time traffic · CPC title